Research And Application Of Non-uniform Illumination Image Enhancement Algorithm Based On Retinex Theory

Image enhancement refers to improving the brightness,contrast,target definition and other characteristics of the image in a certain way to achieve the purpose of obtaining more information.Due to the influence of the surrounding ambient light,the collected images are uneven in brightness,and the details of the target object are hidden in darkness or bright light.This uneven illumination makes people's visual experience of the image worse,and is not conducive to subsequent image analysis.Retinex theory is a classic theory in the field of image enhancement,but when it is used to deal with uneven illumination images,there is a problem of severe local color distortion.This article deeply studies the image enhancement algorithm based on Retinex theory,improves it and proposes a color retention good Image enhancement algorithm of uneven illumination and apply it to the target detection field.The main research contents of this paper are as follows:(1)The multi-scale Retinex algorithm is improved,and combined with the wavelet transform method,a multi-scale Retinex and wavelet transform-based color image enhancement algorithm(AMSR-WT)based on adaptive weights is proposed.The main steps of the algorithm include:first convert the image from RGB space to HSI space,divide the luminance I component into low-frequency images and high-frequency images,and then use the improved adaptive weighting MSR algorithm to enhance the low-frequency images,while The image is enhanced by wavelet denoising,and finally the image is converted back to RGB space through linear color restoration to obtain the final enhanced image.By comparing with other test algorithms,this algorithm can enhance the image visualization while ensuring color fidelity.(2)An SSD(Single Shot MultiBox Detector)target detection method based on AMSR-WT pre-processing enhancement is proposed.This method mainly improves the impact of dark light environment on image target detection accuracy.After that,it is detected by the SSD target detection algorithm.This method has been well verified in the testing process,reducing the missed detection of the target in the dark light environment and improving the detection accuracy of the target.(3)Using Python's powerful GUI tool library PyQt5 to realize the image visualization system of this algorithm.The system implements the two algorithms proposed in this paper and some classic image enhancement algorithms.In addition,the image evaluation module can assist users in choosing image enhancement methods.