A lightweight monocular visual odom system written in Python. This project matches feature points across sequential frames, estimates camera poses, triangulates 3D points, and tracks the 2D trajectory of the camera.
- Detects and matches ORB keypoints across frames using brute-force matching with Lowe's ratio test.
- Recovers camera motion (Rotation and Translation) by estimating the Essential Matrix with RANSAC.
- Triangulates 3D coordinate points from matching 2D keypoints.