摘要
运动目标的检测与跟踪在许多领域有着广泛的应用,是应用视觉研究的焦点之一。设计了一种基于DM642的运动目标检测与实时跟踪系统。运动目标检测采用基于YUV色彩空间的单高斯背景建模结合形态学处理的算法;运动目标跟踪采用基于亮度纹理信息的CAMShift跟踪算法。系统利用跟踪算法计算目标位置信息,通过RS485总线控制摄像机伺服设备,以跟踪运动目标。在DM642 EVM板上实现了该系统,实验表明这种检测和跟踪方法是快速有效的。
Moving target detecting and tracking play important roles in many applications, and became a focus for research of computer vision. In this paper, a system for real-time motion target detection and tracking based on DM642 has been designed. Single-Gaussian background modeling algorithm, which is based on color space YUV, was combined with morphological processing to detect moving target. Then the CAMShift algorithm based on luminance texture information was used to track moving target. The information of the objective location calculated by tracking algorithm was transmitted by RS485 to drive the servo of the camera to track the object. The system was realized on the DM642 EVM board, and the experiment showed that the detection and tracking method is effective.
出处
《现代科学仪器》
2011年第2期30-32,共3页
Modern Scientific Instruments
基金
湖北省自然科学基金(No.2010CDZ057)
