Moving Target Tracking Based on Cam-shift Algorithm

The design of the video transmission system has been applied more and more widely in the daily study and life. In the meantime, implementation and optimization of the system which track moving target's algorithm are also of great significance. In order to adapt the development of daily demand, we build a Wi-Fi hardware platform combined with Raspberry Pi 2B and USB camera in this paper. The OpenCV architecture platform is configured under the Linux system, using C language combined with Cam-shift tracking algorithm to implement real-time moving target tracking. At the same time, Cam-shift algorithm which we use in this paper are both improvement and promotion of the Mean-shift algorithm. Mean-shift is a nonparametric method of the density function gradient estimation, by iterating optimization we can find the probability distribution of extreme value to locate the target. Cam-shift is a continuous adaptive computation of Mean-shift, which can be regarded as one of the best methods that used to track the moving target.