Abstract: Through the analysis of video image in the HSV difference image and its difference histogram，based on pixels distribution characteristic of the object and the shadow，a new adaptive shadow detection algorithm based on Otsu method is proposed in this paper.And the experimental results show that this method can detect motion object shadows efficiently，accurately and is usable.

Key words: shadow detection, HSV color space, maximal variance between clusters

摘要： 在分析视频图像HSV差分图和差分图直方图的基础上，根据差分直方图中目标与阴影的像素分布特点，提出了一种基于最大类间方差的自适应加权阴影检测算法。实验表明，该算法具有检测精度高、效果好、适应性强和易于工程实现的特点。

关键词: 阴影检测, HSV颜色空间, 最大类间方差（Otsu）