Stereo_Vision

Stereo vision project mainly consists of 9 steps"

1. Camera calibration and getting calibration files: https://github.com/Mayakshanesht/Kalman-and-bayesian-filters/blob/master/Calibration.ipynb
2. loading images from both the cameras
3. calculating disparity maps using stereo_block matching approach and stereo semi-global block matching approach
4. load calibration parameters again to get the focal length and baseline length between two cameras
5. calculate Depth map
6. do obstacle detection using YOLOv4(tiny model)
7. get obstacle distance by merging knowledge from depth map and object detection algorithm https://drive.google.com/file/d/1-6WVLLUcc16KsNu3AwL8zHDGl0MEy16C/view?usp=sharing
8. 3D reconstruction
9. pseudo-Lidar visualization