基于双阈值的运动目标检测  

The detection of moving object based on double-threshold

作　　者：李波[1] 卢杨[1] 魏冬梅[1] 

机构地区：[1]四川大学计算机学院图形图像研究所,成都610064

出　　处：《微计算机信息》2008年第27期283-284,282,共3页Control & Automation

基　　金：国家"863"计划基金项目(2006AA12A104)

摘　　要：目前在运动目标实时检测中,主要是运用差分法进行检测。但背景差分受光照、环境影响较大,需要实时更新背景,而帧间差分容易出现空洞和误检。结合背景差分和帧间差分,采用双阈值对运动目标进行分割,能对背景进行实时更新,有效的避免了空洞和误检,并且在机场的运动目标检测中取得了较好的效果。At the present time the general method of the detection of moving object is based on subtraction.Since the background subtraction is great affected by the illumination and environment the background needs be updated.The cavity cavitation and false detectione asily appear in inter-frame subtraction.So we update the background by combining the background subtraction with the interframe subtraction and segment the moving object by 2 threshold.We acquired the good effect in practicality.

关 键 词：运动检测 图像分割 背景更新 背景差分 帧间差分 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]