Robust image dehazing using a duided filter

摘要

Images of outdoor scenes are always degraded because of the existence of the haze. In this paper, we propose a new robust way to boost the effect and reliability of image dehazing by a guided filter. We cleverly apply the maximum filtering to improve the dark channel map and use the edge-preserving smoothing operator to improve the guided image of the guided filter. Our method can boost dehazing effects along edges with large depth jumps and enhance the robustness and effect of haze removal. The experiments demonstrate that our method can obtain a haze-free image with high visual vividness even in highly dense haze regions.