基于DSP的实时自动跟踪系统的研究

A DSP based real-time vision tracking system

在差分得到的二值图像基础上,采用快速搜索算法预测运动物体在帧图内的二维坐标,并自动跟踪运动目标。选择TMS320C642作为视觉跟踪系统的硬件平台,采用摄像机获取场景图像,通过软件捕捉场景图像中的运动物体,识别物体运动的方向和距离,实现了视觉实时跟踪功能。在此基础上,完成了分立器件构成的嵌入式视觉实时跟踪系统的设计,对于视觉技术的实际应用具有一定工程意义。

Having obtained a binary image after difference,the authors confirm the coordinates of the moving object by the fast searching algorithm and track the moving objects automatically.TMS320 DSP is chosen as the system hardware platform,and the scene images are acquired by cameras in the system.Having caught the moving object in scene by software,and identified the direction and distance of the moving object,the authors have realized the real-time vision tracking function.In the whole paper,an embedded real-time vision tracking system is constructed by separate electronic parts.This paper has important significance in the application of vision technology.