Shadow extraction algorithm based on background subtraction

A shadow extraction algorithm based on background subtraction is proposed in this paper. First, the original image is preprocessed, then background subtraction is used to separate foreground and background (shadow); Second, shadow edge information is detected by edge detection method based on mathematical morphology and canny algorithm, and it is used as priori knowledge in original image segmentation and shadow extraction; Finally, actual shadow information can be got by fake shadow elimination. The QuickBird remote sensing image of part area of Songming County is experimented in this paper, and the precision of the result reaches 92.4%, so the algorithm proposed in this paper is proved feasible.