Single Image Haze Removal Using Dark Channel Prior and Adaptive Transmission Rate
Download as PDF
DOI: 10.23977/csic.2018.0921
Author(s)
Lu Guo, Jing Song, Xinrui Li, He Huang, Jingjing Du, And Guangfeng Sheng
Corresponding Author
Jing Song
ABSTRACT
In order to solve the problems of low clarity and brightness, color distortion of the restored image obtained by the traditional haze removal algorithm, this paper proposes a new single image haze removal algorithm using dark channel prior and adaptive transmittance rate. Firstly, the image is partitioned into several local regions to calculate the dark channel and minimum map of haze image. Secondly, the atmospheric light estimation is calculated according to the dark channel map, and the atmospheric dissipation value of each pixel is calculated by the atmospheric dissipation function. Then, combined with the atmospheric dissipation function, the median guide filtering algorithm is used to distinguish the image into the close range and a remote area, and the atmospheric dissipation function value of the close range is corrected to get the accurate estimation of the transmittance. Finally, the restoration image is obtained based on the atmospheric scattering model. The experimental results show that compared with the traditional algorithm, the fuzzy coefficient of the image is reduced by 23.09%, the structure similarity and the tone reduction degree are increased by 14.63% and 12.29% respectively, and the quality of the restored image can be significantly improved.
KEYWORDS
Haze Removal, Dark Channel, Image Processing, Computer Simulation, Algorithm