- Install pyKitti
- Download dataset (odometry 80 gb+metadata) for our project http://www.cvlibs.net/datasets/kitti/eval_odometry.php (sliced part is provided here)
- Play with https://github.com/windowsub0406/KITTI_Tutorial to understand how pyKitti works (example for odometry is here, now)
- Install pyKitti, open3d-python, scikit-learn, opencv_contrib_python and other libraries
- Trajectory projection (velodyne from left cam to 2D)
- Clustering black box
- SLAM with DA (or SAM) with map from previous step
- 3D visualization