An Image Dehazing Method Based on Atmospheric Veil

摘要

At present, most of the image dehazing treatments have residual haze at depth of field and obvious over-enhancement phenomenon in bright area. To solve these problems, we propose a new method of fast image dehazing based on the atmospheric veil. Firstly, we use the total variation model to estimate the atmosphere veil, which can preserve the edge characteristics of image on the premise of improving the computational efficiency. Then, aiming at the failure of bright area of the atmosphere veil, we propose a fault-tolerant mechanism based on haze density feature to improve the universality of the algorithm. The experimental results show that our method has better visual effect than some existing mainstream dehazing techniques, and its efficiency also has obvious advantages.