Improve Transmission by Designing Filters for Image Dehazing

Abstract

Image haze removal techniques have a promising application prospects. However, color distortion by excessive dehazing is a common problem especially for images containing light color surface objects. To solve this problem, an improved image dehazing method based on dark channel prior with color fidelity is proposed in this paper. First, filters (Sobel operator and mean filter) are used to refine the initial transmission map estimated from the dark channel prior. Second, a piecewise constrained function is proposed to keep color fidelity. Experimental results on real world images show that the approach is effective in haze removal and preventing the color distortion from excessive dehazing. Preferable results of recovered images with bright and realistic colors and enhanced details are achieved.