基于颜色空间转换的运动目标检测与阴影去除

针对室内视频监控中运动目标检测常出现的阴影误检,提出了一种基于颜色空间转换的自适应背景建模和阴影消除算法.在RGB空间采用自适应背景差对视频图像进行前景背景分离,并将检测出的前景目标锁定在活动轮廓矩形框内进行目标跟踪,对于误检的虚假目标(即阴影),利用其亮度等信息,在HSV空间去除.经实验验证,该算法对阴影的去除有良好的效果,能准确检测出真实目标.

Moving Target Detection and Shadow Elimination Based on the Color Space Conversion

An algorithm of adaptive background modeling and shadow elimination based on color space conversion aims at the error detection of shadow in moving target detection for indoor video surveillance was proposed .The adaptive background difference was used to separate background and foreground of image and the detected target was locked in active contour (means rectangular box) for target tracking in RGB space. False target (means shadow) of error detection was removed by using its brightness and other information in HSV color space. The experiment results proved that the algorithm has a good effect for shadow elimination. Moreover, it can detect the real target accurately.