基于小波边缘提取的混合运动目标检测算法  

A Hybrid Moving Target Detection Algorithm Based on Wavelet Edge Extraction

作　　者：李孟歆[1] 范静静[1] 张颖[1] 张锐[1] 侯丁丁[1] 许伟靖[1] 

机构地区：[1]沈阳建筑大学信息与控制工程学院,沈阳110168

出　　处：《集成技术》2013年第3期41-44,共4页Journal of Integration Technology

基　　金：辽宁省科学技术计划项目(2012229005)

摘　　要：常用的运动目标检测算法易受到噪声、阴影等因素的影响,检测出来的运动目标边缘比较模糊,本文因此提出一种基于小波边缘提取的运动目标检测算法,将小波边缘检测与帧间差分法和背景差分法相结合来检测运动目标,采用形态学滤波和连通性分析得到准确的运动目标。实验表明,该算法可以准确的将运动目标从视频图像序列中检测出来。The existing moving target detection algorithms are vulnerable to noise and shadows, so the edge of the moving target detection is relatively vague. Combining the inter-frame difference with background subtraction, a wavelet edge detection method is proposed to detect moving targets. The morphological filtering and connectivity analysis are applied to obtain accurate moving target. The experiment shows that the algorithm can be adopted to detect moving targets from the video image sequence accurately.

关 键 词：运动目标检测 小波边缘检测 帧间差分法 背景差分法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]