Abstract: An algorithm based on photometric properties and multi-gradient analysis was proposed to remove the moving shadows. After detecting moving objects accurately, firstly, photometric properties and connected components analysis were used to detect potential shadow regions; secondly, multi-gradient in potential shadow regions was compared with that in the background, and connected components analysis was adopted in order to remove the shadow; lastly, edge removal was imposed to further remove penumbra, and morphological algorithm was used to improve the performance. Compared to the conventional method, the experimental results validate that the algorithm has a high detection rate and a high recognition rate.

Key words: background subtraction, photometric property, gradient analysis, connected components analysis, shadow removal

摘要： 提出了一种利用光度特性和多梯度分析去除运动阴影的算法。该算法是在准确地检测出运动目标的基础上，首先利用像素邻域内光度特性和连接成分分析法检测出候选阴影区域;其次比较分析候选阴影区域和背景中相应区域的梯度，并结合连接成分分析法去除阴影;最后使用边缘去除法进一步去除半影，并应用形态学运算完善结果。实验结果表明，同传统的阴影去除法相比，该算法有高检测率和高识别率。

关键词: 背景差分, 光度特性, 梯度分析, 连接成分分析, 阴影去除