Aiming at the problems of color distortion,uneven brightness and high noise in the en-hancement process of low-illumination images,a low-illumination image enhancement algorithm based on improved dark channel prior is proposed in this paper.The algorithm is based on the fact that the inverted low-illumination image is similar to a foggy image.For the inverted low-illumination im-age,firstly,the guided filter is used to solve the block effect caused by the image using the minimum filter to calculate the dark channel,and then the pixel is eliminated.Calculate the atmospheric light val-ue after interference with the pure white point of 255,and then introduce the refinement coefficient for adaptive correction of the transmittance to make the transmittance smoother,and finally use non-local average filtering to remove noise.Experiments show that the proposed algorithm is suitable for image brightness enhancement and clear details.When testing pictures on the Low-Light verification set,the obtained SSIM value is 20.5％higher than that of the comparison algorithm,and the PSNR value is 19.9％higher.It is superior to other algorithms compared in terms of subjective experience and objective e-valuation indicators.
关键词
低光照图像增强/暗通道先验/透射率自适应修正/非局部平均滤波
Key words
low illumination image enhancement/dark channel prior/adaptive transmission cor-rection/non-local average filtering
维普
万方数据