3D Object Detection & Tracking - with occlusion handling - using a Stereo Camera Setup

In this video, I show how to build a 3D object detection and tracking system using a stereo camera setup, with robust handling of occlusions. The pipeline combines depth estimation, object detection, and multi-object tracking to achieve reliable 3D tracking in challenging scenarios. Full code: https://github.com/ApostolosApostolou/stereo-3d-tracker Algorithms used: - SGBM (Stereo Depth Estimation): https://se.mathworks.com/help/visionhdl/ug/stereoscopic-disparity.html - YOLO (Object Detection): https://www.datacamp.com/blog/yolo-object-detection-explained - BoT-SORT (Multi-Object Tracking): https://arxiv.org/pdf/2206.14651 - Kalman Filter (State Estimation): https://thekalmanfilter.com/kalman-filter-explained-simply/