Variable Filter Retinex Algorithm for Foggy Image Enhancement

A new Retinex algorithm based on variable filters is proposed to improve the existing center-surround Retinex algorithm.The proposed algorithm can enhance details effectively while maintaining the color fidelity in the applications of foggy images enhancement.The existing center-surround Retinex algorithm cannot achieve such function because of the utilizing of fixed filter.First the thresholds are worked out according to the image quality distributing.Then in each sub-block,the algorithm uses the local information and the thresholds to make a corresponding filter to estimate illumination of the current sub-block,and adopt the approach of partially overlapped sub-block to get the whole image's illumination.At last the whole illumination is subtracted from original image.The results show that the proposed algorithm is better than the existing center surround Retinex algorithm in enhancing details and preserving color fidelity.