基于多尺度交互结构卷积神经网络的SAR图像相干斑抑制方法

doi:10.12305/j.issn.1001-506X.2021.12.14

Abstract

Abstract:

Combined with the idea of deep learning, a speckle suppression method for synthetic aperture radar (SAR) images based on multi-scale interactive convolutional neural network (CNN) is proposed. Firstly, a multi-scale extraction module is constructed by convolution kernels and skip connection with different sizes to obtain the features of different receptive fields and speed up the convergence of the network. Then, the network can make the best of shallow texture features by using simplified dense connection between multi-scale interactive feature extraction modules. Finally, the suppressed image is obtained by residual learning strategy. The experimental results show that compared with the existing methods, the proposed method not only uses less calculation parameters, but also ensures the improvement of performance.

Key words: deep learning, synthetic aperture radar (SAR), speckle suppression, convolutional neural network (CNN), residual learning strategy

CLC Number:

TP751

Shiyu SHEN, Xiaodong YE, Hao WANG, Shifei TAO. SAR image speckle suppression method based on muti-scale interactive structure convolutional neural network[J]. Systems Engineering and Electronics, 2021, 43(12): 3526-3532.

Figures/Tables 13

Fig.1

Fig.2

Fig.3

Fig.4

Table 1

Table 2

Table 3

Fig.5

Fig.6

Fig.7

Fig.8

Fig.9

Table 4

References 24

1	GOODMAN J W . Some fundamental properties of speckle[J]. Journal of the Optical Society of America, 1976, 66 (11): 1145- 1150. doi: 10.1364/JOSA.66.001145
2	LEE J S . Refined filtering of image noise using local statistics[J]. Computer Graphics & Image Processing, 1981, 15 (4): 380- 389.
3	王彩云, 胡允侃, 吴淑侠. 基于贝叶斯模型的shearlet域SAR图像去噪方法[J]. 系统工程与电子技术, 2017, 39 (6): 1250- 1255.
	WANG C Y , HU Y K , WU S X . Shearlet domain SAR image denoising method based on Bayesian model[J]. Systems Engineering and Electronics, 2017, 39 (6): 1250- 1255.
4	刘书君, 吴国庆, 张新征, 等. 基于Shearlet域系数处理的SAR图像降噪[J]. 系统工程与电子技术, 2015, 37 (9): 2023- 2028.
	LIU S J , WU G Q , ZHANG X Z , et al. SAR image denoising via the process of shearlet coefficients[J]. Systems Engineering and Electronics, 2015, 37 (9): 2023- 2028.
5	刘帅奇, 胡绍海, 肖扬. 基于局部混合滤波的SAR图像去噪[J]. 系统工程与电子技术, 2012, 34 (2): 396- 402. doi: 10.3969/j.issn.1001-506X.2012.02.34
	LIU S Q , HU S H , XIAO Y . SAR image de-noise base on local hybrid filter[J]. Systems Engineering and Electronics, 2012, 34 (2): 396- 402. doi: 10.3969/j.issn.1001-506X.2012.02.34
6	DELEDALLE C , DENIS L , TUPIN F . Iterative weighted maximum likelihood denoising with probabilistic patch-based weights[J]. IEEE Trans.on Image Processing, 2009, 18 (12): 2661- 2672. doi: 10.1109/TIP.2009.2029593
7	PARRILLI S , PODERICO M , ANGELINO C V , et al. A nonlocal SAR image denoising algorithm based on LLMMSE wavelet shrinkage[J]. IEEE Trans.on Geoscience & Remote Sensing, 2012, 50 (2): 606- 616.
8	CHIERCHIA G, COZZOLINO D. SAR image despeckling through convolutional neural networks[C]//Proc. of the IEEE International Geoscience and Remote Sensing Symposium, 2017: 5438-5441.
9	ZHANG K , ZUO W M . Beyond a Gaussian denoiser: residual learning of deep CNN for image denoising[J]. IEEE Trans.on Image Processing, 2017, 26 (7): 3142- 3155. doi: 10.1109/TIP.2017.2662206
10	WANG P , ZHANG H , PATEL V M . SAR image despeckling using a convolutional neural network[J]. IEEE Signal Processing Letters, 2017, 24 (12): 1763- 1767. doi: 10.1109/LSP.2017.2758203
11	LATTARI F , GONZALEZ L B , ASARO F , et al. Deep learning for SAR image despeckling[J]. Remote Sensing, 2019, 11 (13): 1532. doi: 10.3390/rs11131532
12	ZHANG Q , YUAN Q Q . Learning a dilated residual network for SAR image despeckling[J]. Remote Sensing, 2018, 10 (2): 196. doi: 10.3390/rs10020196
13	GUI Y C , XUE L , LI X H . SAR image despeckling using a dilated densely connected network[J]. Remote Sensing Letters, 2018, 9 (7-9): 857- 866.
14	COZZOLINO D, VERDOLIVA L. Nonlocal SAR image despeckling by convolutional neural networks[C]//Proc. of the IEEE International Geoscience and Remote Sensing Syposium, 2019: 5117-5120.
15	LI S T , KANG X , HU J . Image fusion with guided filtering[J]. IEEE Trans.on Image Processing, 2013, 22 (7): 2864- 2875. doi: 10.1109/TIP.2013.2244222
16	LIU S , LIU T . Convolutional neural network and guided filtering for SAR image denoising[J]. Remote Sensing, 2019, 11 (6): 702. doi: 10.3390/rs11060702
17	SZEGEDY C. Going deeper with convolutions[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2015.
18	XIE S, GIRSHICK R. Aggregated residual transformations for deep neural networks[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 5987-5995.
19	HUANG G, LIU Z, LAURENS V D M, et al. Densely connected convolutional networks[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 2261-2269.
20	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2016: 770-778.
21	ULABY F , DOBSON M C . Handbook of radar scattering statistics for terrain[M]. Norwood: Artech House Publishers, 1989.
22	LEE Y, HWANG J. An energy and GPU-computation efficient backbone network for real-time object detection[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, 2019: 752-760.
23	CHENG G , HAN J , LU X . Remote sensing image scene classification: benchmark and state of the art[J]. Proceedings of the IEEE, 2017, 105 (10): 1865- 1883. doi: 10.1109/JPROC.2017.2675998
24	KINGMA D P, BA J. Adam: a method for stochastic optimization[EB/OL][2020-08-30]. http://arxiv.org/pdf/1412.6980.pdf.

层级	网络层结构	卷积核尺寸	输出通道数
L1	卷积核+ReLU	3×3	64
L2	MIFEM-1	-	64
L3	MIFEM-2	-	64
L4	MIFEM-3	-	64
L5	MIFEM-4	-	64
L6	MIFEM-5	-	64
L7	卷积核	3×3	1

方法	ID-CNN	所提方法
网络参数量	223 553	150 161

视数L	指标	噪声图像	PPB	SAR-BM3D	ID-CNN	本文方法
1	PSNR/dB	8.271	19.143	18.726	22.594	24.331
1	SSIM	0.139	0.443	0.436 1	0.656	0.778
2	PSNR/dB	11.261	20.909	21.734	25.157	25.575
2	SSIM	0.217	0.665	0.658	0.811	0.824
4	PSNR/dB	14.304	22.532	24.134	26.667	27.011
4	SSIM	0.322 5	0.743	0.831	0.851	0.862
10	PSNR/dB	18.283	23.746	26.033	28.721	28.956
10	SSIM	0.482	0.776	0.855	0.896	0.901

地区	红框区域	观测SAR图像	PPB	SAR-BM3D	ID-CNN	本文方法
Flevoland	区域Ⅰ	2.76	62.25	14.16	81.45	154.92
Flevoland	区域Ⅱ	2.61	30.69	10.33	35.51	61.77
Deathvalley	区域Ⅰ	10.71	38.96	24.48	36.74	43.28
Deathvalley	区域Ⅱ	3.35	5.61	4.61	4.53	5.62
San Francisco	区域Ⅰ	1.72	3.01	3.29	4.08	3.99
San Francisco	区域Ⅱ	1.45	2.39	2.69	2.82	2.87

[1]	Xiao HAN, Shiwen CHEN, Meng CHEN, Jincheng YANG. Open-set recognition of LPI radar signal based on reciprocal point learning [J]. Systems Engineering and Electronics, 2022, 44(9): 2752-2759.
[2]	Tian MIAO, Hongcheng ZENG, He WANG, Jie CHEN. A fast extraction method of flood areas based on iterative threshold segmentation using spaceborne SAR data [J]. Systems Engineering and Electronics, 2022, 44(9): 2760-2768.
[3]	Caiyun WANG, Yida WU, Jianing WANG, Lu MA, Huanyue ZHAO. SAR image target recognition based on combinatorial optimization convolutional neural network [J]. Systems Engineering and Electronics, 2022, 44(8): 2483-2487.
[4]	Limin ZHANG, Kaiwen TAN, Wenjun YAN, Yuyuan ZHANG. Radar emitter recognition based on multi-level jumper residual network [J]. Systems Engineering and Electronics, 2022, 44(7): 2148-2156.
[5]	Guodong JIN, Yuanliang XUE, Lining TAN, Jiankun XU. Advances in object tracking algorithm based on siamese network [J]. Systems Engineering and Electronics, 2022, 44(6): 1805-1822.
[6]	Dongning FU, Guisheng LIAO, Yan HUANG, Bangjie ZHANG, Xing WANG. Time-varying narrow-band interference suppression algorithm for SAR based on graph Laplacian embedding [J]. Systems Engineering and Electronics, 2022, 44(6): 1846-1853.
[7]	Minghui GAI, Su ZHANG, Weitian SUN, Yude NI, Lei YANG. Structural-feature enhancement of SAR targets based on complex value compatible total variation [J]. Systems Engineering and Electronics, 2022, 44(6): 1862-1872.
[8]	Xiaofeng ZHAO, Yebin XU, Fei WU, Jiahui NIU, Wei CAI, Zhili ZHANG. Ground infrared target detection method based on global sensing mechanism [J]. Systems Engineering and Electronics, 2022, 44(5): 1461-1467.
[9]	Penghui JI, Dahai DAI, Shiqi XING, Dejun FENG. Dense false moving targets generation method [J]. Systems Engineering and Electronics, 2022, 44(5): 1502-1511.
[10]	Hong ZOU, Chenyang BAI, Peng HE, Yaping CUI, Ruyan WANG, Dapeng WU. Edge service placement strategy based on distributed deep learning [J]. Systems Engineering and Electronics, 2022, 44(5): 1728-1737.
[11]	Dong CHEN, Yanwei JU. Ship object detection SAR images based on semantic segmentation [J]. Systems Engineering and Electronics, 2022, 44(4): 1195-1201.
[12]	Jingming SUN, Shengkang YU, Jun SUN. Pose sensitivity analysis of HRRP recognition based on deep learning [J]. Systems Engineering and Electronics, 2022, 44(3): 802-807.
[13]	Lei YANG, Su ZHANG, Minghui GAI, Cheng FANG. High-resolution SAR imagery with enhancement of directional structure feature [J]. Systems Engineering and Electronics, 2022, 44(3): 808-818.
[14]	Hengyan LIU, Limin ZHANG, Wenjun YAN, Zhaogen ZHONG, Qing LING, Xiaojun LIANG. LDPC decoding based on WBP-CNN algorithm [J]. Systems Engineering and Electronics, 2022, 44(3): 1030-1035.
[15]	Kai SHAO, Miaomiao ZHU, Guangyu WANG. Modulation recognition method based on generative adversarial andconvolutional neural network [J]. Systems Engineering and Electronics, 2022, 44(3): 1036-1043.

SAR image speckle suppression method based on muti-scale interactive structure convolutional neural network

RichHTML

PDF (PC)

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 13

References 24

Related Articles 15

Recommended Articles

Metrics

Comments