复杂动态环境下运动目标自动检测算法  

Method for Auto-detection of Tracking Moving Objects in Complicated Dynamic Environment

作　　者：王燕玲[1] 李广伦[2] 林晓[1] 

机构地区：[1]洛阳师范学院,洛阳471022 [2]洛阳理工学院,洛阳471013

出　　处：《系统仿真学报》2015年第4期715-722,共8页Journal of System Simulation

基　　金：国家自然科学基金项目(U1304616);河南省科技攻关计划项目(2010B520019)

摘　　要：针对复杂动态环境中运动目标难以精确检测的问题,提出了基于HSV颜色空间的运动目标检测算法。采用HSV颜色空间中背景减法解决运动目标颜色与背景相似和运动目标阴影影响导致运动目标难以准确检测问题;将背景区分为稳定背景和临时背景,提出双背景自动更新算法解决背景发生变化造成目标难以精确检测的问题;算法中引入表示背景变化程度的背景变化率,从而部分或全部更新背景。实验结果证明该方法能够准确检测运动目标。An approach was proposed to solve the problem of not accurately identifying the moving targets in the complex dynamic environment. The moving objects were detected by utilizing the method based on the HSV color space and auto-updating the background. The background subtraction method based on the HSV color space was used to solve the problem that the objects could not be accurately identified for the color similarity of objects and background and the influence of the shadow of the moving targets. The background change was divided into the temporary stability and the rapid change, and an automatically updating double background was used to deal with changing of the background. The change rate which indicated the degree of background was introduced to updating all or partial background. Experimental results show that this way accurately identifies the moving objects.

关 键 词：背景减法 背景更新 复杂动态环境 阴影去除 

分 类 号：TN948.6[电子电信—信号与信息处理]