Title:Nighttime Haze Removal with Illumination Correction

Abstract:Haze removal is important for computational photography and computer vision applications. However, most of the existing methods for dehazing are designed for daytime images, and cannot always work well in the nighttime. Different from the imaging conditions in the daytime, images captured in nighttime haze condition may suffer from non-uniform illumination due to artificial light sources, which exhibit low brightness/contrast and color distortion. In this paper, we present a new nighttime hazy imaging model that takes into account both the non-uniform illumination from artificial light sources and the scattering and attenuation effects of haze. Accordingly, we propose an efficient dehazing algorithm for nighttime hazy images. The proposed algorithm includes three sequential steps. i) It enhances the overall brightness by performing a gamma correction step after estimating the illumination from the original image. ii) Then it achieves a color-balance result by performing a color correction step after estimating the color characteristics of the incident light. iii) Finally, it remove the haze effect by applying the dark channel prior and estimating the point-wise environmental light based on the previous illumination-balance result. Experimental results show that the proposed algorithm can achieve illumination-balance and haze-free results with good color rendition ability.