Improved Algorithm For Image Haze Removing Based On Dark Channel Prior And Retinex

Due to the refraction and scattering of atmospheric light, the quality of images captured under foggy conditions will be reduced or distortion. This phenomenon will cause serious interference on the traffic safety, video monitoring, etc. The enhancement algorithm of fog image is an important research subject in the field of modern science and technology. In the study of the algorithm many scholars made a lot of progress. The work of this thesis can be detailed as follows:Firstly, this thesis was analyzed the basic theory of the haze removing algorithm, which based on dark channel prior. However, the method still has some disadvantages. Such as the image edge details is not clear, the brightness of image looks slants dark, and the speed of the algorithm is slower.Secondly, this thesis was selected an image defogging improved algorithm based on adaptive median filtering and dark channel prior to improve these deficiencies of the traditional dark channel prior. Instead of guided filtering, this thesis selects adaptive median filtering to get the refined dark channel image. The core of this algorithm is blocking the image to processing. The simulation result reveals that the improved algorithm can effectively adjust the brightness of image and higher contrast of image.Thirdly, this thesis was analyzed the Retinex theory and algorithm, and in-depth research on Single-Scale Retinex(SSR), Multi-Scale Retinex(MSR) and Multi-Scale Retinex with Color Restoration(MSRC R). To compare the advantages and defects of these three algorithms, we experimentize on the computer.Finally, in this thesis, an improved method for dark channel prior and Multi-Scale Retinex with Color Restoration algorithm was proposed. Firstly, using dark channel prior to get the defogging image, and then go to the image from RGB color space conversion to HSI color space. Finally, using the MSRCR algorithm to adjust the brightness of the image, and stretch the image of saturation adaptively. From a lot of experimental analysis and evaluation values of the processed image, we can see that the improved algorithm has better color constancy and brightness.