光照突变环境下运动目标检测方法研究  

RESEARCH ON MOVING TARGET DETECTION IN AN ENVIRONMENT OF SUDDEN LIGHT CHANGING

作　　者：罗欢[1] 吴怀宇[1] 陈洋[2] 周致富[1] 

机构地区：[1]武汉科技大学信息科学与工程学院,湖北武汉430081 [2]冶金工业过程系统科学湖北省重点实验室(武汉科技大学),湖北武汉430081

出　　处：《计算机应用与软件》2015年第9期201-205,共5页Computer Applications and Software

基　　金：国家自然科学基金项目(61075087;61203331)

摘　　要：针对在传统目标检测中,对于光照突变环境处理能力较低的问题,提出一种结合混合高斯背景建模和三帧差分法的方法。该方法在混合高斯背景建模的基础上,采用统计法,判断是否有光照变化,对待光照突变环境中的目标检测,采用三帧差分法。同时,将检测区域分为静态区域与动态区域,对静态区域实行隔帧检测,减少程序计算量。实验结果表明,该方法能够很好地对光照突变环境下的运动目标进行检测。We proposed an algorithm combining the Gaussian background mixture modelling and three-frame difference to tackle the problem of traditional object detection for its low processing capacity in an environment of sudden light change. Based on Gaussian background mixture modelling, this method determines whether there is the change in light using statistical method, and uses three-frame difference method to detect the targets in an environment of sudden light change, meanwhile it divides the detecting region into two areas, the static and the dynamic. For static area it detects every discontinuous frame so as to reduce computational load. Experimental results show that the proposed algorithm can well detect the moving objects in the environment of sudden light change.

关 键 词：目标检测 高斯混合建模 三帧差分 光照突变 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]