Haze removal without transmission map refinement based on dual dark channels

Abstract

Single image haze removal has been a challenge in the field of image processing. In [1], a haze removal scheme based on dark channel prior (DCP) is presented and is getting popular because of its satisfactory performance for most of cases. However, it is known that the scheme proposed in [1] suffers from two problems: high computational cost and over-exposure when a bright area shown in images. This paper proposes a dehazing algorithm with dual dark channels where the soft matting to refine the transmission map is avoided and the atmospheric light is estimated directly from a dark channel. The objective of proposed dehazing algorithm (PDA) is to alleviate or remove the two problems found in [1]. Several examples are given to verify the PDA and to compare it with the DCP scheme The simulation results show that the PDA is about 2.5 times, on average, faster than the DCP since the soft matting is avoided. Similar visual quality to the DCP is found in the PDA which is of better color situation and without over-exposure problem as in the DCP scheme.