一种改进的基于区域匹配的图像阴影去除算法

An Improved Image Shadow Removal Algorithm Based on Region Matching

阴影使图像出现降质现象,影响图像的定量分析。针对阴影去除时丢失纹理信息的问题,本文提出了一种基于区域匹配的阴影去除方法。首先估计出阴影区域的中心位置,使用区域生长和边缘膨胀算法得到阴影区域;其次,将阴影中的子区域与亮度区域中的子区域进行匹配;最后,对配对的阴影部分的子区域采用光照转移技术去除阴影。实验表明,该方法能有效去除阴影区域并保留内部的纹理特征。

Shadow degrades the image,and affects the quantitative analysis of the image.Aiming at the problem of losing texture information during shadow removal,this paper proposes a shadow removal method based on region matching.Firstly,the center position of the shadow area is estimated,and the area growth and edge expansion algorithms are used to obtain the shadow area.Secondly,the subregion in the shadow matches with the subregion in the brightness area.Finally,a light shift technique is used to remove shadows from subregion of the paired shadow portion.Experiments show that this method can effectively remove shadow areas and retain internal texture features.