结合局部二元图特征的运动目标阴影抑制方法  

Shadow suppression method for moving object based on the characteristics of binary pattern

作　　者：戴璐平[1] 刘海英[1] 郑宽磊[1] 

机构地区：[1]武汉工程大学电子信息学院,湖北武汉430205

出　　处：《华中科技大学学报（自然科学版）》2016年第10期119-122,共4页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：湖北省教育厅资助项目(B2016059)

摘　　要：针对基于局部二元图纹理的阴影区分算法不能有效抑制暗阴影边缘的问题,提出了一个结合图像边缘信息的改进方案.当进行阴影抑制时,对不在边缘上的像素直接通过比较前景像素及其相应背景像素的局部二元图纹理值进行区分;否则,以待判定的像素为中心,设定一个滑窗,对滑窗内相同位置的前景像素与背景像素的局部二元图纹理值进行比较,统计滑窗内纹理值相同的像素对的数量.若该数量大于一定的阈值,则判定中心像素为背景像素,否则为目标像素.实验结果表明:改进方案能适应阴影的明暗程度变化,消除仅使用局部二元图时存在的伪阴影边缘,所增加的计算量与阴影区域的大小成比例.An improved image segmentation algorithm which combining image edge information was proposed in order to solve the problem that the shadow discrimination algorithm based on local binary pattern texture works badly on dark shadow edges suppression. In shadows suppression step, for the pixels on edges, the proposed algorithm clustered pixels into foreground and background based on their local binary pattern texture values directly; for other pixels, a sliding window with the undeter- mined pixel as its center was used, and the local binary pattern texture values of background pixel and foreground pixel at same place in the sliding window were compared. Then the number of pixel pairs with same binary pattern texture value among the sliding window was counted. If the number was greater than certain threshold, then the center pixel was classified as background pixel, otherwise it was classified as foreground pixel. The experimental results show that the improved algorithm is adaptive to shading degree change, which can eliminate false shadow edges compared to the algorithms only using local binary pattern. The computation complexity is proportional to the size of the shadow area.

关 键 词：目标分割 阴影抑制 背景建模 混合高斯模型 局部二元图 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]