Low-Cost-Mocap

Why we included this project

Getting accurate indoor position data for autonomous drones usually means paying for a commercial motion capture rig. This project swaps that for a few inexpensive cameras wired into a room-scale tracking system, combining structure from motion, epipolar geometry, and bundle adjustment to compute 3D marker positions that multiple drones share for autonomous flight. A Python backend ingests the camera streams and a web frontend shows the tracking live, so it doubles as a readable reference for anyone putting vision-based tracking into a robot. Expect real DIY work: you compile OpenCV with the SFM module, calibrate your own camera parameters, and the docs are admittedly thin, but the architecture is clear enough to adapt to your own cameras and space.