基于改进高斯模型的目标检测与阴影去除方法  

Target detection and shadow removal method based on the improved Gaussian model

作　　者：王红茹[1] 童伟[1] 

机构地区：[1]江苏科技大学机械工程学院,镇江212003

出　　处：《江苏科技大学学报（自然科学版）》2017年第2期172-177,共6页Journal of Jiangsu University of Science and Technology:Natural Science Edition

基　　金：江苏科技大学高级人才基金资助项目(35020902)

摘　　要：针对复杂环境下传统目标检测方法不能够准确检测出运动目标,容易将运动阴影误检为运动目标的问题,提出一种基于改进自适应混合高斯模型与颜色空间相结合的目标检测与阴影去除方法.该方法通过三帧差分获取当前帧目标的粗略区域,使用改进混合高斯模型方法区分出包含阴影的运动区域与背景显露区域,采用不同的自适应更新策略更新建模参数;然后进一步利用基于YUV颜色空间特性去除阴影;最后通过形态学处理提取出准确的运动目标区域.对比实验表明,所提方法不仅能够有效抑制阴影和光照变化的影响,而且具有良好的实时性.To cope with low accuracy of traditional moving object detection and the problem that a moving shadow is easily mistaken for moving target,a novel moving target detection and shadow removal method is proposedbased on improved adaptive Gaussian mixture model and color space. Firstly,we extract rough region of the current frame by three frame difference,divide the suspicious motion region into the exposed background region and motion region by improved mixed Gauss algorithm,and update Gaussian parameters of the two regions by different strategies. Then shadow is suppressed by the YUV color space feature. Finally,the complete and accurate moving target area is detected out by morphological closing operation. Comparative experiments indicate that the proposed method not only can effectively restrain the influence of shadow and lighting changes,but also has good real-time performance.

关 键 词：运动目标检测 自适应 混合高斯模型 彩色空间 阴影 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]