Research On Low Illumination Image Enhancement Algorithm

In the era of Internet artificial intelligence,massive image data exists.Under the influence of various environmental factors,there are a lot of low illuminance in the image data,which affects the presentation effect of machine vision to a certain extent.Therefore,it is of great significance to develop low illumination image enhancement methods.(1)Aiming at the halo problem of traditional Retinex algorithm when processing low-illumination images,a low-illumination image enhancement algorithm based on TopHat weighted guided filtering was proposed.Firstly,the input image is converted from RGB to HSV space,which can effectively avoid the color distortion in RGB space.Secondly,the method of weighted filtering regularization parameter selection was improved,using TopHat transform to extract bright area,a low illumination image obtained by convolution operation then the weighting factor of local average after processing,and use it as a weighted guide regularization parameter of the filter,in the realization of accurate estimate of light weight and at the same time to achieve light weight of edge enhancement,Reduce the image blur phenomenon.Finally,adaptive Gamma correction is used to adjust the illumination component and saturation component to improve the details and color components of the image light source,and the Retinex algorithm is used to calculate the reflection component,and finally obtain the enhanced low-illumination image.The experimental results show that the brightness and clarity of the image are greatly improved after the algorithm is processed.Compared with the original image,the information entropy is improved by 48.84%,the average gradient is improved by 295.12%,and the NRSS is improved by 249.36%.(2)Aiming at the similarity between low-illumination inversion image and foggy image,this paper proposes a new method to enhance low-illumination image by using defogging model.In this method,the atmospheric light value is calculated by normalizing the low-illumination inversion image,and the mathematical relation between Retinex theory and atmospheric scattering model is derived.Then to simplify the solution of the transmittance of inverted image saturation solution,then through simple saturation stretching function can estimate the saturation of the image enhancement algorithm using the inverted image of bright channels used in the saturation stretching function,realized the saturation of the adaptive stretching,avoid the image exposure of the problem.Experiments show that the algorithm can improve the brightness and contrast of the image,and retain the details of the image,improve the contrast of the image,and reduce the complexity of the calculation.Figure [48] Table [7] Reference [80]...