Suche

21_station_wedel_21.3 (Letzte Änderung: Mon Mar 20 20:44:35 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'21_station_wedel_21.3'* ### Description for Sequence Group *'21_station_wedel'* The ego train is standing on a platform at the end of a track. To the left of it is another train. Common object classes are track and person. ### Statistics for Sequence *'21_station_wedel_21.3'* Number of Multisensor-Frames: `40` Total file size: `3.8 GB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 1 672 crowd 0 train 280 wagons 0 bicycle 32 group of bicycles 80 motorcycle 0 road vehicle 122 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 280 transition 0 switch 0 catenary pole 0 signal pole 0 signal 130 signal bridge 0 buffer stop 200 flame 0 smoke 0 ------------------ --------------------- total 2 796

21_station_wedel_21.2 (Letzte Änderung: Mon Mar 20 20:44:21 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'21_station_wedel_21.2'* ### Description for Sequence Group *'21_station_wedel'* The ego train is standing on a platform at the end of a track. To the left of it is another train. Common object classes are track and person. ### Statistics for Sequence *'21_station_wedel_21.2'* Number of Multisensor-Frames: `100` Total file size: `9.4 GB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 3 607 crowd 0 train 700 wagons 0 bicycle 60 group of bicycles 200 motorcycle 0 road vehicle 354 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 700 transition 0 switch 0 catenary pole 0 signal pole 0 signal 331 signal bridge 0 buffer stop 500 flame 0 smoke 0 ------------------ --------------------- total 6 452

21_station_wedel_21.1 (Letzte Änderung: Mon Mar 20 20:44:07 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'21_station_wedel_21.1'* ### Description for Sequence Group *'21_station_wedel'* The ego train is standing on a platform at the end of a track. To the left of it is another train. Common object classes are track and person. ### Statistics for Sequence *'21_station_wedel_21.1'* Number of Multisensor-Frames: `100` Total file size: `9.4 GB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 2 441 crowd 0 train 700 wagons 0 bicycle 0 group of bicycles 200 motorcycle 0 road vehicle 176 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 700 transition 0 switch 0 catenary pole 0 signal pole 0 signal 330 signal bridge 0 buffer stop 500 flame 0 smoke 0 ------------------ --------------------- total 5 047

20_vegetation_squirrel_20.1 (Letzte Änderung: Mon Mar 20 20:43:45 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'20_vegetation_squirrel_20.1'* ### Description for Sequence Group *'20_vegetation_squirrel'* The ego train is heading for a curve in a section of track with a lot of vegetation. An animal (squirrel) is on the tracks of the ego train. Common object classes are track and catenary pole. ### Statistics for Sequence *'20_vegetation_squirrel_20.1'* Number of Multisensor-Frames: `10` Total file size: `848 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 9 crowd 0 train 0 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 0 animal 50 group of animals 0 wheelchair 0 drag shoe 0 track 50 transition 0 switch 0 catenary pole 145 signal pole 0 signal 0 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 254

19_vegetation_curve_19.1 (Letzte Änderung: Mon Mar 20 20:43:31 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'19_vegetation_curve_19.1'* ### Description for Sequence Group *'19_vegetation_curve'* The ego train runs through a curve in a section of track with a lot of vegetation. Apart from the track, no other objects can be seen. ### Statistics for Sequence *'19_vegetation_curve_19.1'* Number of Multisensor-Frames: `10` Total file size: `843 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 0 crowd 0 train 0 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 0 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 78 transition 0 switch 0 catenary pole 0 signal pole 0 signal 0 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 78

18_vegetation_switch_18.1 (Letzte Änderung: Mon Mar 20 20:43:17 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'18_vegetation_switch_18.1'* ### Description for Sequence Group *'18_vegetation_switch'* The ego train travels through a section of track with a lot of vegetation towards a tunnel. Common object classes are track and catenary pole. ### Statistics for Sequence *'18_vegetation_switch_18.1'* Number of Multisensor-Frames: `10` Total file size: `894 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 0 crowd 0 train 0 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 0 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 180 transition 80 switch 40 catenary pole 883 signal pole 62 signal 94 signal bridge 0 buffer stop 50 flame 0 smoke 0 ------------------ --------------------- total 1 389

17_signal_bridge_17.1 (Letzte Änderung: Mon Mar 20 20:43:04 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'17_signal_bridge_17.1'* ### Description for Sequence Group *'17_signal_bridge'* The ego train travels through a narrow section of track with a noise barrier on the left and vegetation on the right towards a bridge. Common object classes are track and signal. ### Statistics for Sequence *'17_signal_bridge_17.1'* Number of Multisensor-Frames: `10` Total file size: `848 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 0 crowd 0 train 0 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 0 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 70 transition 0 switch 30 catenary pole 115 signal pole 110 signal 120 signal bridge 40 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 485

16_under_bridge_16.1 (Letzte Änderung: Mon Mar 20 20:42:52 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'16_under_bridge_16.1'* ### Description for Sequence Group *'16_under_bridge'* The ego train is standing next to a bridge in an urban section of track. Two oncoming trains can be seen. Common object classes are track and catenary pole. ### Statistics for Sequence *'16_under_bridge_16.1'* Number of Multisensor-Frames: `10` Total file size: `820 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 0 crowd 0 train 130 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 0 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 294 transition 58 switch 50 catenary pole 639 signal pole 106 signal 170 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 1 447

15_construction_vehicle_15.1 (Letzte Änderung: Mon Mar 20 20:37:49 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'15_construction_vehicle_15.1'* ### Description for Sequence Group *'15_construction_vehicle'* The ego train passes two other trains through a section of track with a lot of vegetation. One of the two trains is a construction vehicle. Common object classes are track, signal and catenary pole. ### Statistics for Sequence *'15_construction_vehicle_15.1'* Number of Multisensor-Frames: `10` Total file size: `951 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 64 crowd 0 train 142 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 73 animal 20 group of animals 0 wheelchair 0 drag shoe 0 track 170 transition 10 switch 10 catenary pole 370 signal pole 150 signal 372 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 1 381

14_signals_station_14.3 (Letzte Änderung: Mon Mar 20 20:37:36 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'14_signals_station_14.3'* ### Description for Sequence Group *'14_signals_station'* The ego train passes a platform. Since the camera is looking backwards in this sequence, the ego train moves further away from the platform with each subsequence. A time interval between the subsequences cannot be determined. Common object classes are track, signal and person (passenger). ### Statistics for Sequence *'14_signals_station_14.3'* Number of Multisensor-Frames: `10` Total file size: `816 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 24 crowd 0 train 130 wagons 70 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 631 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 310 transition 0 switch 0 catenary pole 949 signal pole 399 signal 647 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 3 160

14_signals_station_14.2 (Letzte Änderung: Mon Mar 20 20:37:20 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'14_signals_station_14.2'* ### Description for Sequence Group *'14_signals_station'* The ego train passes a platform. Since the camera is looking backwards in this sequence, the ego train moves further away from the platform with each subsequence. A time interval between the subsequences cannot be determined. Common object classes are track, signal and person (passenger). ### Statistics for Sequence *'14_signals_station_14.2'* Number of Multisensor-Frames: `10` Total file size: `917 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 88 crowd 0 train 110 wagons 40 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 100 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 320 transition 20 switch 116 catenary pole 1 020 signal pole 709 signal 1 461 signal bridge 0 buffer stop 70 flame 0 smoke 0 ------------------ --------------------- total 4 054

14_signals_station_14.1 (Letzte Änderung: Mon Mar 20 20:37:08 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'14_signals_station_14.1'* ### Description for Sequence Group *'14_signals_station'* The ego train passes a platform. Since the camera is looking backwards in this sequence, the ego train moves further away from the platform with each subsequence. A time interval between the subsequences cannot be determined. Common object classes are track, signal and person (passenger). ### Statistics for Sequence *'14_signals_station_14.1'* Number of Multisensor-Frames: `10` Total file size: `890 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 981 crowd 0 train 78 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 9 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 234 transition 0 switch 91 catenary pole 760 signal pole 334 signal 940 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 3 427

13_station_ohlsdorf_13.1 (Letzte Änderung: Mon Mar 20 20:35:25 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'13_station_ohlsdorf_13.1'* ### Description for Sequence Group *'13_station_ohlsdorf'* The ego train is heading for a platform. Other stationary trains can be seen. Common object classes are rails, signals and people (passengers). ### Statistics for Sequence *'13_station_ohlsdorf_13.1'* Number of Multisensor-Frames: `10` Total file size: `960 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 421 crowd 0 train 103 wagons 0 bicycle 64 group of bicycles 0 motorcycle 0 road vehicle 0 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 335 transition 0 switch 177 catenary pole 0 signal pole 295 signal 644 signal bridge 0 buffer stop 20 flame 0 smoke 0 ------------------ --------------------- total 2 059

12_vegetation_steady_12.1 (Letzte Änderung: Mon Mar 20 20:35:06 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'12_vegetation_steady_12.1'* ### Description for Sequence Group *'12_vegetation_steady'* The ego train stands on a section of track, which is surrounded by trees on both sides. Common object classes are track and signal. ### Statistics for Sequence *'12_vegetation_steady_12.1'* Number of Multisensor-Frames: `98` Total file size: `10.2 GB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 730 crowd 0 train 0 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 0 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 1 666 transition 0 switch 0 catenary pole 0 signal pole 1 176 signal 2 542 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 6 114

11_main_station_11.1 (Letzte Änderung: Mon Mar 20 20:34:50 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'11_main_station_11.1'* ### Description for Sequence Group *'11_main_station'* The ego train pulls up to a platform in a main station. There are other stationary trains in the station. A common object class is person (passenger). ### Statistics for Sequence *'11_main_station_11.1'* Number of Multisensor-Frames: `10` Total file size: `822 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 848 crowd 50 train 117 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 0 animal 29 group of animals 0 wheelchair 0 drag shoe 0 track 40 transition 0 switch 0 catenary pole 0 signal pole 0 signal 83 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 1 167

10_station_suelldorf_10.1 (Letzte Änderung: Mon Mar 20 20:34:17 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'10_station_suelldorf_10.1'* ### Description for Sequence Group *'10_station_suelldorf'* The ego train first travels through a section of forest and then passes a platform. The camera looks backwards in this sequence. Common object classes are track, signal and persons (passenger). ### Statistics for Sequence *'10_station_suelldorf_10.1'* Number of Multisensor-Frames: `10` Total file size: `874 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 411 crowd 0 train 0 wagons 0 bicycle 70 group of bicycles 39 motorcycle 0 road vehicle 0 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 80 transition 0 switch 40 catenary pole 0 signal pole 227 signal 401 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 1 268

9_station_ruebenkamp_9.7 (Letzte Änderung: Mon Mar 20 20:33:58 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'9_station_ruebenkamp_9.7'* ### Description for Sequence Group *'9_station_ruebenkamp'* The ego train is located at a platform. On the left side one can see a noise barrier. Depending on the partial sequence, the train is located shortly before, next to or behind the platform. In 9.1 and 9.2, the ego train travels on the middle track, in the other subsequences it is on the left track. Furthermore, in some subsequences the weather is rainy, and the number of passengers also differs. Common object classes are track, catenary pole and in some subsequences person (passenger). ### Statistics for Sequence *'9_station_ruebenkamp_9.7'* Number of Multisensor-Frames: `10` Total file size: `845 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 448 crowd 0 train 18 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 0 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 180 transition 0 switch 0 catenary pole 360 signal pole 130 signal 110 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 1 246

9_station_ruebenkamp_9.6 (Letzte Änderung: Mon Mar 20 20:33:43 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'9_station_ruebenkamp_9.6'* ### Description for Sequence Group *'9_station_ruebenkamp'* The ego train is located at a platform. On the left side one can see a noise barrier. Depending on the partial sequence, the train is located shortly before, next to or behind the platform. In 9.1 and 9.2, the ego train travels on the middle track, in the other subsequences it is on the left track. Furthermore, in some subsequences the weather is rainy, and the number of passengers also differs. Common object classes are track, catenary pole and in some subsequences person (passenger). ### Statistics for Sequence *'9_station_ruebenkamp_9.6'* Number of Multisensor-Frames: `10` Total file size: `846 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 60 crowd 0 train 110 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 0 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 170 transition 0 switch 0 catenary pole 250 signal pole 210 signal 268 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 1 068

9_station_ruebenkamp_9.5 (Letzte Änderung: Mon Mar 20 20:33:28 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'9_station_ruebenkamp_9.5'* ### Description for Sequence Group *'9_station_ruebenkamp'* The ego train is located at a platform. On the left side one can see a noise barrier. Depending on the partial sequence, the train is located shortly before, next to or behind the platform. In 9.1 and 9.2, the ego train travels on the middle track, in the other subsequences it is on the left track. Furthermore, in some subsequences the weather is rainy, and the number of passengers also differs. Common object classes are track, catenary pole and in some subsequences person (passenger). ### Statistics for Sequence *'9_station_ruebenkamp_9.5'* Number of Multisensor-Frames: `10` Total file size: `818 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 50 crowd 0 train 119 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 75 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 170 transition 0 switch 0 catenary pole 310 signal pole 116 signal 113 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 953

9_station_ruebenkamp_9.4 (Letzte Änderung: Mon Mar 20 20:33:11 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'9_station_ruebenkamp_9.4'* ### Description for Sequence Group *'9_station_ruebenkamp'* The ego train is located at a platform. On the left side one can see a noise barrier. Depending on the partial sequence, the train is located shortly before, next to or behind the platform. In 9.1 and 9.2, the ego train travels on the middle track, in the other subsequences it is on the left track. Furthermore, in some subsequences the weather is rainy, and the number of passengers also differs. Common object classes are track, catenary pole and in some subsequences person (passenger). ### Statistics for Sequence *'9_station_ruebenkamp_9.4'* Number of Multisensor-Frames: `10` Total file size: `741 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 1 054 crowd 56 train 79 wagons 0 bicycle 27 group of bicycles 0 motorcycle 0 road vehicle 89 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 189 transition 20 switch 20 catenary pole 380 signal pole 264 signal 278 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 2 456

9_station_ruebenkamp_9.3 (Letzte Änderung: Mon Mar 20 20:32:55 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'9_station_ruebenkamp_9.3'* ### Description for Sequence Group *'9_station_ruebenkamp'* The ego train is located at a platform. On the left side one can see a noise barrier. Depending on the partial sequence, the train is located shortly before, next to or behind the platform. In 9.1 and 9.2, the ego train travels on the middle track, in the other subsequences it is on the left track. Furthermore, in some subsequences the weather is rainy, and the number of passengers also differs. Common object classes are track, catenary pole and in some subsequences person (passenger). ### Statistics for Sequence *'9_station_ruebenkamp_9.3'* Number of Multisensor-Frames: `10` Total file size: `850 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 1 027 crowd 30 train 32 wagons 0 bicycle 27 group of bicycles 0 motorcycle 0 road vehicle 78 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 181 transition 0 switch 50 catenary pole 365 signal pole 281 signal 296 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 2 367

9_station_ruebenkamp_9.2 (Letzte Änderung: Mon Mar 20 20:32:37 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'9_station_ruebenkamp_9.2'* ### Description for Sequence Group *'9_station_ruebenkamp'* The ego train is located at a platform. On the left side one can see a noise barrier. Depending on the partial sequence, the train is located shortly before, next to or behind the platform. In 9.1 and 9.2, the ego train travels on the middle track, in the other subsequences it is on the left track. Furthermore, in some subsequences the weather is rainy, and the number of passengers also differs. Common object classes are track, catenary pole and in some subsequences person (passenger). ### Statistics for Sequence *'9_station_ruebenkamp_9.2'* Number of Multisensor-Frames: `10` Total file size: `776 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 71 crowd 0 train 78 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 30 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 250 transition 0 switch 0 catenary pole 505 signal pole 229 signal 174 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 1 337

9_station_ruebenkamp_9.1 (Letzte Änderung: Mon Mar 20 20:32:17 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'9_station_ruebenkamp_9.1'* ### Description for Sequence Group *'9_station_ruebenkamp'* The ego train is located at a platform. On the left side one can see a noise barrier. Depending on the partial sequence, the train is located shortly before, next to or behind the platform. In 9.1 and 9.2, the ego train travels on the middle track, in the other subsequences it is on the left track. Furthermore, in some subsequences the weather is rainy, and the number of passengers also differs. Common object classes are track, catenary pole and in some subsequences person (passenger). ### Statistics for Sequence *'9_station_ruebenkamp_9.1'* Number of Multisensor-Frames: `10` Total file size: `823 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 100 crowd 0 train 20 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 20 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 200 transition 0 switch 0 catenary pole 350 signal pole 60 signal 50 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 800

8_station_altona_8.3 (Letzte Änderung: Mon Mar 20 14:26:27 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'8_station_altona_8.3'* ### Description for Sequence Group *'8_station_altona'* The ego train is standing at a platform in an underground station. About 50 seconds after the end of the second subsequence in 8.3, an oncoming train has also arrived and obscures the view of the right platform. Common object classes are tack and person (passenger). ### Statistics for Sequence *'8_station_altona_8.3'* Number of Multisensor-Frames: `10` Total file size: `849 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 740 crowd 0 train 60 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 0 animal 260 group of animals 0 wheelchair 0 drag shoe 0 track 75 transition 0 switch 30 catenary pole 0 signal pole 0 signal 90 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 1 255

8_station_altona_8.2 (Letzte Änderung: Mon Mar 20 14:26:05 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'8_station_altona_8.2'* ### Description for Sequence Group *'8_station_altona'* The ego train is standing at a platform in an underground station. About 50 seconds after the end of the second subsequence in 8.3, an oncoming train has also arrived and obscures the view of the right platform. Common object classes are tack and person (passenger). ### Statistics for Sequence *'8_station_altona_8.2'* Number of Multisensor-Frames: `100` Total file size: `7.9 GB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 19 231 crowd 0 train 905 wagons 0 bicycle 300 group of bicycles 0 motorcycle 0 road vehicle 0 animal 1 474 group of animals 0 wheelchair 0 drag shoe 0 track 1 101 transition 0 switch 200 catenary pole 0 signal pole 0 signal 588 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 23 799

8_station_altona_8.1 (Letzte Änderung: Mon Mar 20 14:25:44 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'8_station_altona_8.1'* ### Description for Sequence Group *'8_station_altona'* The ego train is standing at a platform in an underground station. About 50 seconds after the end of the second subsequence in 8.3, an oncoming train has also arrived and obscures the view of the right platform. Common object classes are tack and person (passenger). ### Statistics for Sequence *'8_station_altona_8.1'* Number of Multisensor-Frames: `100` Total file size: `7.9 GB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 18 652 crowd 0 train 899 wagons 0 bicycle 265 group of bicycles 0 motorcycle 0 road vehicle 0 animal 1 365 group of animals 0 wheelchair 0 drag shoe 0 track 1 100 transition 0 switch 199 catenary pole 0 signal pole 0 signal 600 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 23 080

7_approach_underground_station_7.3 (Letzte Änderung: Mon Mar 20 14:25:27 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'7_approach_underground_station_7.3'* ### Description for Sequence Group *'7_approach_underground_station'* The ego train is heading for a large station. In 7.3, it has arrived there and stands in front of a multi-lane tunnel system. A time interval between the partial sequences cannot be determined. In the first two subsequences, train, track, catenary pole and signal are among the most frequent object classes. In 7.3, it is track, switch and signal. ### Statistics for Sequence *'7_approach_underground_station_7.3'* Number of Multisensor-Frames: `10` Total file size: `912 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 76 crowd 0 train 50 wagons 0 bicycle 89 group of bicycles 0 motorcycle 14 road vehicle 543 animal 90 group of animals 0 wheelchair 0 drag shoe 0 track 280 transition 270 switch 386 catenary pole 30 signal pole 196 signal 394 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 2 418

7_approach_underground_station_7.2 (Letzte Änderung: Mon Mar 20 14:25:09 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'7_approach_underground_station_7.2'* ### Description for Sequence Group *'7_approach_underground_station'* The ego train is heading for a large station. In 7.3, it has arrived there and stands in front of a multi-lane tunnel system. A time interval between the partial sequences cannot be determined. In the first two subsequences, train, track, catenary pole and signal are among the most frequent object classes. In 7.3, it is track, switch and signal. ### Statistics for Sequence *'7_approach_underground_station_7.2'* Number of Multisensor-Frames: `10` Total file size: `843 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 111 crowd 0 train 150 wagons 0 bicycle 54 group of bicycles 0 motorcycle 0 road vehicle 152 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 290 transition 40 switch 80 catenary pole 800 signal pole 581 signal 863 signal bridge 272 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 3 393

7_approach_underground_station_7.1 (Letzte Änderung: Mon Mar 20 14:24:43 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'7_approach_underground_station_7.1'* ### Description for Sequence Group *'7_approach_underground_station'* The ego train is heading for a large station. In 7.3, it has arrived there and stands in front of a multi-lane tunnel system. A time interval between the partial sequences cannot be determined. In the first two subsequences, train, track, catenary pole and signal are among the most frequent object classes. In 7.3, it is track, switch and signal. ### Statistics for Sequence *'7_approach_underground_station_7.1'* Number of Multisensor-Frames: `10` Total file size: `736 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 20 crowd 0 train 435 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 0 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 268 transition 35 switch 131 catenary pole 1 114 signal pole 171 signal 336 signal bridge 0 buffer stop 58 flame 0 smoke 0 ------------------ --------------------- total 2 568

6_station_klein_flottbek_6.2 (Letzte Änderung: Mon Mar 20 14:24:02 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'6_station_klein_flottbek_6.2'* ### Description for Sequence Group *'6_station_klein_flottbek'* In 6.1, the ego train approaches a platform with waiting passengers. In 6.2 it has already arrived there, a time interval between the two partial sequences cannot be determined. Common object classes are track, person (passenger) and signal. ### Statistics for Sequence *'6_station_klein_flottbek_6.2'* Number of Multisensor-Frames: `100` Total file size: `8.6 GB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 3 580 crowd 443 train 0 wagons 0 bicycle 688 group of bicycles 46 motorcycle 0 road vehicle 398 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 400 transition 0 switch 0 catenary pole 0 signal pole 416 signal 1 090 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 7 061

6_station_klein_flottbek_6.1 (Letzte Änderung: Mon Mar 20 14:23:35 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'6_station_klein_flottbek_6.1'* ### Description for Sequence Group *'6_station_klein_flottbek'* In 6.1, the ego train approaches a platform with waiting passengers. In 6.2 it has already arrived there, a time interval between the two partial sequences cannot be determined. Common object classes are track, person (passenger) and signal. ### Statistics for Sequence *'6_station_klein_flottbek_6.1'* Number of Multisensor-Frames: `10` Total file size: `912 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 166 crowd 80 train 0 wagons 0 bicycle 33 group of bicycles 30 motorcycle 0 road vehicle 656 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 146 transition 0 switch 30 catenary pole 0 signal pole 120 signal 163 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 1 424

5_station_bergedorf_5.2 (Letzte Änderung: Mon Mar 20 14:23:04 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'5_station_bergedorf_5.2'* ### Description for Sequence Group *'5_station_bergedorf'* The ego train is heading for a platform. There is a time gap of 57 seconds between 5.1 and 5.2, which means that the ego train has already arrived at the platform in the second subsequence. Common object classes in the first frame area are track, train as well as catenary pole. In the second subsequence, it is mainly person (passenger), track and signal. ### Statistics for Sequence *'5_station_bergedorf_5.2'* Number of Multisensor-Frames: `10` Total file size: `928 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 897 crowd 190 train 70 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 159 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 120 transition 63 switch 60 catenary pole 228 signal pole 245 signal 775 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 2 807

5_station_bergedorf_5.1 (Letzte Änderung: Mon Mar 20 14:22:49 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'5_station_bergedorf_5.1'* ### Description for Sequence Group *'5_station_bergedorf'* The ego train is heading for a platform. There is a time gap of 57 seconds between 5.1 and 5.2, which means that the ego train has already arrived at the platform in the second subsequence. Common object classes in the first frame area are track, train as well as catenary pole. In the second subsequence, it is mainly person (passenger), track and signal. ### Statistics for Sequence *'5_station_bergedorf_5.1'* Number of Multisensor-Frames: `100` Total file size: `9.1 GB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 0 crowd 453 train 2 698 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 485 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 1 353 transition 0 switch 0 catenary pole 9 616 signal pole 1 840 signal 4 008 signal bridge 0 buffer stop 1 378 flame 0 smoke 0 ------------------ --------------------- total 21 831

4_station_pedestrian_bridge_4.5 (Letzte Änderung: Mon Mar 20 14:22:36 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'4_station_pedestrian_bridge_4.5'* ### Description for Sequence Group *'4_station_pedestrian_bridge'* The ego train is standing at a platform facing the end of the platform. Common object classes are track, person (passenger, worker), signal and road vehicle. The subsequences differ only with respect to the visible people. ### Statistics for Sequence *'4_station_pedestrian_bridge_4.5'* Number of Multisensor-Frames: `76` Total file size: `6.9 GB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 2 029 crowd 0 train 0 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 1 368 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 532 transition 0 switch 0 catenary pole 0 signal pole 910 signal 2 204 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 7 043

4_station_pedestrian_bridge_4.4 (Letzte Änderung: Mon Mar 20 14:22:19 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'4_station_pedestrian_bridge_4.4'* ### Description for Sequence Group *'4_station_pedestrian_bridge'* The ego train is standing at a platform facing the end of the platform. Common object classes are track, person (passenger, worker), signal and road vehicle. The subsequences differ only with respect to the visible people. ### Statistics for Sequence *'4_station_pedestrian_bridge_4.4'* Number of Multisensor-Frames: `100` Total file size: `9.1 GB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 2 598 crowd 0 train 0 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 1 800 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 700 transition 0 switch 0 catenary pole 0 signal pole 1 198 signal 2 898 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 9 194

4_station_pedestrian_bridge_4.3 (Letzte Änderung: Tue Feb 28 09:56:37 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'4_station_pedestrian_bridge_4.3'* ### Description for Sequence Group *'4_station_pedestrian_bridge'* The ego train is standing at a platform facing the end of the platform. Common object classes are track, person (passenger, worker), signal and road vehicle. The subsequences differ only with respect to the visible people. ### Statistics for Sequence *'4_station_pedestrian_bridge_4.3'* Number of Multisensor-Frames: `100` Total file size: `9.1 GB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 3 117 crowd 0 train 0 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 1 785 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 700 transition 0 switch 0 catenary pole 0 signal pole 1 183 signal 2 897 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 9 682

4_station_pedestrian_bridge_4.2 (Letzte Änderung: Tue Feb 28 09:56:19 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'4_station_pedestrian_bridge_4.2'* ### Description for Sequence Group *'4_station_pedestrian_bridge'* The ego train is standing at a platform facing the end of the platform. Common object classes are track, person (passenger, worker), signal and road vehicle. The subsequences differ only with respect to the visible people. ### Statistics for Sequence *'4_station_pedestrian_bridge_4.2'* Number of Multisensor-Frames: `10` Total file size: `924 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 259 crowd 0 train 0 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 236 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 70 transition 0 switch 0 catenary pole 0 signal pole 90 signal 343 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 998

4_station_pedestrian_bridge_4.1 (Letzte Änderung: Tue Feb 28 09:56:10 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'4_station_pedestrian_bridge_4.1'* ### Description for Sequence Group *'4_station_pedestrian_bridge'* The ego train is standing at a platform facing the end of the platform. Common object classes are track, person (passenger, worker), signal and road vehicle. The subsequences differ only with respect to the visible people. ### Statistics for Sequence *'4_station_pedestrian_bridge_4.1'* Number of Multisensor-Frames: `10` Total file size: `924 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 270 crowd 0 train 0 wagons 0 bicycle 50 group of bicycles 0 motorcycle 0 road vehicle 110 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 62 transition 0 switch 0 catenary pole 0 signal pole 80 signal 330 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 902

3_fire_site_3.4 (Letzte Änderung: Tue Feb 28 09:56:00 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'3_fire_site_3.4'* ### Description for Sequence Group *'3_fire_site'* The ego train passes through an area with a noise barrier on the right and vegetation on the left. Further ahead one can see a fire. The ego train travels on different tracks depending on the subsequence and the distance to the fire is different. In 3.1 and 3.2 it is on the right track, between the two subsequences there is a time gap of 24 seconds. In 3.3 the train is also on the right track in close proximity to the fire. In 3.4 it is on the middle track, to the left of it another train can be seen. Common object classes are track, person (passenger, worker, other), signal and catenary pole. ### Statistics for Sequence *'3_fire_site_3.4'* Number of Multisensor-Frames: `10` Total file size: `841 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 312 crowd 0 train 77 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 0 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 200 transition 10 switch 20 catenary pole 470 signal pole 215 signal 305 signal bridge 0 buffer stop 0 flame 30 smoke 0 ------------------ --------------------- total 1 639

3_fire_site_3.3 (Letzte Änderung: Tue Feb 28 09:55:52 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'3_fire_site_3.3'* ### Description for Sequence Group *'3_fire_site'* The ego train passes through an area with a noise barrier on the right and vegetation on the left. Further ahead one can see a fire. The ego train travels on different tracks depending on the subsequence and the distance to the fire is different. In 3.1 and 3.2 it is on the right track, between the two subsequences there is a time gap of 24 seconds. In 3.3 the train is also on the right track in close proximity to the fire. In 3.4 it is on the middle track, to the left of it another train can be seen. Common object classes are track, person (passenger, worker, other), signal and catenary pole. ### Statistics for Sequence *'3_fire_site_3.3'* Number of Multisensor-Frames: `10` Total file size: `822 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 350 crowd 0 train 0 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 0 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 180 transition 30 switch 50 catenary pole 340 signal pole 160 signal 239 signal bridge 0 buffer stop 0 flame 30 smoke 0 ------------------ --------------------- total 1 379

3_fire_site_3.2 (Letzte Änderung: Tue Feb 28 09:55:45 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'3_fire_site_3.2'* ### Description for Sequence Group *'3_fire_site'* The ego train passes through an area with a noise barrier on the right and vegetation on the left. Further ahead one can see a fire. The ego train travels on different tracks depending on the subsequence and the distance to the fire is different. In 3.1 and 3.2 it is on the right track, between the two subsequences there is a time gap of 24 seconds. In 3.3 the train is also on the right track in close proximity to the fire. In 3.4 it is on the middle track, to the left of it another train can be seen. Common object classes are track, person (passenger, worker, other), signal and catenary pole. ### Statistics for Sequence *'3_fire_site_3.2'* Number of Multisensor-Frames: `10` Total file size: `832 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 252 crowd 0 train 0 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 0 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 237 transition 0 switch 80 catenary pole 220 signal pole 155 signal 174 signal bridge 0 buffer stop 0 flame 50 smoke 50 ------------------ --------------------- total 1 218

3_fire_site_3.1 (Letzte Änderung: Tue Feb 28 09:55:37 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'3_fire_site_3.1'* ### Description for Sequence Group *'3_fire_site'* The ego train passes through an area with a noise barrier on the right and vegetation on the left. Further ahead one can see a fire. The ego train travels on different tracks depending on the subsequence and the distance to the fire is different. In 3.1 and 3.2 it is on the right track, between the two subsequences there is a time gap of 24 seconds. In 3.3 the train is also on the right track in close proximity to the fire. In 3.4 it is on the middle track, to the left of it another train can be seen. Common object classes are track, person (passenger, worker, other), signal and catenary pole. ### Statistics for Sequence *'3_fire_site_3.1'* Number of Multisensor-Frames: `100` Total file size: `8.2 GB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 4 252 crowd 0 train 0 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 0 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 1 801 transition 0 switch 0 catenary pole 2 598 signal pole 1 372 signal 2 267 signal bridge 0 buffer stop 0 flame 300 smoke 138 ------------------ --------------------- total 12 728

2_station_berliner_tor_2.1 (Letzte Änderung: Tue Feb 28 09:55:27 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'2_station_berliner_tor_2.1'* ### Description for Sequence Group *'2_station_berliner_tor'* The ego train passes a platform inside a city. Common object classes are track, person (passenger, worker), signal and catenary pole. ### Statistics for Sequence *'2_station_berliner_tor_2.1'* Number of Multisensor-Frames: `10` Total file size: `926 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 1 082 crowd 50 train 0 wagons 0 bicycle 0 group of bicycles 0 motorcycle 0 road vehicle 33 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 139 transition 0 switch 0 catenary pole 308 signal pole 145 signal 222 signal bridge 0 buffer stop 0 flame 0 smoke 0 ------------------ --------------------- total 1 979

1_calibration_1.2 (Letzte Änderung: Tue Feb 28 09:55:14 GMT 2023)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'1_calibration_1.2'* ### Description for Sequence Group *'1_calibration'* The ego train runs onto a siding consisting of three tracks with buffer stops at the end. In subsequence 1.1, the ego train is on the right track. In 1.2 it runs on the left track. The distance to the buffer stops is the same. Common object classes are person (worker), road vehicle, catenary pole and track. ### Statistics for Sequence *'1_calibration_1.2'* Number of Multisensor-Frames: `100` Total file size: `9.2 GB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 1 155 crowd 0 train 0 wagons 0 bicycle 0 group of bicycles 29 motorcycle 0 road vehicle 2 867 animal 0 group of animals 0 wheelchair 0 drag shoe 0 track 1 782 transition 0 switch 1 025 catenary pole 4 041 signal pole 379 signal 2 250 signal bridge 0 buffer stop 1 598 flame 0 smoke 0 ------------------ --------------------- total 15 126

1_calibration_1.1 (Letzte Änderung: Wed Dec 21 10:54:02 GMT 2022)

  *Source: own illustration based on image / data by DB Netz AG* ## Project Info    The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) has been created in a joint research project by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de), [Digitale Schiene Deutschland / DB Netz AG](http://digitale-schiene-deutschland.de), and [FusionSystems GmbH](https://www.fusionsystems.de). The [Research report (10.48755/dzsf.230012.01)](https://doi.org/10.48755/dzsf.230012.01) and the [Labeling Guide (10.48755/dzsf.230012.05)](https://doi.org/10.48755/dzsf.230012.05) can be obtained from the [DZSF website](https://www.dzsf.bund.de/SharedDocs/Standardartikel/DZSF/Projekte/Projekt_70_Reale_Datensaetze.html). The data set consists of 45 sequences of annotated multi-sensor data (color camera, infrared camera, lidar, radar, localization, IMU). Data have been collected on different railway tracks in Hamburg, Germany. ## License Info The "Open Sensor Data for Rail 2023" (OSDaR23, [10.57806/9mv146r0](https://doi.org/10.57806/9mv146r0)) is published by the [German Centre for Rail Traffic Research at the Federal Railway Authority (DZSF)](https://www.dzsf.bund.de). Annotation data (file type `.json`) are published under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/legalcode). Sensor data (file types `.png`, `.pcd`, and `.csv`) are published under [CC BY-SA 3.0 de](https://creativecommons.org/licenses/by-sa/3.0/de/legalcode). ## Further Info The data set can be used in Python with the [RailLabel](https://github.com/DSD-DBS/raillabel) package published by DB Netz AG. The data set can be viewed, for example, with the [WebLabel Player](https://github.com/Vicomtech/weblabel) published by [Vicomtech Research Foundation](https://www.vicomtech.org/en/). *(Disclaimer: Vicomtech was not part of the research project and there are currently no further relationships between DZSF and Vicomtech.)*  ## Sequence Info for *'1_calibration_1.1'* ### Description for Sequence Group *'1_calibration'* The ego train runs onto a siding consisting of three tracks with buffer stops at the end. In subsequence 1.1, the ego train is on the right track. In 1.2 it runs on the left track. The distance to the buffer stops is the same. Common object classes are person (worker), road vehicle, catenary pole and track. ### Statistics for Sequence *'1_calibration_1.1'* Number of Multisensor-Frames: `10` Total file size: `946 MB` (compressed archive is 10-15% smaller) Statistic of annotation objects: object class number of annotations ------------------ --------------------- person 166 crowd 0 train 0 wagons 0 bicycle 20 group of bicycles 20 motorcycle 0 road vehicle 320 animal 0 group of animals 0 wheelchair 0 drag shoe 79 track 160 transition 0 switch 32 catenary pole 540 signal pole 60 signal 200 signal bridge 0 buffer stop 165 flame 0 smoke 0 ------------------ --------------------- total 1 762