基于亮度直方图匹配的运动目标检测算法  

Moving object detection algorithm based on matching of brightness histogram

作　　者：薛丽霞[1] 刘煌[1] 王佐成[2] 

机构地区：[1]重庆邮电大学计算机科学与技术学院,重庆400065 [2]重庆邮电大学软件学院,重庆400065

出　　处：《计算机工程与应用》2013年第12期148-150,共3页Computer Engineering and Applications

基　　金：重庆市教育委员会基金项目(No.KJ080521)

摘　　要：提出了一种亮度直方图匹配的运动目标检测算法,采用相邻两帧各子块的直方图匹配程度检测运动目标,通过自适应的最大阈值方差法来选取阈值,在HSV色度空间下检测阴影,用形态学方法进行后处理,得到准确的运动目标;实验结果表明该方法是快速有效的。This paper introduces a moving object detection algorithm based on the brightness histogram match. It uses the matching degree of the adjacent two frames block＇ s brightness histogram to detect the moving object, selects the threshold through the adaptive biggest threshold variance method, detects shadow in the HSV color space, and gets the exact moving object after pro- cessing by the morphological method. The experiment results show that the presented method is fast and effective.

关 键 词：直方图匹配 最大阈值方差法 HSV色度空间 形态学方法 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]