Shadow Detection Based on Mixed Feature and HSI Color Space

In order to effectively detect the shadow area in video image sequences, this paper proposes a new shadow detection algorithm based on mixed feature and HSI color space. After the foreground figures are extracted, this algorithm proposes to use extended moment invariants to get global feature of image and use Gabor wavelet transform to get local feature. Then, combining the global feature and the local feature will fuse into a new feature to be processed. Finally, there is a further analysis of HSI color space to search real shadow area. Experimental results shows that the proposed algorithm owns high performance of shadow detection.