一种改进的三帧差分运动目标检测  

An improved moving object detection based on three-frame difference

作　　者：刘红[1] 周晓美[1] 张震[1] 

机构地区：[1]安徽大学计算智能与信号处理教育部重点实验室,安徽合肥230039

出　　处：《安徽大学学报（自然科学版）》2014年第6期55-59,共5页Journal of Anhui University(Natural Science Edition)

基　　金：安徽大学博士科研启动基金资助项目(33190049)

摘　　要：针对当前常用运动目标检测方法易受到光照和噪声影响、不易提取完整运动目标,提出一种改进的三帧差分运动目标检测算法.首先取连续三帧图像,对前两帧进行滤波及边缘提取,将提取结果进行异或运算,保留前两帧之间不同的边缘信息,然后对后两帧图像进行差分二值化,并进行膨胀处理,粗略得到运动目标,最后将该结果与异或结果进行与运算即可得到较为完整的运动目标边缘图.实验表明,该算法简单易行、可实时检测、能有效提取出运动目标边缘,改善了差分法常出现的运动目标信息丢失问题.Generally,changes in illumination and noise could influence the integrity and effectiveness of moving object detection. An improved method based on three-frame difference was proposed.Firstly,three frames were extracted continuously;then the former two frames were then filtered and extracted edge respectively.After that,XOR processing was conducted to get the edge difference.Secondly,the later two frames were processed by difference and binarization.Then morphological processing was used,we could obtain a rough moving object.Finally,and-operation was used to it and the result of XOR,then we could get the integral edge of moving object.Experiments showed that comparing to the three-frame difference method,this method was realized easily, which could achieve detection result timely and solve the problem of the missing information of moving targets.

关 键 词：运动目标检测 边缘提取 帧间差 异或运算 形态学处理 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]