摘要
为实现复杂视频中前景目标的分割,需要解决前景目标准确提取难题,但在光照情况下,会受到阴影影响。为解决这一难题,提出一种结合高斯混合模型的HSV颜色空间阴影检测算法。对HSV颜色空间阴影检测进行修正,消除对非运动目标区域阴影的误检,加入运动目标轮廓检测,消除运动目标边缘阴影误检,得到运动目标阴影的准确检测。实验结果表明,该算法能有效检测复杂背景下的阴影目标,为获得准确分割前景目标奠定基础。
To realize the segmentation of foreground objects in complex video,it is needed to extract the foreground objects accurately,which is affected by the shadow because of light.To solve this problem,an HSV color space shadow detection algorithm based on Gaussian mixture model was proposed.HSV color space shadow detection was corrected,false detection of non-moving target area shadow was corrected,and the contour detection which eliminated the false detection of moving object edge shadow was adopted to achieve accurate detection of the moving shadow.Experimental results show that the algorithm can effectively detect the shadow targets in complex background,which lays the foundation for accurate segmentation of foreground targets.
出处
《计算机工程与设计》
北大核心
2018年第1期255-259,共5页
Computer Engineering and Design
基金
重庆市高校优秀成果转化基金项目(KJZH14219)
关键词
阴影检测
高斯混合模型
HSV颜色空间
轮廓检测
视频分割
shadow detection
Gaussian mixture model
HSV color space
contour detection
video segmentation
