Here is the source code of some exercise. They are based on KITTI datasets.
The following packets are needed:
- utm (Convert GPS coordinates in mercator coordinates)
- pykitti (Read KITTI's datasets)
- open3D (Display 3D objects and point cloud)
- opencv (Image processing)
- numpy
- matplotlib
List of exercises
- Display coordinate frames for cameras, IMU and velodyne
- Display the GNSS trajectory for the next 30 seconds in a Velodyne pointcloud.
- Display the GNSS tragectory for the next 30 seconds in the image of one of the front camera
Download rectified raw dataset and calibrations set on the RAW Data page of KITTI. First exercice can be run with reperes.py, the second with lidarGNSS.py and the last with camGNSS.py. A bonus one, camToLidar.py, display the lidar pointcloud with the color from the camera image.