Shadow Removal Based on Physical Features and Morphological Methods

In this paper an algorithm has been presented to remove the shadow from images, which includes two steps. First step is to detect shadow regions using physical features of shadow for an initial result with less shadow. Two properties of shadow, luminance and gray density are taken into account, to get the primary result. The second step is based on morphological methods. Using the similarity of gray density and the gradient value, a refined result can be obtained with the help of real edge of the object and the direction to expand.