中北大学学报（自然科学版）2025，Vol.46Issue(5)：592-600,9.DOI:10.62756/jnuc.issn.1673-3193.2024.10.0014
基于Max-min先验与强度先验的图像盲去模糊方法
Image Blind Deblurring Method Based on Max-min Prior and Intensity Prior
摘要
Abstract
Image deblurring aims to remove the blurring phenomenon caused by shaking,lighting and other reasons in the process of taking pictures,then obtain a clear image.In view of the lack of process-ing image details and edge information in deblurring with prior knowledge,we proposed a method combin-ing Max-min prior and intensity prior.The algorithm extracted the prominent edges and details of the blurred image by intensity prior,used Max-min prior to suppress blur and noise interference,used a multi-scale strategy to obtain the blurred kernel,and used the estimated intermediate restored image and the blurred kernel to iterate repeatedly to obtain a clear image.Experimental results show that compared with the existing methods,the proposed method can estimate the blurred kernel more accurately,and the inter-mediate image contains clearer and sharper edge information,and the artifacts are also improved,and the performance evaluation indexes PSNR and SSIM are also improved.The PSNR value can be increased by 0.782～3.6754 dB,and the SSIM value can be increased by 4.69%.These results show that the pro-posed method has better deblurring performance than the existing methods in the direct sensory and numerical evaluation indexes of the obtained clear images.关键词
Max-min先验/强度先验/图像去模糊/最大后验概率Key words
Max-min prior/intensity prior/image deblurring/maximum a posteriori probability分类
信息技术与安全科学引用本文复制引用
刘成琼,续婷,程蓉,白艳萍..基于Max-min先验与强度先验的图像盲去模糊方法[J].中北大学学报（自然科学版）,2025,46(5):592-600,9.基金项目
山西省基础研究计划资助项目(202103021224195,202103021224212,202103021223189,20210302123019) （202103021224195,202103021224212,202103021223189,20210302123019）
山西省回国留学人员科研项目(2021-108). （2021-108）
