基于高斯曲率和加权图总变分正则化的遥感图像盲去模糊算法

doi:10.13229/j.cnki.jdxbgxb.20230140

摘要/Abstract

摘要：

针对遥感图像在采集过程中出现的运动模糊现象，本文从图像曲面的几何性质和图像像素的代数性质出发，设计了一种基于高斯曲率和加权图总变分正则化的遥感图像盲去模糊算法。首先，将加权图总变分先验与高斯曲率先验相结合以获得骨架图像，骨架图像保留图像的梯度以及锐利边缘信息，并去除中间潜在图像中的有害结构；然后，利用骨架图像估计模糊核，进而利用非盲去模糊算法获得清晰图像；最后，在8张不同场景下的模糊遥感图像上进行仿真验证。结果表明，相比于其他先进的图像盲去模糊算法，本文提出的去模糊算法复原效果的峰值信噪比平均值分别高于对比算法2.76、1.84、3.11、2.79、3.35、2.76 dB，结构相似性平均值分别高于对比算法0.0792、0.0604、0.0873、0.0801、0.0997、0.0906。本文算法复原的遥感图像具有清晰的边缘轮廓和局部细节，提升了遥感图像的清晰度。

关键词: 计算数学, 遥感图像, 高斯曲率, 加权图总变分, 图像盲去模糊

Abstract:

To tackle the motion blur in the process of acquiring remote sensing images， an algorithm for blind deblurring of remote sensing images is designed. The algorithm is based on the geometric property of image surfaces and the algebraic property of image pixels， and it utilizes Gaussian curvature and reweighted graph total variation. First， the reweighted graph total variational prior and Gaussian curvature prior were combined to obtain the skeleton image which not only retains the gradient and sharp edge information， but also removes the harmful structural information in the latent clean image. Then， the skeleton image is used to estimate the fuzzy kernel， and then the non-blind deblurring algorithm is used to obtain the clear image. Finally, simulation validation was conducted on 8 fuzzy remote sensing images in different scenarios, and the results showed that, compared with other advanced image deblurring algorithms， the peak signal-to-noise ratio of the recovery effect of the deblurring algorithm proposed is higher than that of the comparison algorithm by 2.76， 1.84， 3.11， 2.79， 3.35， 2.76 dB， respectively. The structure similarity is higher than that of the comparison algorithm by 0.0792、0.0604、0.0873、0.0801、0.0997、0.0906， respectively. The remote sensing images recovered by our proposed algorithm have clear edge contours and local details while improving the clarity of remote sensing images.

Key words: computational mathematics, remote sensing image, Gaussian curvature, reweighted graph total variation, blind image deblurring

中图分类号:

TP751

蔡志丹,方明,李喆,许佳路. 基于高斯曲率和加权图总变分正则化的遥感图像盲去模糊算法[J]. 吉林大学学报(工学版), 2023, 53(9): 2649-2658.

Zhi-dan CAI,Ming FANG,Zhe LI,Jia-lu XU. Blind remote sensing image deblurring algorithm based on Gaussian curvature and reweighted graph total variation[J]. Journal of Jilin University(Engineering and Technology Edition), 2023, 53(9): 2649-2658.

图/表 8

图1

图2

图3

图4

图5

图6

表1

图7

参考文献 31

1	王李祺, 候宇超, 高翔, 等. 基于深度学习与狮群SVM算法的遥感场景分类[J]. 吉林大学学报: 理学版, 2023, 61(4): 863-874.
	Wang Li-qi, Hou Yu-chao, Gao Xiang, et al. Remote sensing scene classfication based on deep learning and lion swarm SVM algorithm[J]. Journal of Jilin University(Science Edition), 2023, 61(4): 863-874.
2	张玉波, 王建阳, 韩爽, 等. 一种非对称的轻量级图像盲与模糊网络[J]. 吉林大学学报: 理学版, 2023, 61(2): 362-370.
	Zhang Yu-bo, Wang Jian-yang, Han Shuang, et al. An asymmetric lightweight image blind deblurring network[J]. Journal of Jilin University(Science Edition), 2023, 61(2): 362-370.
3	Lai W S, Ding J J, Lin Y Y, et al. Blur kernel estimation using normalized color-line priors[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Boston, USA, 2015: 64-72.
4	Liu J, Yan M, Zeng T Y. Surface aware blind image deblurring[J]. IEEE transactions on Pattern Analysis and Machine Intelligence, 2021, 43(3): 1041-1055.
5	Nah S, Kim T H, Lee K M. Deep multi-scale convolutional neural network for dynamic scene deblurring[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Hawaii, USA, 2017: 257-265.
6	Kupyn O, Budzan V, Mykhailych M, et al. DeblurGAN: blind motion deblurring using conditional adversarial neworks[C]∥IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Salt Lake City, USA, 2018: 8183-8192.
7	Zhang J W, Pan J S, Ren J, et al. Dynamic scene deblurring using spatially variant recurrent neural networks[C]∥IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR), Salt Lake City, USA, 2018: 2521-2529.
8	Krishnan D, Tay T, Fergus R. Blind deconvolution using a normalized sparsity measure[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Colorado Springs, USA, 2011: 233-240.
9	Xu L, Zheng S C, Jia J Y. Unnatural L0 sparse representation for natural image deblurring[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Portland, USA, 2013: 1107-1114.
10	Pan J, Su Z. Fast regularized kernel estimation for robust motion deblurring[J]. IEEE Signal Processing Letters, 2013, 20(9): 841-844.
11	Pan J S, Sun D Q, Pfister H, et al. Deblurring images via dark channel prior[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 40(10): 2315-2328.
12	岳有军, 云赛, 王红君, 等. 基于暗通道与低秩先验的运动模糊图像盲复原[J]. 电光与控制, 2019, 26(11): 95-98, 110.
	Yue You-jun, Yun Sai, Wang Hong-jun, et al. Blind restoration of motion-blurred image based on dark channel and low-rank prior[J]. Electronics Optics ＆ Control, 2019, 26(11): 95-98, 110.
13	Yan Y Y, Ren W Q, Gou Y F, et al. Image deblurring via extreme channels prior[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, USA, 2017: 6978-6986.
14	Ge X, Tan J, Zhang L. Blind image deblurring using a non-linear channel prior based on dark and bright channels[J]. IEEE Transactions on Image Processing, 2021, 30: 6970-6984.
15	Chen L, Fang F, Wang T, et al. Blind image deblurring with local maximum gradient prior[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition(CVPR), Long Beach, USA, 2019: 1742-1750.
16	Ren Wen-qi, Cao Xiao-chun, Pan Jin-shan, et al. Image deblurring via enhanced low rank prior[J]. IEEE Transactions on Image Processing, 2016, 25(7): 3426-3437.
17	Dong J X, Pan J S, Su Z X. Blur kernel estimation via salient edges and low rank prior for blind image deblurring[J]. Signal Processing: Image Communication, 2017, 58: 134-145.
18	Ge X Y, Tan J Q, Zhang L, et al. Blind image deblurring with Gaussian curvature of the image surface[J]. Signal Processing: Image Communication, 2022, 100: No.116531.
19	Tang Y, Xue Y, Chen Y, et al. Blind deblurring with sparse representation via external patch priors[J]. Digital Signal Processing, 2018, 78: 322-331.
20	Yu J, Chang Z, Xiao C. Blur kernel estimation using sparse representation and cross-scale self-similarity[J]. Multimedia Tools and Applications, 2019, 78(13): 18549-18570.
21	Berger P, Hannak G, Matz G. Graph signal recovery via primaldual algorithms for total variation minimization[J]. IEEE Journal of Selected Topics in Signal Processing, 2017, 11(6): 842-855.
22	Bai Y, Cheung G, Liu X, et al. Graph based blind image deblurring from a single photograph[J]. IEEE Transactions on Image Processing, 2019, 28(3): 1404-1418.
23	Wen F, Ying R D, Liu Y P, et al. A simple local minimal intensity prior and an improved algorithm for blind image deblurring[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2021, 31(8): 2923-2937.
24	Zhang Z Y, Zheng L L, Piao Y J, et al. Blind remote sensing image deblurring using local binary pattern prior[J]. Remote Sensing, 2022, 14(5): No.1276.
25	Li Z Y, Guo J Y, Zhang Y T, et al. Reference based multi-level features fusion deblurring network for optical remote sensing images[J]. Remote Sensing, 2022, 14(11): No.2520.
26	Lim H, Yu S, Park K, et al. Texture aware deblurring for remote sensing images using L0-based deblurring and L2-based fusion[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2020, 13: 3094-3108.
27	王琪瑶, 胡琸悦, 李潇雁, 等. 基于局部最大和最小强度先验的遥感图像盲去模糊[J]. 激光与光电子学进展, 2023, 60(4): 429-436.
	Wang Qi-yao, Hu Zhuo-yue, Li Xiao-yan, et al. Blind deblurring of remote sensing images based on local maximum and minimum intensity prior[J]. Laser and Optoelectronics Progress, 2022, 60(4): 429-436.
28	刘晨辉, 尹增山, 高爽. 基于遥感图像序列的运动去模糊算法研究[J]. 激光与光电子学进展, 2022, 59(8): 496-504.
	Liu Chen-hui, Yin Zeng-shan, Gao Shuang. Motion deblurring algorithm based on remote sensing image sequence[J]. Laser and Optoelectronics Progress, 2022, 59(8): 496-504.
29	Fergus R, Singh B, Hertzmann A, et al. Removing camera shake from a single photograph[J]. ACM Transactions on Graphics, 2006, 25(3): 787-794.
30	Levin A, Weiss Y, Durand F, et al. Understanding and evaluating blind deconvolution algorithms[C]∥IEEE Conference on Computer Vision and Pattern Recognition, Miami, USA, 2009: No.10836014.
31	Krishnan D, Fergus R. Fast image deconvolution using hyper-Laplacian priors[J]. Advances in Neural Information Processing Systems, 2009, 22: 1033-1041.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

图像	文献［11］		文献［22］		文献［15］		文献［14］		文献［23］		文献［18］		本文算法
图像	PSNR	SSIM	PSNR	SSIM	PSNR	SSIM	PSNR	SSIM	PSNR	SSIM	PSNR	SSIM	PSNR	SSIM
海滩	36.21	0.9596	37.43	0.9621	34.84	0.9534	35.27	0.9558	34.61	0.9518	36.86	0.9620	37.45	0.9634
沙漠	34.15	0.9936	34.24	0.9939	34.23	0.9936	34.34	0.9938	33.97	0.9934	35.48	0.9950	35.49	0.9950
学校	22.25	0.8122	22.03	0.7903	21.75	0.7933	21.47	0.7849	22.41	0.8147	21.80	0.7840	25.59	0.8775
度假村	18.75	0.6046	21.96	0.7186	18.87	0.6090	18.49	0.5924	16.35	0.4946	18.32	0.5849	22.03	0.7209
机场	26.59	0.8785	25.97	0.8453	26.74	0.8819	27.81	0.9002	27.13	0.8876	25.78	0.8618	27.86	0.8901
山地	15.45	0.4885	15.50	0.4648	15.08	0.4823	15.49	0.4914	15.53	0.4900	15.99	0.4948	20.27	0.7045
广场	20.64	0.7387	22.39	0.7696	20.56	0.7337	20.92	0.7479	20.12	0.7197	20.15	0.7148	25.05	0.8587
高架桥	24.94	0.6236	26.85	0.7047	24.11	0.5873	24.97	0.6253	24.16	0.5831	24.62	0.6110	27.34	0.7226

[1]	马为駽,张䶮,马传香,朱飒. 不同光照条件下含噪遥感图像边缘检测算法[J]. 吉林大学学报(工学版), 2023, 53(1): 241-247.
[2]	朱冰,李紫薇,李奇. 基于改进SegNet的遥感图像建筑物分割方法[J]. 吉林大学学报(工学版), 2023, 53(1): 248-254.
[3]	李雄飞,吴佳婧,张小利,王泽宇,冯云丛. 基于相对总变差结构提取的遥感图像融合[J]. 吉林大学学报(工学版), 2021, 51(5): 1775-1784.
[4]	李雄飞,宋璐,张小利. 基于协同经验小波变换的遥感图像融合[J]. 吉林大学学报(工学版), 2019, 49(4): 1307-1319.
[5]	邬开俊, 王春丽, 单亚州, 杜三山, 鲁怀伟. 噪声作用下的化学突触耦合神经元系统的同步[J]. 吉林大学学报(工学版), 2017, 47(5): 1554-1560.
[6]	于德新, 刘春宇, 郑坤, 郭亚娟, 赵鑫. 基于遥感图像的车辆空间占有率估计[J]. 吉林大学学报(工学版), 2016, 46(3): 764-769.
[7]	包磊, 徐其志. 基于PCA变换和光谱补偿的遥感影像融合方法[J]. 吉林大学学报(工学版), 2013, 43(增刊1): 88-91.
[8]	张锐，唐志平，郑航 . 离散元与有限元耦合的时空多尺度计算方法[J]. 吉林大学学报(工学版), 2009, 39(02): 408-0412.
[9]	陈浩，张磊，张晔 . 基于JPWL的遥感图像压缩传输错误保护[J]. 吉林大学学报(工学版), 2007, 37(06): 1441-1446.
[10]	张国敏，殷建平，祝恩，强永刚 . 遥感图像中基于视觉显著性的分层目标检测[J]. 吉林大学学报(工学版), 2007, 37(03): 625-0629.