An autonomous target-tracking algorithm based on visual feature

An autonomous target-tracking algorithm based on visual feature is proposed, which combines the Camshift (Continuously Adaptive Mean-Shift) algorithm with the SIFT(Scale-Invariant Feature Transform) one. SIFT algorithm is stable and it is well suited for tracking the object that contains complex color background and more SIFT points. Camshift algorithm is suited for the case of simplex color. In this paper, the advantages of the two algorithms are fused to implement a real-time system for autonomous moving object detecting and target tracking, which shows good performance in the case with any color distribution. Compared to the traditional separated Camshift and SIFT algorithm, the proposed algorithm is validated by the experiments.