A Framework in Shadow Detection and Compensation of Images

Shadow detection and removal is a crucial and unavoidable task in computer vision algorithms. In this paper detection and reconstruction processes are employed separately on an image. In the detection process, the shadow and non-shadow parts of an image are detected separately using thresholding, histogram equalization and shadow masking. The reconstruction process can be split into two steps. The first step is the calculation of mean and standard deviation for the shadow and nonshadow parts and the second step is finding the difference between the shadow and non-shadow parts and applying these differences to the shadow part. The result is a quality image with non- shadow region.