Shadow elimination by chrominance distortion and local kullback-leibler divergence
基于色度畸变和局部交叉熵的阴影消除

Extracting Moving objects accurately plays an important role in intelligent video-surveillance systems.Remnant shadows will bring many problems such as distortion of the object shapes and the unexpected adjacency between objects,and these problems can make the posterior work of object tracking and recognition very difficult.First,this paper analyse the optical characteristic of cast shadows,then propose a shadow elimination method based on chrominance distortion and local Kullback-Leibler divergence.At last,the experiment results are presented.