An accurate and robust adaptive motion shadow detection algorithm

In order to improve the results of shadow detection and remedy the defect that fixed threshold detection methods are difficult to adapt to different lighting environment, and to improve the detection and tracking accuracy of moving target, we propose an accurate, robust and adaptive motion shadow detection algorithm. The algorithm takes an advantage of the inner relationship between pixels pair and pixels. Firstly, the algorithm detects the first shadow candidate region by using the local color constancy (LCC) feature, and gets the region's largest connected domain with region growing method, and estimates the c1c2c3 detection thresholds based on the parameter estimation theory. In addition, we update the detection threshold in combination with an updating weight, so that it can automatically adapt to all kinds of ambient lighting environment. With the experiments on standard test set in different lighting conditions, we get the results that the proposed algorithm can effectively improve the results of sampling parameters estimation, besides, it has good robustness and improved detection accuracy compared with other methods, and the comprehensive evaluation index has reached above 94%.