基于相关滤波器的嵌入式动态目标跟踪系统设计

针对空间非合作运动目标监控，提出了一种基于嵌入式技术的动态目标实时检测与跟踪方法，并完成了跟踪系统的研制设计和测试。该视觉目标跟踪系统以嵌入式处理器为控制器，通过相机模组获取实时视频图像，图像经过视觉检测与跟踪算法程序分析，得到目标的位置信息后，控制伺服运动系统调整相机姿态实现对动态目标跟踪。提出了基于相关滤波原理的跟踪算法，搭建了树莓派嵌入式测试系统，完成实时检测与跟踪的实验评估。实验结果表明，嵌入式处理器中检测与跟踪程序的平均运行帧率达到25FPS，伺服机构在水平360°和俯仰±60°范围内实现了对移动速度90°/s的目标的实时、稳定监控与跟踪。

Design of Embedded Moving Object Tracking System Based on Correlation Filter

To monitor a non-cooperative moving object, we propose a method of real-time detection and tracking of moving object. Based on the embedded technology, we develop and test the tracking system. The embedded processor is the controller of the designed object tracking system. The camera module is employed to obtain the real-time video images. The acquired images are analyzed by the tracking algorithm. Therefore, the system is to obtain the position information of an object, and the servo motion system is to adjust the camera to track the object. The tracking algorithm is based on correlation filtering. We built the testing system with raspberry PI module to achieve the experimental evaluation of real-time detection and tracking. Experimental results indicate that the average frame rate of the detection and tracking program is as high as 25FPS. The system can conduct real-time and stable monitoring and tracking of a moving target with the speed of 90°/s in the range of horizontal 360° and pitch ±60°.