Research On Image Defog Reduction Technology

As China's Internet era,more and more applications will be used in image processing technology,which,in face recognition,license plate recognition,video surveillance and other fields more and more obvious role.This requires that in practice the process to fully restore the true image for recognition,but in the actual situation,when the haze,the scattering effect of atmospheric particles in the air,will lead to the shooting out of the image blur,contrast and color visual effect will be greatly reduced.Therefore,in the bad weather,how to remove the impact of haze,improve image quality is particularly important,which is the focus of this paper discussed.This paper discusses two kinds of image defogging techniques,a kind of image enhancement,image enhancement technology mainly in image enhancement,frequency enhancement and color constancy enhancement algorithm based on Retinex,which focused on and in different situations the defogging algorithm based on Retinex image;The other is based on image restoration technology.Through the analysis of image degradation model and the statistics and experiment of large amounts of non fog images,an image fog removal technology based on dark primary color is formed.This paper makes a detailed analysis of the causes of haze and the degradation model of concrete image,and through the test the effect of image dark channel prior to fog technology based on the lack of timeliness showed the experimental results in the correction of,take a guided wave transmittance filter optimization algorithm to improve the efficiency of the algorithm based on.Finally,through the analysis and Research on the quality model of the drop of the image,the rapid dehazing algorithm Gauss filter and adaptive median filter based on not only improve the quality of image to the fog,increased compared with the dark channel prior algorithm of image imaging speed.Experiments of the algorithm are based on MATLAB 2012 operation,for the final results of the specific contrast through the contrast,the average gradient and the information entropy of the three quantitative parameters,and analyze the advantages and disadvantages of the algorithm.