基于YUV颜色空间的视频运动检测  

Motion detection in video based on YUV color space

作　　者：马国峰[1] 杨俊红[1] 周兵[2] 

机构地区：[1]郑州铁路职业技术学院信息工程系 [2]郑州大学信息工程学院,河南郑州450052

出　　处：《计算机工程与设计》2008年第14期3700-3702,3708,共4页Computer Engineering and Design

基　　金：国家自然科学基金项目(60075006)

摘　　要：给出了一种结合YUV颜色空间色度和亮度进行运动检测的算法,该算法首先采用单高斯背景建模,然后利用当前帧和背景帧像素的色度分量差分进行运动检测,并将膨胀后的色度检测结果和亮度检测结果进行与运算,得到色度和亮度联合检测结果。最后利用数学形态学闭运算和连通区域面积阈值化的方法对检测结果进行后处理,实现运动物体内部空洞的填充和周围较大噪声点的消除。实验结果表明,该算法能克服亮度变化和阴影的影响,取得较好的检测结果。An algorithm for motion detection in video based on chroma and intensity of YUV color space is presented, which firstly used single Gaussian background modeling, and then detected moving object via the chroma difference of the current and background image, by doing AND operation with the dilation of chroma and the binary image of intensity, the combined detection result of chroma and intensity is obtained. Finally, using mathematical morphology closing operation and connected region area threshold to fill the holes in the moving object and eliminate the greater noises. Experiments show that the algorithm can overcome illumination change and the impact of the shadow and achieve better results.

关 键 词：颜色空间 视频 运动检测 背景模型 色度 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]