运动目标检测是计算机视觉研究的重要组成部分。为了改进传统运动目标检测算法存在空洞、阴影和虚假边缘现象，提出一种静态背景下基于五帧差分与背景边缘差分相结合的运动目标检测算法。该算法首先利用优化的Canny边缘检测算法快速简单地构建稳定的背景边缘模型，同时利用背景边缘检测差确定运动目标边缘；之后与五帧差分法检测的结果相累加；最后通过同化填充和后期相关处理提取完整、准确的运动目标区域。实验结果表明，该算法准确率高，连通性好，能满足实时性检测的要求。

Moving object detection is an important part of computer vision research. In order to eliminate the phenomenon of inanition, shadow and false edge, an novel algorithm of moving object detection based on five-frame-differencing and background edge-characteristic difference in the static background, is presented in this paper. Firstly, a background model is based on improved algorithm of Canny edge detection, the edge of motion object is gained by the background edge-characteristic difference, and then, it is combined with five-frame differencing to detect the motion object. Finally the complete and accurate object is extracted with morphologic filtering and connectivity detecting. The result shows the present algorithm runs quickly and veraciously and fits for real time detection.