Shadow Removal Combining Invariant Color Features and Gaussian Mixture Shadow Model with Edge Detection

摘要

Shadow removal is an important step in the moving object detection. By combining Gaussian Mixture Shadow Model (GMSM) and edge detection technique, a new shadow removal method based on invariant color features is proposed in this paper. First, the edge difference image of current background and current frame is binarized by an adaptive threshold. This image is called BEDI in this paper. Secondly, based on invariant color features and GMSM, those shadows in the result of moving object detection are removed. Then, the binary edge image of the shadow removal image is obtained, called BESRI. After that, a binary edge image can be obtained by using "And logic operation" to BEDI and BESRI. Finally, the object contour is extracted using a shrinkable active contour model to get the moving object without shadows. Experimental results demonstrate that this algorithm is fast, in computation and effective in shadow removal.