Shadow Detection of Urban Color Aerial Images Based on Partial Differential Equations

Shadow detection plays an important role in digital aerial image processing. In this paper, shadow is treated as a special kind of image degradation. Shadow region commonly has lower brightness than non-shadow region. And, shadow region tends to have a slow gradient change in luminance value, or means to have some conformability. On the basis of these analyses, an edge detector is proposed firstly. Then, inspired by the general principle of the partial differential equations used in image restoration, a new shadow detection algorithm based on the PDES is presented, which uses the gradient values to be the parameter of edge detector. After the experiments with several urban color aerial images, it shows that the presented algorithm is effective for shadow detection, and no additional information is required except for the image itself.