基于区域的运动阴影检测方法

为了使阴影检测结果更加准确和鲁棒,提出了一种基于区域的运动阴影检测方法。该方法从阴影具有的物理特性出发,考虑了区域内所有像素的总体特征。将每帧图像进行合理的分块,并且采用基于梯度的方法对运动区域边缘的小块进行合并。对每个小块根据阴影区域和对应的背景区域之间具有较强的结构相似性和色度近似性的特点进行阴影检测。实验结果表明,阴影检测准确率超过90%,其结果明显比基于颜色不变量的方法有效。该方法能够准确而鲁棒地检测出运动阴影。

Region-based moving shadow detection approach

A region-based moving shadow detection approach was developed with improved accuracy and robustness.The analysis starts with the physical properties of the shadow,and takes into account the features of all the pixels in a region.Every image is segmented into blocks with the blocks at the edge of moving regions combined based on a gradient algorithm.Shadow detection is performed in every block based on the structure similarity and color approximations between the shadow region and the corresponding background region.The shadow detection accuracy is above 90% and is more efficient than approaches based on invariant color features.Test results show that the approach accurately and robustly detects moving shadows.