视频监控系统中运动目标检测的阴影去除方法  

Approach for Shadow Removal of Moving Object Detection in Surveillance System

作　　者：付萍[1] 方帅[2] 徐心和[1] 薛定宇[1] 

机构地区：[1]东北大学人工智能与机器人研究所,沈阳110004 [2]合肥工业大学计算机与信息学院,合肥230009

出　　处：《计算机工程》2007年第10期22-24,共3页Computer Engineering

基　　金：国家"863"计划基金资助项目(2001AA422270);国防预研基金资助项目

摘　　要：在分析了阴影产生的原因及视觉特性的基础上,针对固定场景的运动目标检测问题,提出了一种具有实际应用价值的阴影检测与去除方法。该方法通过背景建模和背景差分方法分割出运动目标及其可能存在的阴影,采用光照评估方法判断阴影是否存在,进而运用多梯度分析和二值图像的快速聚类算法检测出阴影,去除阴影并分割运动目标。实验验证了该方法的有效性。The method of the shadow detection and remove which has application value is proposed based on analyzing visual properties concerning shadows for visual surveillance purposes. The method proposed is essentially based on background modeling and background subtraction which aim to segment moving object and the corresponding shadow. It decides whether there are shadows in a given foreground figures. A method based on illumination assessment is developed for this purpose. Multi-gradient analysis and a fast clustering algorithm for thresholded image are employed to detect the shadow. Shadow is removed and the object is segmented. Experimental results are given to demonstrate the effectiveness and practicality of the method.

关 键 词：视频监控系统 阴影去除 运动目标检测 梯度分析 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]