Nighttime Image Enhancement Algorithm Research Based On Retinex Theory

With the extensive development and application of artificial intelligence technology represented by deep learning,the demands of computer vision processing on the quality of video images are becoming increasingly higher.Therefore,higher performance and smarter digital image processing equipment and technology are particularly important.Night image enhancement is one of the key technologies in image processing.The enhanced nighttime image allows users to get more picture content and details.Night image enhancement can be improved from both brightness and saturation of the image.However,the traditional nighttime image enhancement algorithm may suffer from over-enhancement,halo artifacts,or severe edge sharpening.These phenomena will lead to a decline in image quality,How to better enhance the night image to be more suitable for human eye observation,and the image that can be further analyzed and utilized is very important.The main work of this paper is as follows:(1)A nighttime image enhancement algorithm based on Retinex theory(BORET1)is proposed.The algorithm combines the background image and the global image,and uses the gamma transform function to obtain the enhancement coefficient,which can brighten the original image brightness.(2)The BORET2 algorithm is proposed to solve the problem that the enhancement of the BORET1 algorithm enhancement effect image in the dark area is not obvious.The main principle of the BORET2 algorithm is to use the optimized transform function to enhance the overall image first,and then enhance the dark region.The BORET2 algorithm and the other six algorithms are subjectively evaluated and objectively evaluated.In terms of subjective evaluation,the BORET2 algorithm has improved brightness compared with other algorithms,and the details are clearer.On the objective side,the evaluation results show that the BORET2 algorithm is superior to other algorithms in the no-reference evaluation.The biggest difference between the BORET2 algorithm and other algorithms is that the BORET2 algorithm only uses power and addition and subtraction,which can reduce the complexity of the BORET2 algorithm and shorten the running time.