Make should you have installed the JRDB at https://jrdb.erc.monash.edu/ and pasted to data folder with the following structure:
├── data
│ ├── 2011_09_26
│ │ └── 2011_09_26_drive_0001_sync
│ └── train_dataset_with_activity
│ ├── calibration
│ │ └── indi2stitch_mappings
│ ├── detections
│ │ ├── detections_2d
│ │ ├── detections_2d_stitched
│ │ └── detections_3d
│ ├── images
│ ├── labels
│ └── pointclouds
│ ├── lower_velodyne
│ └── upper_velodyne
└── src
Then clone this repo and https://github.com/KoaBou/JRDB_publishers.git to src folder and run
colcon build
to build the model.
If you can't install ros-foxy-vision-msgs then clone this repo to your src https://github.com/ros-perception/vision_msgs
Make sure you source the workspace with
source install/setup.bash
Finally, launch the package with
ros2 launch pedestrian_detection coffe_detection.launch.py