Visibility Improvement of Hazy Images Using Manipulation of Convex Combination Coefficients of Equi-Hue Planes' Vertices in the RGB Color Space

Abstract Images captured in foggy or hazy environments can exhibit whitening and reduced contrast. Various methods have been proposed to enhance the visibility of hazy images with degraded quality. However, these conventional methods may cause degradation, such as unnatural color changes and artifacts, in the resultant image. To address this problem, we propose a haze-removal method for image visibility improvement using manipulation of convex combination coefficients of equi-hue planes' vertices, which correspond to white, black, and pure colors in the RGB color space. The proposed method begins with a pre-processing step containing contrast enhancement of the lightness component using a multi-scale image enhancement method with S-shaped functions. Then, for saturation enhancement, the convex combination coefficients of white are reduced and those of pure colors are increased in the lightness-enhanced image to achieve haze removal. Experiments were conducted using real hazy images and artificially synthesized images, and the effectiveness of the proposed method was verified by comparison with conventional methods. From the experimental results, we confirmed that the proposed method sufficiently improves the visibility of hazy images while suppressing unnatural color changes and artifacts.