Retinex Enhancement Algorithm Based On Frequency-domain To Divide Scale Under Poor Illumination

Multi-scale Retinex(MSR)image enhancement algorithm is widely used as an algorithm to effectively eliminate the influence of high dynamic light and improve image quality.However,traditional multi-scale MSR is based on red(R),green(G),blue(B).In this paper,the Retinex enhancement algorithm based on the frequency domain partitioning scale is introduced,and the experimental verification is carried out.Aiming at the problem of halo,image gray-scale and pseudo-color generated by the traditional multi-scale Retinex algorithm according to R,G and B division scale at low illumination,a multi-scale Retinex enhancement algorithm based on frequency domain partitioning is proposed.According to the characteristics of the visual system,the image processing process is transformed from the airspace to the frequency domain,and the image is divided into three scales according to the frequency.Combining with the characteristics of the image after the image is combined with the scale factor,the three scales are single scale Retinex image enhancement,and finally the fuzzy set of three scales of the image weighted normalized processing,highlighting the details of the image to enhance the overall image contrast,to solve the problem of fog,pseudo color.In this paper,we use low peak image data source,peak signal to noise ratio,mean square error and entropy as the quality evaluation index to verify the algorithm.