基于TMS320DM642的PPM跟踪算法设计与实现

文中设计了一个实时目标跟踪控制系统，采用PPM（后验概率指标）目标跟踪算法，辅以模板自适应算法和目标轨迹预测算法。系统可以进行360。场景的巡逻监控，并在手动捕获待跟踪的运动物体后，实时跟踪目标。利用目标偏离视场中心的偏差生成控制信号，以此控制云台的运动，确保目标始终处于摄像机的视场中央。经过优化的代码在SEEDVPM642平台上的运行结果表明，该系统能实时可靠地跟踪选定的目标。

Design and Implementation of PPM Tracking Algorithm Based on TMS320DM642

In this paper, a visual target tracking system using PPM tracking algorithm er with template updating and trace prediction is proposed and implemented on the SEED VPM642 platform. It achieves the basic operations, such as scanning and monitoring, the working range of vision is 360~. After selecting the tracking target manually, this system can carry out real-time target tracking under the panoramic vision of complex back- ground. Based on divergence between target position and center region of view field, the DSP generates proper PELCO-D signal to make target staying within the center region of view field. Results of the optimized code run- ning on the SEED VPM642 platform shows that the system can track the selected target reliably.