基于局部二元图的视频对象阴影检测方法

针对在视频对象分割时,运动投影常被误分为视频对象,给出一种新的视频阴影检测方法,该算法基于在灰度图像中阴影区域和背景相应位置具有相同纹理这一事实,其中利用自适应高斯混合模型进行背景建模,利用局部二元图(local binary patterns,LBP)来表征纹理。首先,进行基于自适应高斯混合模型的背景提取,获得包含运动投影的前景分割,分割时加入了LBP纹理相似性判断,减少了分割出的目标内的孔洞提高了分割的精确度。然后利用阴影区域和已获取背景相应位置的LBP纹理相似性,可较好的对视频阴影进行检测。通过实验,获得了不错的阴影检测实验结果,可较好地应用于运动目标检测分割及跟踪等领域。

Shadow detection method in video sequences based on LBP

To prevent moving cast shadows from being misunderstood as part of moving objects in video object segmentation,a novel approach to the shadow detection based on LBP(local binary patterns)is proposed.This novel method is based on the observation that shadow regions present same textural characteristics in each frame of the gray-level video-sequence and in the corresponding adaptive background model.The background is modeled using adaptive gaussian mixture models,and texture of little patches is represented by local binary patterns.First,video object segmentation is achieved after background subtraction based on adaptive gaussian mixture models,judgment of the similarity of the texture represented by LBP is incorporated in this step,it improves the correctnesss of the video object segmentation.Then,shadow is detected based on the similarity of texture represented by LBP between shadow region and corresponding region in the background.The scheme has been tested and better experimental results have been given.It can be applied to moving object segmentation,detection and etc.