Vehicle-Detection-and-Tracking

Why we included this project

Vehicle detection and tracking is a common first project in computer vision, and this repo shows the whole pipeline in a compact, readable form. It pairs a lightweight SSD-MobileNet detector from the TensorFlow Object Detection API with a Kalman filter that predicts and corrects bounding-box states, then assigns detections to tracks and manages track lifecycles in just a few Python files. The code is simple and well commented, which makes it a useful teaching reference for engineers who want to see how detection and tracking connect before moving on to heavier trackers like SORT or DeepSORT. It ships with a pre-trained COCO model and targets cars, buses, and trucks, so it fits traffic monitoring, autonomous driving research, and vehicle-counting work. One caveat: it is a demonstration pipeline from 2017, so expect to adapt it to a current TensorFlow version before using it in production.