CAO Shengfang, HU Hongping, WANG Wenke
（School of Mathematics, North University of China, Taiyuan 030051, China）
Abstract： Aiming at the problem of ringing artifacts existing in the edge of image in traditional blind image deblurring methods, l1/l2 regularization-based blind image deblurring method is proposed. The latent image is constrained by l1/l2 regularization, and the two-norm constraint is applied to the blur kernel to remove the noise of the blur kernel. During the solution process, the latent image and the blur kernel are updated alternately and iteratively, and the deblurred image is finally obtained by combining the finest estimated blur kernel with the non-blind deblurring method. The experimental results show that the proposed method improves the quality of image deblurring and effectively removes some ringing artifacts. It has a good restoration effect on natural blurred images.
Key words： blind image deblur； regularization； fast iterative contraction threshold； fast Fourier transform
References
［1］CUI G, ZHAO J, GAO X, et al. High quality image-pair-based deblurring method using edge mask and improved residual deconvolution. Optical Review, 2017, 24(2): 128-138.
［2］YANG T J, JIANG Y Q, GUO X Y. Blind image restoration algorithm based on tikhonov regularization. Computer Technology and Development, 2022, 32(5): 29-35.
［3］LI H B, SHAO W Z. A survey of blind image deblurring: from variation to depth models and extension. Journal of Nanjing University of Posts and Telecommunications (Journal of Natural Science), 2020, 40(5): 84-94.
［4］XU L, ZHENG S Z, JIA J. Unnatural L0 sparse representation for natural image deblurring//IEEE Conference on Computer Vision and Pattern Recognition, Jun. 23-28, 2013, Portland, Oregon, USA. New York: IEEE, 2013: 1107-1114.
［5］ZHENG B W, LI B, LI Y R. A depth image inpainting method for statistical consistency. Journal of Signal, 2020, 36(9): 1511-1524
［6］HE Y J, ZENG W Q, YU A M. An effective edge detection method for medical noisy image. Electronic Design Engineering, 2016, 24(10): 180-183.
［7］ZHU P, ZHOU Y. Blind image deblurring algorithm based on gradient l0 sparse regularization. Chinese Journal of Stereology and Image Analysis. 2015, 20(4): 361-368.
［8］PAN J S, SUND Q, PFISTER H, et al. Blind image deblurring using dark channel prior//IEEE Conference on Computer Vision and Pattern Recognition, Jun. 26-Jul. 1, 2016, Las Vegas, Nevada, USA. New York: IEEE, 2016: 1628-1636.
［9］WEN F, YING R D, LIU Y P, et al. A simple local minimal intensity prior and an improved algorithm for blind image deblurring. IEEE Transactions on Circuits and Systems for Video Technology, 2021, 31(8): 2923-2937.
［10］CHEN Q J, HU Q N, LI J Y. Image deblurring using multi-scale alternating connected residual networks. Optics and Precision Engineering, 2021, 29(7): 1686-1694.
［11］GUO K X, BAI Y F, FU X Q, et al. Ghost imaging of the low or high frequency based on the corresponding spatial frequency of the reference pattern. Optics Communications, 2019, 444: 120-126.
［12］KRISHNAN D, TAY T, FERGUS R. Blind deconvolution using a normalized sparsity measure//IEEE Conference on Computer Vision and Pattern Recognition, Jun. 20-25, 2011, Colorado Springs, CO, USA. New York: IEEE, 2011: 233-240.
［13］QIN F, FANG S, WANG L, et al. Kernel learning for blind image recovery from motion blur. Multimedia Tools Applications, 2020, 79(29-30): 21873-21887
［14］ZHANG T M. Non-convex trace lasso method for image saliency segmentation. Computer Era, 2021(11): 5-8.
［15］PAN S L, YAN K, LAN H Q, et al. A sparse spike deconvolution algorithm based on a recurrent neural network and the iterative shrinkage-thresholding algorithm. Energies, 2020, 13(12): 3074
［16］YU Y F, LI X, HE Y S, et al. Research on underwater image quality improvement method based on contrast Experiment. Application of Electronic Technology, 2021, 47(6): 82-86+91.
［17］YANG Y F, PANG Z F, SHI B L, et al. Split bregman method for the modified lot model in image denoising. Applied Mathematics and Computation. 2011, 217(12): 5392-5403.
［18］PAN J S, LIU R S, SU Z X, et al. Kernel estimation from salient structure for robust motion deblurring. Signal Processing Image Communication, 2013, 28(9): 1156-1170.
［19］PAN J S, HU Z, SU Z X, et al. Deblurring text images via l0-regularized intensity and gradient prior//IEEE Conference on Computer Vision and Pattern Recognition, Jun. 23-28, 2014, Columbus, OH, USA. New York: IEEE, 2014: 2901-2908
［20］BAI Y, CHEUNG G, LIU X, et al. Graph-based blind image deblurring from a single photograph. IEEE Transactions on Image Processing, 2018, 28(3): 1404-1418.
［21］LIU J J, LOU Y F, NI G X, et al. An image sharpening operator combined with framelet for image deblurring. Inverse Problem, 2020, 36(4): 045015.
［22］CHEN J, XIE T F, YANG Y, et al. SPADMM algorithm based on L0-L1norm. Journal of Inner Mongolia University (Natural Science Edition), 2022, 53(4): 351-358.
［23］CHEN L, FANG F, WANG T, et al. Blind image deblurring with local maximum gradient prior//Conference on Computer Vision and Pattern Recognition. Jun. 15-20, 2019, Long Beach, CA, USA. New York: IEEE, 2019: 1742-1750.
基于l1/l2正则化的图像盲去模糊方法
曹胜芳， 胡红萍， 王文科
（中北大学 数学学院， 山西 太原 030051）
摘要：针对传统图像盲去模糊方法中出现的图像边缘振铃伪影问题， 本文提出了一种基于l1/l2正则化的图像盲去模糊方法。 该方法利用清晰图像边缘信息的稀疏性， 采用l1/l2范数对未知图像潜影进行约束， 对模糊核采用l2范数约束以去除模糊核的噪声。 在求解过程中， 通过交替迭代的方式更新潜影和模糊核， 以估计出的最精细模糊核结合非盲去模糊方法， 从而得到去模糊图像。 实验结果表明， 提出的方法对图像去模糊质量有所提升， 有效地去除了一些振铃伪影， 对于自然的模糊图像也有较好的复原效果。
关键词：图像盲去模糊； 正则化； 快速迭代收缩阈值； 快速傅里叶变换
引用格式：CAO Shengfang, HU Hongping, WANG Wenke． Blind image deblurring method based on l1/l2-norm regularization． Journal of Measurement Science and Instrumentation， 2023， 14（2）： 182-188. DOI： 10．3969／j．issn．1674-8042．2023．02．007
[full text view]
