Research on the image matching and tracking algorithm for the end of infrared target tracking

The real-time target tracking has become the key technologies of the image processing technique for the IR imaging seeker. In the paper, a new real-time tracking algorithm is proposed, which is based on corner of feature points matching and correlation tracking algorithm. In this method, a small set of feature points are extracted from IR image using a SUSAN arithmetic operator. Therefore we can select the initial template according to the corner feature points, and then we can obtain the reference template to track target in real-time. In this paper, we explain in detail the principle for selecting template and improving optimization feature point of the initial template. The experimental results show that our tracking algorithm can resolve the two problems of existing algorithm very well: (1) the sliding of tracking point; (2) the locked framework excursion. Consequently, our algorithm could satisfy the requirement of real-time and tracking stability.