基于MeanShift算法和改进PID算法的图像追踪研究

针对MeanShift算法可以实现图像聚类的这一特性,并结合改进的PID算法,对某二维追踪实验台实现了有效的跟踪控制.由MeanShift算法得到的坐标偏差经过自适应积分分离的PID控制器校正后,作为实验台电机的输入速度信号从而实现有效追踪.实验证明该系统可以实现左右和俯仰两个方向及其复合方向的追踪,具有较高的跟踪实时性.

Image Tracking Study Based on the meanshift Algorithm and the Improved PID Algorithm

According to the feature of MeanShift algorithm realizing image clustering,a two-dimensional tracking testing tower is controlled to realize effective tracking integrated with the improved PID algorithm. The deviations of the coordinate through the MeanShift algorithm are adjusted by self-adaption PID controller, then are input to the motor of the testing tower as the rate signal to realize effective tracking of the image. Experiments proves that the system can realize tracking of left-right and pitch as well as the direction of them and has high real-time tracking.