The main aim of this project is to calculate the visual odometry of the cars in the KITTI Dataset using Python and OpenCV. This notebook contains the following topics:
- Camera Coordinate system
- Projection matrix
- Disparity maps
- Depth map from Stereo Camera
- Depth from poinclouds
- SIFT (Scale Invariant Feature Transform)
- HOG (Histogram of Oriented Gradients)
- RANSAC (Random Sample Consensus)
- Visual Odometry
- OpenCV
- Visual Odometry KITTI dataset (details can be found in the VO_example.ipynb)
Run the VO_example.ipynb for demo. For detailed explanation follow visual_odometry.ipynb.
The black lines indicate the ground truth whereas the red lines indicate the path calculated by us.
This work is done by following the video and the corresponding github repository.