Research on Underwater Image Restoration Combining Dark Channel and Retinex Algorithm
Aiming at the problems of blurred details and color distortion in underwater images,based on the un-derwater optical imaging model,this paper proposes a method that integrates dark channel prior and Retinex algorithm.Firstly,in order to avoid the influence of bright white areas in the scene,a quadtree decomposition-based method is proposed to improve the estimation of background light,further estimate the transmittance by channel,intro-duce an adaptive tolerance compensation mechanism,adaptively correct the transmittance according to the brightness of the image area,and refine the transmittance by using an improved guide filter,eliminate the image block effect,and clarify the image by using the optimized dark channel a priori algorithm.The recovered image is obtained by using the optimized dark channel prior algorithm to clarify the image;secondly,the Retinex algorithm is used to improve the contrast of the underwater image and correct the color distortion;then the pixel-level fusion is performed based on the characteristics of the recovered image and the enhanced image by the Retinex algorithm,and the recovered underwater image is finally obtained.To quantitatively evaluate the recovery algorithm,information entropy,average gradient and UIQM quantified evaluation factors are selected.The experimental results show that the proposed algorithm outperforms the comparison algorithm in both subjective and objective evaluations,which provides a research basis for subsequent underwater target detection.
NETL
NSTL
万方数据
