基于非局部自相似性的低对比度图像降噪算法

doi:10.13229/j.cnki.jdxbgxb.20240594

吉林大学学报(工学版) ›› 2025, Vol. 55 ›› Issue (7): 2378-2382.doi: 10.13229/j.cnki.jdxbgxb.20240594

• 计算机科学与技术 • 上一篇

基于非局部自相似性的低对比度图像降噪算法

刘迪仁(),马澳

长江大学地球物理与石油资源学院，武汉 430100

收稿日期:2024-05-29 出版日期:2025-07-01 发布日期:2025-09-12
作者简介:刘迪仁（1965-），男，教授，博士.研究方向：地球物理测井及光纤传感技术.E-mail： liudr@yangtzeu.edu.cn
基金资助:
国家重点研发计划项目子课题(2018YFC060330502)

Low contrast image denoising algorithm based on non local self similarity

Di-ren LIU(),Ao MA

College of Geophysics and Petroleum Resources，Yangtze University，Wuhan 430100，China

Received:2024-05-29 Online:2025-07-01 Published:2025-09-12

摘要/Abstract

摘要：

为改善低对比度图像边缘平滑性并提高其清晰度，本文设计了一种基于非局部自相似性的低对比度图像降噪算法。该算法基于低对比度图像中图像块的差异度构建自相似数据集，在此基础上计算图像块相似度，并结合二值标识矩阵区分噪声点与有效像素点，引入加权系数对像素点实施加权平均运算，得到过渡图像。由此，通过计算原始图像和过渡图像的绝对差分和判断降噪是否完成，若绝对差分和≤0，则实现降噪。实验结果表明，本文算法的结构相似度稳定在0.9左右，且去噪后的图像更加清晰真实，说明本文算法可以更好地保留原始信息和结构信息，使得降噪图像在视觉上更加自然与真实。

关键词: 非局部自相似性, 低对比度图像, 图像降噪, 过渡图像, 椒盐噪声

Abstract:

In order to improve the edges smoothness and clarity of low contrast image， a low contrast image denoising algorithm based on non local self similarity is designed in this paper. Based on the difference between image blocks in low contrast images， a self similarity dataset is constructed. On this basis， the similarity of image blocks is calculated， and a binary identification matrix is used to distinguish noise points and effective pixel， weight coefficients are introduced to perform weighted averaging on the pixels， resulting in a transition image. Therefore， based on the absolute difference sum between the original image and the transition image is calculated to determined whether has completed denoising. If the absolute difference sum is less than or equal to 0， denoising is achieved. The experimental results show that the SSIM of the proposed algorithm has remained around 0.9， and the denoised image is clearer and more realistic. It is shown that the proposed algorithm can better preserve the original information and structural information， making the denoised images more natural and realistic visually.

Key words: non local self similarity, low contrast images, image denoising, transition image, salt and pepper noise

中图分类号:

TP391

刘迪仁,马澳. 基于非局部自相似性的低对比度图像降噪算法[J]. 吉林大学学报(工学版), 2025, 55(7): 2378-2382.

Di-ren LIU,Ao MA. Low contrast image denoising algorithm based on non local self similarity[J]. Journal of Jilin University(Engineering and Technology Edition), 2025, 55(7): 2378-2382.

图/表 5

图1

表1

图2

图3

图4

参考文献 16

[1]	蒋澎涛, 欧阳建权. 改进稀疏去噪算法下图像自适应融合研究[J]. 计算机仿真, 2023, 40(5): 224-227, 266.
	Jiang Peng-tao, Ouyang Jian-quan. Research on adaptive image fusion based on improved sparse denoising algorithm[J].Computer Simulation, 2023, 40(5):224-227, 266.
[2]	王蕾, 王永波, 边兆英, 等. 基于非局部能谱相似特征的基物质分解方法用于双能CT图像去噪[J]. 南方医科大学学报, 2022, 42(5): 724-732.
	Wang Lei, Wang Yong-bo, Bian Zhao-ying,et al.A nonlocal spectral similarity-induced material decomposition method for noise reduction of dual-energy CT images[J].Journal of Southern Medical University,2022, 42(5): 724-732.
[3]	陈晓军, 李芬, 徐少平, 等. 基于无监督深度图像生成的盲降噪模型[J]. 计算机应用研究, 2022, 39(7): 2224-2229.
	Chen Xiao-jun, Li Fen, Xu Shao-ping, et al. Blind denoising model based on unsupervised deep image generation[J].Application Research of Computers,2022, 39(7): 2224-2229.
[4]	孙晓辉, 蔡永洪, 林雁飞. 基于局部结构形态改进图像边缘限幅滤波算法研究[J]. 计量学报, 2022, 43(1): 21-25.
	Sun Xiao-hui, Cai Yong-hong, Lin Yan-fei. Improved image edge clipping and filtering algorithm based on local structural shape[J]. Acta Metrologica Sinica, 2022, 43(1): 21-25.
[5]	管继祥, 李林, 张超, 等. 基于组合滤波的CMOS图像固定模式噪声降噪算法[J]. 应用激光, 2022, 42(4): 173-180.
	Guan Ji-xiang, Li Lin, Zhang Chao, et al. Fixed pattern noise removal algorithm for CMOS image based on combined filter[J]. Applied Laser, 2022, 42(4): 173-180.
[6]	郑艳, 何欢, 卜丽静, 等. 自相似性和边缘保持分解的超分辨率重建算法[J]. 测绘通报, 2022(7): 54-59.
	Zheng Yan, He Huan, Bu Li-jing,et al.Super-resolution reconstruction method based on self-similarity and edge-preserving decomposition[J].Bulletin of Surveying and Mapping, 2022(7): 54-59.
[7]	Dhillon D, Chouhan R. Edge-preserving image denoising using noise-enhanced patch-based non-local means[J]. Multimedia Systems, 2023, 29(3): 1025-1041.
[8]	张莹莹, 任超, 朱策. 基于形状自适应非局部回归和非局部梯度正则的深度图像超分辨[J]. 计算机应用, 2022, 42(6): 1941-1949.
	Zhang Ying-ying, Ren Chao, Zhu Ce. Depth image super-resolution based on shape-adaptive non-local regression and non-local gradient regularization[J]. Journal of Computer Applications, 2022, 42(6): 1941-1949.
[9]	Hongyu W, Ying L, Songtao D,et al.Adaptive denoising for magnetic resonance image based on nonlocal structural similarity and low-rank sparse representation[J]. Cluster Computing, 2023, 26(5):2933-2946.
[10]	Abdelkader R, Ramou N, Khorchef M.Welding defects detection in radiographic images using an improved denoising technique combined with an enhanced Chan-Vese model[J].International Journal of Engineering Research in Africa, 2022, 60: 155-172.
[11]	龙超, 金恒, 黎玲, 等. 基于特征融合的非局部均值CT图像降噪[J]. 光学学报, 2022, 42(11): 281-291.
	Long Chao, Jin Heng, Li Ling,et al. CT image denoising with non-local means based on feature fusion[J]. Acta Optica Sinica,2022, 42(11): 281-291.
[12]	Zhang H, He Z, Wang X.A novel mesh denoising method based on relaxed second-order total generalized variation[J]. SIAM Journal on Imaging Sciences, 2022, 15(1): 1-22.
[13]	王春智, 牛宏侠. 基于直方图均衡化与MSRCR的沙尘降质图像增强算法[J]. 计算机工程, 2022, 48(9): 223-229.
	Wang Chun-zhi, Niu Hong-xia.Sand-dust degraded image enhancement algorithm based on histogram equalization and MSRCR[J]. Computer Engineering, 2022, 48(9): 223-229.
[14]	Popa J, Lou Y, Minkoff S E.Low-rank tensor data reconstruction and denoising via ADMM: Algorithm and convergence analysis[J].Journal of Scientific Computing, 2023, 97(2): 1-26.
[15]	黄小莉, 陈春梅, 刘桂华. 混合二阶全变分的抗核辐射图像降噪方法[J]. 重庆邮电大学学报: 自然科学版, 2022, 34(4): 585-594.
	Huang Xiao-li, Chen Chun-mei, Liu Gui-hua.A hybrid second-order total variational noise reduction method for radiation-resistant images[J].Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition), 2022, 34(4): 585-594.
[16]	An B, Kim Y.Image link through adaptive encoding data base and optimized GPU algorithm for real-time image processing of artificial intelligence[J].Journal of Web Engineering, 2022, 21(2): 459-496.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

参数名称	参数值
图像块大小	8×8
差异度阈值	100
搜索窗口大小	32×32
相似度阈值	0.7
二值化阈值	20%
滤波窗口大小	5×5
ADS阈值	1 000

基于非局部自相似性的低对比度图像降噪算法

Low contrast image denoising algorithm based on non local self similarity

RICH HTML

PDF (PC)

摘要/Abstract

引用本文

使用本文

图/表 5

参考文献 16

相关文章 2

Metrics

本文评价

推荐阅读 0

[1]	郭昕刚,何颖晨,程超. 抗噪声的分步式图像超分辨率重构算法[J]. 吉林大学学报(工学版), 2024, 54(7): 2063-2071.
[2]	张田, 孙延奎, 田小林. 二进小波与扩散滤波结合的光学相干层析图像降噪[J]. 吉林大学学报(工学版), 2013, 43(增刊1): 340-344.