Improved dark channel prior image dehazing algorithm
Download as PDF
DOI: 10.23977/meimie.2019.43031
Author(s)
Peng Gao, Lixia Du
Corresponding Author
Peng Gao
ABSTRACT
Aiming at the influence of haze on image such as low contrast, low saturation and color distortion, an improved dark channel prior image defogging method is proposed. Starting from atmospheric scattering model, this method optimizes the dark channel priori rule and obtains the dark channel priori values of two different color spaces. Secondly, the atmospheric light values are estimated for two different color spaces respectively. The average transmittance of RGB color space and Y-channel transmittance of YCbCr color space are calculated. Different weights are given to the two transmittances to restore the image. This method evaluates the experimental results based on structural similarity index, brightness and information entropy. The results show that compared with the algorithm and Fattal algorithm, the defogging image obtained by this algorithm has better visual effect; compared with Zhu algorithm, the calculation is simpler and the color restoration is more accurate.
KEYWORDS
Ark channel prior, transmission map, image defogging, color space, image reconstruction, atmospheric light value