Robust vehicle tracking based on Scale Invariant Feature Transform

Vehicle tracking is a challenging problem in Intelligent Transport System. This paper presents a vehicle tracking approach combining blob based tracking and feature based tracking. First objects are detected as blobs using codebook(CB) algorithm and scale invariant feature transform(SIFT) features are extracted from the blobs. Then vehicles are tracked by using SIFT to match the vehicles frame-by-frame. The method is robust to partial occlusion, partial affine distortion, changing in illumination, shape and size of vehicle. The experiments show that it is effective for vehicle tracking.