视频监控中一种完整提取运动目标的检测算法

An Accurate Moving Object Detection Algorithm for Video Surveillance

提出一种视频监控中完整、精确提取运动目标前景的检测算法。首先对彩色图像建立混合高斯模型,由背景差分法得到基本准确的前景图像;然后和对称差分法图像综合,得到完整可靠的运动目标图像;再利用亮度信息消除运动目标阴影;最后利用形态学滤波和连通区域面积检测进行后处理。实验结果表明,该算法检测的运动目标前景信息完整准确,对固定场景下的视频监控系统具有一定实用价值。

An accurate moving object detection algorithm in video surveillance is proposed in this paper. First, the rationale behind this algorithm is based on mixture of Guassian （MoG）. Through applying MoG on each pixel, a coarse foreground image can be obtained. Then combining with the image information using symmetrical differencing method, a more reliable foreground image can be gained. This approach involves procedure of eliminating the shadow of moving objects using, lightness information. The last step is post processing, including morphology filtering and area detection of connected region. The experimental results demonstrate the algorithm is quite accurate and efficient for moving object detection in fixed scenes.