采用深度学习的极化SAR地物分类方法综述

doi:10.19665/j.issn1001-2400.2023.03.015

摘要/Abstract

摘要：

极化合成孔径雷达 (PolSAR) 能够实现全天时、全天候的成像,因此该数据成为遥感数据的主要来源之一。其中地物分类是极化SAR数据解译的重要研究内容,已成为该研究领域的热点之一,目前在军事和民事领域都有着广泛的应用。近年来深度学习已在多个研究领域取得了显著成果,并且在极化SAR图像解译领域也获得了一定的成效。与传统的图像分类方法相比,深度学习方法具有自动提取特征、泛化性能强及获取较高准确率等优势。围绕极化SAR数据解译中的地物分类问题,对现有采用深度学习的极化SAR图像地物分类方法进行综述。根据深度学习中不同的网络模型,主要从三方面对极化SAR地物分类研究进行了详细叙述,即基于深度信念网络,稀疏自编码网络以及卷积神经网络的图像分类模型。最后,通过与经典的极化SAR分类方法进行性能评估和比较,总结采用深度学习的极化SAR地物分类方法的优势与不足,同时对该领域未来的发展趋势进行分析和探讨。

关键词: 极化合成孔径雷达, 图像分类, 深度学习, 研究综述

Abstract:

Polarimetric synthetic aperture radar (PolSAR) is one of the main sources of remote sensing data,because it can realize all-day and all-weather imaging.Terrain classification is an important research in the field of PolSAR data interpretation,which has become one of the hotspots in the research field and has been widely used in both military and civilian applications.In recent years,deep learning has achieved remarkable results in many research fields,some of which have been made in the field of PolSAR image processing.Compared with traditional image classification methods,the deep learning method has the advantages of automatic extracting deep features,strong generalization and high accuracy.In this paper,the existing terrain classification methods for the PolSAR image based on deep learning are reviewed.According to the different network models in deep learning,the research on PolSAR terrain classification is described in detail from three aspects,that is,deep belief network,sparse autoencoder network and convolutional neural network.Finally,the advantages and disadvantages of PolSAR terrain classification based deep learning are summarized in comparison with classical classification methods.Meanwhile,the development trend of PolSAR terrain classification is analyzed and discussed.

Key words: polarimetric synthetic aperture radar, image classification, deep learning, research review

中图分类号:

谢雯,滑文强,焦李成,王若男. 采用深度学习的极化SAR地物分类方法综述[J]. 西安电子科技大学学报, 2023, 50(3): 151-170.

XIE Wen,HUA Wenqiang,JIAO Licheng,WANG Ruonan. Review on polarimetric SAR terrain classification methods using deep learning[J]. Journal of Xidian University, 2023, 50(3): 151-170.

图/表 10

图1

图2

表1

图3

图4

表2

图5

图6

表3

表4

参考文献 109

[1]	FABIAN S P, PAN J Y, ADAM T D. Estimation of Significant Wave Height of Near-Range Traveling Ocean Waves Using Sentinel-1 SAR Images[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2019, 12(4):1067-1075. doi: 10.1109/JSTARS.4609443
[2]	CORNRLIU O D, GOTTFRIED S, MIHAI D. SAR Image Land Cover Datasets for Classification Benchmarking of Temporal Changes[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2018, 11(5):1571-1592. doi: 10.1109/JSTARS.4609443
[3]	LIU G W, ZHANG X, MENG J M. A Small Ship Target Detection Method Based on Polarimetric SAR[J]. Remote Sensing, 2019, 11(24):2938. doi: 10.3390/rs11242938
[4]	张强, 杨欣朋, 赵世祥, 等. 注意力机制的SAR图像车辆目标检测网络[J]. 西安电子科技大学学报, 2023, 50(1):36-47.
	ZHANG Qiang, YANG Xinpeng, ZHAO Shixiang, et al. Vehicle-Target Detection Network for SAR Images Based on the Attention Mechanism[J]. Journal of Xidian University, 2023, 50(1):36-47.
[5]	王明春, 张嘉峰, 杨子渊, 等. Bata分布下基于白化滤波的极化SAR图像海绵舰船CFAR检测方法[J]. 电子学报, 2019, 47(9):1883-1890. doi: 10.3969/j.issn.0372-2112.2019.09.010
	WANG Mingchun, ZHANG Jiafeng, YANG Ziyuan, et al. A CFAR Detection Method of Ship Targets in Polarimetric SAR Imagery Based on Whitening Filter Under Beta Distributed Texture[J]. Acta Electronica Sinica, 2019, 47(9):1883-1890. doi: 10.3969/j.issn.0372-2112.2019.09.010
[6]	焦李成, 刘芳, 李玲玲, 等. 遥感影像深度学习智能解译与识别[M]. 西安: 西安电子科技大学出版社, 2019.
[7]	焦李成, 冯婕, 刘芳, 等. 高分辨遥感影像学习与感知[M]. 北京: 科学出版社, 2017.
[8]	LIM Y X, BURGIN M S, VANZYL J J. An Optimal Nonnegative Eigenvalue Decomposition for the Freeman and Durden Three-Component Scattering Model[J]. IEEE Transactions on Geoscience and Remote Sensing, 2017, 55(4):2167-2176. doi: 10.1109/TGRS.2016.2637882
[9]	LEE J S, AINSWORTH T L, WANG Y, et al. Polarimetric SAR Speckle Filtering and the Extended Sigma Filter[J]. IEEE Transactions on Geoscience and Remote Sensing, 2015, 53(3):1150-1160. doi: 10.1109/TGRS.2014.2335114
[10]	WANG H P, FENG X, JIN Y Q, et al. A Review of PolSAR Image Classification:From Polarimetry to Deep Learning[C]//Proceeding of IEEE International Geoscience and Remote Sensing Symposium (IGARSS). Piscataway:IEEE, 2019:3189-3192.
[11]	HUA W Q, WANG S, LIU H Y, et al. Semi-Supervised PolSAR Image Classification Based on Improved Co-Training[J]. IEEE Journal of Selected Topics in Applied Observations and Remote Sensing, 2017, 10(11):4971-4986.
[12]	曹芳, 洪文, 吴一戎. 基于Cloude-Pottier目标分解核聚合的层次聚类算法的全极化SAR数据的非监督分类算法研究[J]. 电子学报, 2008, 36(3):543-546.
	CAO Fang, HONG Wen, WUYirong. An Unsupervised Classification for Fully Polarimetric SAR Data Using Cloude Pottier Decomposition and Agglomerative Hierarchical Clustering Algorithm[J]. Acta Electronica Sinica, 2008, 36(3):543-546.
[13]	YANG J, PENG Y N, YAMAGUCHI Y, et al. On Huynen’s Decomposition of Kennaugh Matrix[J]. IEEE Geoscience and Remote Sensing Letters, 2006, 3(3):369-372. doi: 10.1109/LGRS.2006.873229
[14]	FREEMAN A, DURDEN S L. A Three-Component Scattering Model for Polarimetric SAR Data[J]. IEEE Transactions on Geoscience and Remote Sensing, 1998, 36(3):963-973. doi: 10.1109/36.673687
[15]	YAMAGUCHI Y, MORIYAMA T, ISHIO M. Four-Component Scattering Model for Polarimetric SAR Image Decomposition[J]. IEEE Transactions on Geoscience and Remote Sensing, 2005, 43(8):1699-1706. doi: 10.1109/TGRS.2005.852084
[16]	闫丽丽, 张继贤, 高井祥, 等. 一种适合方位建筑物的基于物理散射模型的极化SAR影像四分量分解方法[J]. 电子学报, 2015, 43(1):203-208. doi: 10.3969/j.issn.0372-2112.2015.01.032
	YAN Lili, ZHANGJixian, GAO Jinxiang, et al. Four-Component Model-Based Decomposition of Polarimetric SAR Data for Oriented Urban Buildings[J]. Acta Electronica Sinica, 2015, 43(1):203-208. doi: 10.3969/j.issn.0372-2112.2015.01.032
[17]	POTTIER E, CLOUDE S R. Application of the H/A/a Polarimetric Decomposition Theorems for Land Classification[J]. The International Society for Optical Engineering, 1997:132-143.
[18]	MANICKAM S, BHATTACHARYA A, SINGH G, et al. Estimation of Snow Surface Dielectric Constant from Polarimetric SAR Data[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2017, 10(1):211-218. doi: 10.1109/JSTARS.4609443
[19]	WANG Y H, LIU HW, JIU B. PolSAR Coherency Matrix Decomposition Based on Constrained Sparse Representation[J]. IEEE Transactions on Geoscience and Remote Sensing, 2017, 55(12):6683-6694. doi: 10.1109/TGRS.2017.2727067
[20]	宋婉莹, 李明, 张鹏, 等. 基于加权合成核与三重Markov场的极化SAR图像分类方法[J]. 电子学报, 2016, 44(3):520-526. doi: 10.3969/j.issn.0372-2112.2016.03.004
	SONG Wanying, LI Ming, ZHANG Peng, et al. A Classification Method of PolSAR Image Based on Weighted Composite Kernel and Triplet Markov Field[J]. Acta Electronica Sinica, 2016, 44(3):520-526. doi: 10.3969/j.issn.0372-2112.2016.03.004
[21]	UHLMANN S, KIRANYAZ S. Integrating Color Features in Polarimetric SAR Image[J]. IEEE Transactions on Geoscience and Remote Sensing, 2014, 55(12):6683-6694. doi: 10.1109/TGRS.2017.2727067
[22]	HINTON G E, SALAKHUTDINOV R R. Reducing the Dimensionality of Data with Neural Networks[J]. Science, 2006, 313(5786):504-507. doi: 10.1126/science.1127647 pmid: 16873662
[23]	HINTON G E. Learning Multiple Layers of Representation[J]. Trends in Cognitive Sciences, 2007, 11(10):428-434. pmid: 17921042
[24]	NG. A.Sparse Autoencoder. CS294A Lecture notes 72 (2011)[R/OL].[2011-12-31]. http://graphics.stanford.edu/courses/cs233-22-spring/ReferencedPapers/SAE.pdf.
[25]	LECUN Y, BENGIO Y, HINTON G. Deep Learning[J]. Nature, 2015, 521(7553):436-444. doi: 10.1038/nature14539
[26]	CHEN Y Q, LI YY, JIAO L C, et al. Adversarial Reconstruction-Classification Networks for PolSAR Image Classification[J]. Remote Sensing, 2019, 11:415. doi: 10.3390/rs11040415
[27]	LIU H Y, LUO R Y, SHANG F H, et al. Semi-Supervised Deep Metric Learning Networks for Classification of Polarimetric SAR Data[J]. Remote Sensing, 2020, 12:1-14. doi: 10.3390/rs12010001
[28]	LIU M C, HU Y, WANG S, et al. Fully Convolutional Semi-Supervised GAN for PolSAR Classification[C]//Proceeding of IEEE International Geoscience and Remote Sensing Symposium (IGARSS). Piscataway:IEEE, 2018:621-624.
[29]	SUN Q G, LI X F, LI L L, et al. Semi-Supervised Complex-Valued GAN for Polarimetric SAR Image Classification[C]//Proceeding of IEEE International Geoscience and Remote Sensing Symposium (IGARSS). Piscataway:IEEE, 2019:3245-3248.
[30]	LIU F, JIAO LC, TANG X. Task-Oriented GAN for PolSAR Image Classification and Clustering[J]. IEEE Transactions on Neural Networks and Learning Systems, 2019, 9(30):2707-2719.
[31]	周晓光, 匡纲要, 万建伟. 极化SAR图像分类综述[J]. 信号处理, 2008, 24(5):806-812.
	ZHOU Xiaoguang, KUANGGangyao, WAN Jianwei. A Review of Polarimetric SAR Image Classification[J]. Journal of Signal Processing, 2008, 24(5):806-812.
[32]	邹斌, 张腊梅, 孙德明, 等. PolSAR图像信息提取技术及应用的发展[J]. 遥感技术与应用, 2009, 3(26):263-273.
	ZOU Bin, ZHANGLamei, SUN Deming, et al. Development and Application of Information Extraction Using Polarimetric SAR Data[J]. Remote Sensing Technology and Application, 2009, 3(26):263-273.
[33]	WANG H P, FENG X, YA Q J. A Review of PolSAR Image Classification:From Polarimetric to Deep Learning[C]//Proceeding of IEEE International Geoscience and Remote Sensing Symposium (IGARSS). Piscataway:IEEE, 2019:3189-3192.
[34]	聂祥丽, 黄夏渊, 张波, 等. 极化合成孔径雷达图像相干斑抑制和分类方法综述[J]. 自动化学报, 2019, 8(45):1419-1438.
	NIE Xiangli, HUANG Xiayuan, ZHANG Bo, et al. Review on PolSAR Image Speckle Reduction and Classiflcation Methods[J]. Acta Automatica Sinica, 2019, 8(45):1419-1438.
[35]	魏丹, 李渊, 黄丹. 极化SAR图像地物分类方法综述[J]. 计算机系统与应用, 2020, 29(11):29-39.
	WEI Dan, LI Yuan, HUANG Dan. Overview onMethods of Land Classification Based on Polarimetric SAR Images[J]. Computer Systems & Applications, 2020, 29 (11):29-39.
[36]	毕海霞, 魏志强. 深度学习在极化SAR图像分类上的应用综述(英文)[J]. 雷达科学与技术, 2021, 19(5):539-551.
	BI Haixia, WEI Zhiqiang. A Survey:The Application of Deep Learning in PolSAR Image Classification[J]. Radar Science and Technology, 2021, 19(5):539-551.
[37]	邓磊, 付珊珊, 张儒侠. 深度置信网络在极化SAR图像分类中的应用[J]. 中国图象图形学报, 2016, 27(7):933-941.
	DENG Lei, FU Shanshan, ZHANG Ruxia. Application of Deep Belief Network in Polarimetric SAR Image Classification[J]. Journal of Image and Graphics, 2016, 27(7):933-941.
[38]	HOU B, GUO X P, HOU W D, et al. PolSAR Image Classification Based on DBN and Tensor Dimensionality Reduction[C]//Proceeding of IEEE International Geoscience and Remote Sensing Symposium (IGARSS). Piscataway:IEEE, 2018:8448-8451.
[39]	GE S J, LU J C, GU H, et al. Polarimetric SAR Image Classification Based on Deep Belief Network and Superpixel Segmentation[C]//Proceeding of IEEE International Conference on Frontiers of Signal Processing (ICFSP). Piscataway:IEEE, 2017:114-119.
[40]	GUO Y H, WANG S, GAO C Q, et al. Wishart RBM Based DBN for Polarimetric Synthetic Radar Data Classification[C]//Proceeding of IEEE International Geoscience and Remote Sensing Symposium (IGARSS). Piscataway:IEEE, 2015:1841-1844.
[41]	LIU F, JIAO L C, HOU B, et al. Pol-SAR Image Classification Based on Wishart DBN and Local Spatial Information[J]. IEEE Transactions on Geoscience and Remote Sensing, 2016, 54(6):3292-3308. doi: 10.1109/TGRS.2016.2514504
[42]	JIAO L C, LIU F. Wishart Deep Stacking Network Fast PolSAR Image Classification[J]. IEEE Transactions on Image Processing, 2016, 25(7):3273-3286. doi: 10.1109/TIP.2016.2567069
[43]	HINYON G E, SALAKHUTDINOV R R. Reducing the Dimensionality of Data with Neural Networks[J]. Science, 2006, 313(28):504-507. doi: 10.1126/science.1127647
[44]	XIE H M, WANG S, LIU K, et al. Multilayer Feature Learning for Polarimetric Synthetic Radar Data Classification[C]//Proceeding of IEEE International Geoscience and Remote Sensing Symposium (IGARSS). Piscataway:IEEE, 2014:2818-2821.
[45]	张姝茵, 候彪, 焦李成, 等. 基于稀疏自编码器和边缘保持的Wishart马尔科夫随机场的极化SAR图像分类[J]. 红外线与毫米波学报, 2018, 37(2):177-183.
	ZHANG Shuyin, HOU Biao, JIAO Licheng, et al. PolSAR Image Classification Based on Sparse Autoencoder and Boundary-Preserved WMRF[J]. Journal of Infrared and Millimeter Waves, 2018, 37(2):177-183.
[46]	HOU B, KOU H D, JIAO L C, et al. Classification of Polarimetric SAR Images Using Multilayer Autoencoders and Superpixels[J]. IEEE Journal of Selected Topics in Applied Observations and Remote Sensing, 2016, 9(7):3072-3081.
[47]	MORI G, REN X, EFROS A A, et al. Recovering Human Body Configurations:Combining Segmentation and Recognition[C]//Proceeding of IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway:IEEE, 2004:326-333.
[48]	ZHANG L, MA W P, ZHANG D. Stacked Sparse Autoencoder in PolSAR Data Classification Using Local Spatial Information[J]. IEEE Geoscience and Remote Sensing Letters, 2016, 13(9):1359-1363. doi: 10.1109/LGRS.2016.2586109
[49]	ZHANG L, JIAO L C, MA W P, et al. PolSAR Image Classification Based on Multi-Scale Stacked Sparse Autoencoder Sparse[J]. Neurocomputing, 2019, 351:167-179. doi: 10.1016/j.neucom.2019.03.024
[50]	HU YY, FAN J C, WANG J. Classification of PolSAR Images Based on Adaptive Nonlocal Stacked Sparse Autoencoder[J]. IEEE Geoscience and Remote Sensing Letters, 2018, 15(7):1050-1054. doi: 10.1109/LGRS.2018.2829182
[51]	BUADES A, COLL B, MOREL J M. A Non-Local Algorithm for Image Denoising[C]//Proceeding of IEEE Computer Society Conference on Computer Vision and pattern recognition (CVPR). Piscataway:IEEE, 2005:60-65.
[52]	DE S, RATHA D, RATHA D, et al. Tensorization of Multifrequency PolSAR Data for Classification Using an Autoencoder Network[J]. IEEE Geoscience and Remote Sensing Letters, 2018, 15(4):542-546. doi: 10.1109/LGRS.2018.2799875
[53]	GENG J, MA X R, FAN J C, et al. Semisupervised Classification of Polarimetric SAR Image via Superpixel Restrained Deep Neural Network[J]. IEEE Geoscience and Remote Sensing Letters, 2018, 15(1):122-126. doi: 10.1109/LGRS.2017.2777450
[54]	XIE W, JIAO L C, HOU B, et al. PolSAR Image Classification via Wishart-WAE Model or Wishart-CAE Model[J]. IEEE Journal of Selected Topics in Applied Observations and Remote Sensing, 2017, 10(8):3604-3615.
[55]	XIE W, MA G N, HUA W Q, et al. Complex-Valued Wishart Stacked Auto-Encoder Network for PolSAR Image Classification[C]//Proceeding of IEEE International Geoscience and Remote Sensing Symposium (IGARSS). Piscataway:IEEE, 2019:3193-3196.
[56]	WANG J, HOU B, JIAO L C, et al. PolSAR Image Classification Based on Modified Stacked Autoencoder Network and Data Distribution[J]. IEEE Transactions on Geoscience and Remote Sensing, 2020, 58(3):1678-1695. doi: 10.1109/TGRS.36
[57]	CHEN Y Q, JIAO L C, LI Y Y, et al. Multilayer Projective Dictionary Pair Learning and Sparse Autoencoder for PolSAR Image Classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2017, 55(12):6683-6694. doi: 10.1109/TGRS.2017.2727067
[58]	石俊飞, 刘芳, 林耀海, 等. 基于深度学习和层次语义模型的极化SAR分类[J]. 自动化学报, 2017, 2(43):215-226.
	SHI Junfei, LIU Fang, LIN Yaohai, et al. Polarimetric SAR Image Classiflcation Based on Deep Learning and Hierarchical Semantic Model[J]. Acta Automatica Sinica, 2017, 2(43):215-226.
[59]	WU J, LIU F, ZHANG X R, et al. Local Maximal Homogeneous Region Search for SAR Speckle Reduction with Sketch-Based Geometrical Kernel Function[J]. IEEE Transactions on Geoscience and Remote Sensing, 2014, 52(9):5751-5764. doi: 10.1109/TGRS.2013.2292081
[60]	LIU F, SHI J F, JIAO L C, et al. Hierarchical Semantic Model and Scattering Mechanism Based PolSAR Image Classification[J]. Pattern Recognition, 2016, 59:325-342. doi: 10.1016/j.patcog.2016.02.020
[61]	CHEN X D, DENG J H. A Robust Polarimetric SAR Terrain Classification Based on Sparse Deep Autoencoder Model Combined with Wavelet Kernel-Based Classifier[J]. IEEE Access, 2020, 8:64810-64819. doi: 10.1109/Access.6287639
[62]	GENG J, FAN J, WANG H, et al. High-Resolution SAR Image Classification via Deep Convolutional Autoencoders[J]. IEEE Geoscience and Remote Sensing Letters, 2015, 12(11):2351-2355. doi: 10.1109/LGRS.8859
[63]	ZHOU Y, WANG H P, XU F et al. Polarimetric SAR Image Classification Using Deep Convolutional Neural Networks[J]. IEEE Geoscience and Remote Sensing Letters, 2016, 13(12):1935-1939. doi: 10.1109/LGRS.2016.2618840
[64]	CHEN S W, TAO C S. PolSAR Image Classification Using Polarimetric-Feature-Driven Deep Convolutional Neural Network[J]. IEEE Geoscience and Remote Sensing Letters, 2018, 15(4):627-631. doi: 10.1109/LGRS.2018.2799877
[65]	韩萍, 孙丹丹. 特征选择与深度学习相结合的极化SAR图像分类[J]. 信号处理, 2019, 6(35):972-978.
	HAN Ping, SUN Dandan. Classification ofPolarimetric SAR Image with Feature Selection and Deep Learning[J]. Journal of Signal Processing, 2019, 6(35):972-978.
[66]	YANG C, HOU B, REN B, et al. CNN-Based Polarimetric Decomposition Feature Selection for PolSAR Image Classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(11):8796-8812. doi: 10.1109/TGRS.36
[67]	HUA W, WANG X, ZHANG C, et al. Attention-Based Multiscale Sequential Network for PolSAR Image Classification[J]. IEEE Geoscience and Remote Sensing Letters, 2022, 19:1-5.
[68]	RUI Y, XIN X, RONG G, et al. Composite Sequential Network with POA Attention for PolSAR Image Analysis[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60:1-15.
[69]	FANG Z, ZHANG G, DAI Q, et al. PolSAR Image Classification Based on Complex-Valued Convolutional Long Short-Term Memory Network[J]. IEEE Geoscience and Remote Sensing Letters, 2022, 19:1-5.
[70]	CUI Y H, LIU F, JIAO L C, et al. Polarimetric Multipath Convolutional Neural Network for PolSAR Image Classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60:1-18.
[71]	GAO F, HUANG T, WANG J, et al. Dual-Branch Deep Convolution Neural Network for Polarimetric SAR Image Classification[J]. Applied Ences, 2017, 7(5):447.
[72]	张佳琪, 张继贤, 赵争. 深度神经网络的高分三号全极化SAR图像分类方法[J]. 测绘科学, 2019, 2(44):6-11.
	ZHANG Jiaqi, ZHANGJixian, ZHAO Zhen. Classification Method of GF-3 Polarimetric SAR Image Using Deep Neural Network[J]. Science of Surveying and Mapping, 2019, 2(44):6-11.
[73]	HE C, HE B K, TU M X, et al. Fully Convolutional Networks and A Manifold Graph Embedding-Based Algorithm for PolSAR Image Classification[J]. Remote Sensing, 2020, 12:1467. doi: 10.3390/rs12091467
[74]	SHELHAMER E, LONG J, DARRELL T. Fully Convolutional Networks for Semantic Segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 39(4):640-651. doi: 10.1109/TPAMI.2016.2572683
[75]	HE C, TU M X, XIONG D H, et al. Nonliner Manifold Learning Integrated with Fully Convolutional Networks for PolSAR Image Classification[J]. Remote Sensing, 2020, 12:665. doi: 10.3390/rs12040665
[76]	陈嘉跃, 李飞. 基于特征融合的全卷积网络极化SAR分类方法[J]. 电子测量技术, 2022, 45(1):104-110.
	CHEN Jiayue, LI Fei. Fully Convolutional Network PolSAR Classification Based on Features Fusion[J]. Electronic Measurement Technology, 2022, 45(1):104-110.
[77]	DING L, ZHENG K, LIN D, et al. MP-ResNet:Multipath Residual Network for the Semantic Segmentation of High-Resolution PolSAR Images[J]. IEEE Geoscience and Remote Sensing Letters, 2021, 19:1-5.
[78]	王云艳, 罗冷坤, 王重阳. Deeplab网络的极化合成孔径雷达图像分类[J]. 测绘科学, 2020, 45(6):1-29.
	WANG Yunyan, LUO Lengkun, WANG Chongyang. PolSAR Image Classification Based on Deeplab Network[J]. Science of Surveying and Mapping, 2020, 45(6):1-29.
[79]	CHEN L, PAPANDREOU G, KOKKINOS I, et al. Deeplab:Semantic Image Segmentation with Deep Convolutional Nets,Atrous Convolution,And Fully Connected CRFs[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018, 40(4):834-848. doi: 10.1109/TPAMI.2017.2699184
[80]	张腊梅, 陈泽茜, 邹斌. 基于3D卷积神经网络的PolSAR图像精细分类[J]. 红外与激光工程, 2018, 47(7):1-8.
	ZHANG Lamei, CHEN Zeqian, ZOU Bin. Fine Classification of Polarimetric SAR Images Based on 3D Convolutional Neural Network[J]. Infrared and Laser Engineering, 2018, 47(7):1-8.
[81]	DONG H W, ZHANG L M, ZOU B. PolSAR Image Classification with Lightweight 3D Convolutional Networks[J]. Remote Sensing, 2020, 12:396. doi: 10.3390/rs12030396
[82]	王睿川, 王岩飞. 基于半监督空间-通道选择性卷积核网络的极化SAR图像地物分类[J]. 雷达学报, 2021, 10(4):516-530.
	WANG Ruichuan, WANG Yanfei. Terrain Classification of Polarimetric SAR Images Using Semi-Supervised Spatial-Channel Selective Kernel Network[J]. Journal of Radars, 2021, 10(4):516-530.
[83]	ZHANG S, YIN Q, NI J, et al. PolSAR Image Classification with Small Sample Learning Based on CNN and CRF[C]//Proceeding of Asia-Pacific Conference on Synthetic Aperture Radar (APSAR). Piscataway:IEEE, 2019:1-5.
[84]	BI H X, XU F, WEI Z Q, et al. An Active Deep Learning Approach for Minimally Supervised PolSAR Image Classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(11):9378-9395. doi: 10.1109/TGRS.36
[85]	JIAO C Z, WANG X L, GOU S P, et al. Self-Paced Convolutional Neural Network for PolSAR Image Classification[J]. Remote Sensing, 2019, 11:424. doi: 10.3390/rs11040424
[86]	REN B, ZHAO YY, HOU B, et al. A Mutual Information-Based Self-Supervised Learning Model for PolSAR Land Cover Classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2021, 59(11):9224-9237. doi: 10.1109/TGRS.2020.3048967
[87]	ZHANG L M, ZHANG S Y, ZOU B, et al. Unsupervised Deep Representation Learning and Few-Shot Classification of PolSAR Images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2020, 60:1-16.
[88]	BI H X, SUN J, XU Z B. A Graph-Based Semisupervised Deep Learning Model for PolSAR Image Classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(4):2116-2132. doi: 10.1109/TGRS.36
[89]	HOU B, GUAN JJ, WU Q, et al. Semi-Supervised Classification of PolSAR Image Incorporating Labels’ Semantic Priors[J]. IEEE Geoscience and Remote Sensing Letters, 2019:1-5.
[90]	GUO Y W, SUN ZZ, QU R, et al. Fuzzy Superpixels Based Semi-Supervised Similarity-Constrained CNN for PolSAR Image Classification[J]. Remote Sensing, 2020, 12:1-18. doi: 10.3390/rs12010001
[91]	ZHANG Z M, WANG H P, XU F, et al. Complex-Valued Convolutional Neural Network and Its Application in Polarimetric SAR Image Classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2017, 55(12):7177-7188. doi: 10.1109/TGRS.2017.2743222
[92]	QIN XX, HU T, ZOU H X, et al. PolSAR Image Classification via Complex-valued Convolutional Neural Network Combining Measured Data and Artificial Features[C]//Proceeding of IEEE International Geoscience and Remote Sensing Symposium (IGARSS). Piscataway:IEEE, 2019:3209-3212.
[93]	XIE W, MA G N, ZHAO F, et al. PolSAR Image Classification via a Novel Semi-Supervised Recurrent Complex-Valued Convolution Neural Network[J]. Neurocomputing, 2020, 388:255-268. doi: 10.1016/j.neucom.2020.01.020
[94]	ADUGNA G, MULLISSA C P, ALFRED S. PolSARNet:A Deep Fully Convolutional Network for Polarimetric SAR Image Classification[J]. IEEE Journal of Selected Topics in Applied Observations and Remote Sensing, 2019, 12(12):5300-5309.
[95]	TAN X F, LI M, ZHANG P, et al. Complex-Valued 3-D Convolutional Neural Network for PolSAR Image Classification[J]. IEEE Geoscience and Remote Sensing Letters, 2020, 17(6):1022-1026. doi: 10.1109/LGRS.8859
[96]	CAO Y C, WU Y, ZHANG P, et al. Pixel-Wise PolSAR Image Classification via a Novel Complex-Valued Deep Fully Convolutional Network[J]. Remote Sensing, 2019, 11:2653. doi: 10.3390/rs11222653
[97]	MULLISSA A G, PERSELLO C, REICHE J. Despeckling Polarimetric SAR Data Using a Multistream Complex-Valued Fully Convolutional Network[J]. IEEE Geoscience and Remote Sensing Letters, 2021, 19:1-5.
[98]	LIU X, JIAO L C, TANG X, et al. Polarimetric Convolutional Network for PolSAR Image Classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(5):3040-3054. doi: 10.1109/TGRS.36
[99]	SHANG R H, WANG GG, MICHAEL A, et al. Complex-Valued Convolutional Autoencoder and Spatial Pixel-Squares Refinement for Polarimetric SAR Image Classification[J]. Remote Sensing, 2019, 11:522. doi: 10.3390/rs11050522
[100]	肖东凌, 刘畅. 基于精调的膨胀编织-交叉卷积网络(FDGC-CNN)的极化SAR地物分类[J]. 雷达学报, 2019, 8(4):479-488.
	XIAO Dongling, LIU Chang. PolSAR Terrain Classification Based on Fine-Tuned Dilated Group-Cross Convolution Neural Network[J]. Journal of Radars, 2019, 8(4):479-488.
[101]	TAN X F, LI M, ZHANG P, et al. Deep Triplet Complex-Valued Network for PolSAR Image Classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2021, 59(12):10179-10196. doi: 10.1109/TGRS.2021.3053013
[102]	GUO Y H, WANG S, SONG G X, et al. PolSAR Terrain Classification Based on Denoising-CNN[C]//Proceeding of IEEE International Geoscience and Remote Sensing Symposium (IGARSS). Piscataway:IEEE, 2019:3213-3216.
[103]	LIU F, WANG J, TANG X, et al. Adaptive Graph Convolutional Network for PolSAR Image Classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2021, 60:1-14.
[104]	DONG H W, ZOU B, ZHANG L M, et al. Automatic Design of CNNs via Differentiable Neural Architecture Search for PolSAR Image Classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2020, 58(9):6362-6375. doi: 10.1109/TGRS.36
[105]	ZHANG X Z, XIA J L, TAN X H, et al. PolSAR Image Classification via Learned Superpixels and QCNN Integrating Color Features[J]. Remote Sensing, 2019, 11:1831. doi: 10.3390/rs11151831
[106]	王云艳, 何楚, 赵守能, 等. 基于多层反卷积网络的SAR图像分类[J]. 武汉大学学报:自然科学版, 2015, 40(10):1371-1346.
	WANG Yunyan, HE Chu, ZHAO Shouneng, et al. Classification of SAR Images Based on Deep Deconvolutional Network[J]. Geomatics and Information Science of Wuhan University, 2015, 40(10):1371-1346.
[107]	赵泉华, 谢凯浪, 王光辉, 等. 全卷积网络和条件随机场相结合的全极化SAR土地覆盖分类[J]. 测绘学报, 2020, 1(49):65-78.
	ZHAO Quanhua, XIE Kailiang, WANG Guanghui, et al. Land Cover Classification of Polarimetric SAR with Fully Convolution Network and Conditional Random Field[J]. Acta Geodaetica et Cartographica Sinica, 2020, 1(49):65-78.
[108]	HUANG K, NIE W, LUO N X. Fully Polarized SAR imagery Classification Based on Deep Reinforcement Learning Method Using Multiple Polarimetric Features[J]. IEEE Journal of Selected Topics in Applied Observations and Remote Sensing, 2019, 12(10):3719-3730.
[109]	NIE W, HUANG K, YANG J, et al. A Deep Reinforcement Learning-Based Framework for PolSAR Imagery Classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60(60):1-15.

种类	方法
种类	Wishart	SVM	DBN^[37]	WRBM^[40]	WBRBM^[41]	GRBM^[40]
Stem beans	93.36	84.06	76.76	94.78	96.60	95.65
Rapeseed	61.41	62.56	51.89	80.42	86.71	83.26
Bare soil	87.14	92.86	94.95	93.64	96.99	97.96
Potatoes	77.66	82.53	87.61	98.11	88.88	92.69
Beet	83.70	93.95	87.54	96.30	95.81	97.45
Wheat 2	73.61	71.67	81.03	94.66	88.17	82.19
Peas	86.52	91.38	91.44	92.07	96.80	95.97
Wheat 3	90.77	95.67	97.66	97.82	93.77	98.02
Lucerne	94.61	79.21	95.70	92.45	96.11	96.91
Barley	96.95	95.31	59.28	98.88	98.06	91.80
Wheat	95.17	90.23	94.34	93.86	92.32	98.14
Grasses	62.58	72.95	10.03	81.52	90.93	67.64
Forest	93.80	90.57	92.52	95.34	91.41	95.64
Water	40.33	90.99	68.95	99.74	99.60	89.59
Building	87.35	1.63	23.95	82.45	85.26	77.69
OA	81.66	79.70	74.24	92.80	93.08	90.71
Kappa	0.800 7	0.778 6	0.719 1	0.921 5	0.924 9	0.898 7

种类	方法
种类	DBN^[37]	WRBM^[40]	SAE^[44]	SAE- BWMRF^[45]	SSAE- SL^[48]	WAE^[54]	CV-WSAE^[55]	CAE^[62]	MAE^[56]
Stem beans	79.05	95.17	84.22	98.91	98.80	92.42	78.89	78.98	95.91
Rapeseed	39.33	64.31	74.19	66.49	87.82	78.80	72.90	83.32	84.11
Bare soil	0.00	39.95	93.78	99.88	97.73	96.06	76.70	79.24	92.62
Potatoes	93.74	87.00	81.55	97.33	93.74	93.70	96.30	98.48	89.64
Beet	93.25	89.33	88.96	98.79	96.52	89.06	94.37	96.72	95.77
Wheat 2	89.52	62.63	64.88	93.51	87.21	90.62	87.60	85.93	81.02
Peas	93.77	93.33	92.85	97.91	95.72	78.20	41.94	49.22	96.42
Wheat 3	97.70	95.45	94.32	94.15	96.70	85.69	79.52	85.02	95.06
Lucerne	80.60	93.84	94.72	96.88	96.91	93.52	92.13	93.83	95.34
Barley	75.98	93.51	94.38	88.86	94.79	93.04	97.11	97.76	95.98
Wheat	92.01	90.18	85.14	92.08	93.55	91.23	94.76	96.12	91.57
Grasses	42.66	56.55	77.20	80.94	93.36	92.39	96.60	98.64	86.41
Forest	94.08	89.52	93.57	90.49	93.95	93.48	87.56	88.49	91.13
Water	91.77	83.64	99.30	35.70	99.61	86.89	90.37	92.20	98.02
Building	0.00	8.57	37.69	81.77	87.48	85.39	59.74	72.99	84.09
OA	70.86	76.19	83.78	87.58	94.26	89.97	83.10	86.46	92.01
Kappa	0.681 7	0.740 4	0.823 1	0.865 1	0.937 8	0.885 7	0.818 2	0.854 0	0.912 7

种类	方法
种类	DBN^[37]	SAE^[44]	CNN^[63]	3D- CNN^[80]	FCN- SPUO^[82]	MI-SSL^[86]	KGraph- CNN^[88]	CNNSP^[89]	CV-3D- CNN^[95]	PCN^[98]
Stem beans	79.05	84.22	83.81	97.52	99.98	99.95	99.07	98.58	98.63	92.11
Rapeseed	39.33	74.19	67.52	90.03	98.87	99.64	99.15	94.44	97.48	95.76
Bare soil	0.00	93.78	60.72	97.67	99.91	99.43	100.00	99.67	92.74	98.71
Potatoes	93.74	81.55	90.75	90.78	99.44	99.97	99.63	99.73	93.60	96.49
Beet	93.25	88.96	95.25	91.47	98.93	99.95	99.76	97.11	95.21	91.07
Wheat 2	89.52	64.88	86.34	86.08	98.07	99.75	85.86	99.32	95.73	96.27
Peas	93.77	92.85	98.46	96.45	99.03	99.86	99.96	99.55	87.65	97.52
Wheat 3	97.70	94.32	97.01	97.46	99.78	99.95	99.01	98.21	99.44	96.87
Lucerne	80.60	94.72	80.54	95.31	97.45	99.98	98.10	92.33	84.81	96.13
Barley	75.98	94.38	85.62	97.46	99.17	99.60	98.08	97.29	84.14	92.23
Wheat	92.01	85.14	91.36	92.92	99.56	99.82	99.71	96.87	98.79	96.62
Grasses	42.66	77.20	64.86	91.54	99.68	99.51	92.50	93.41	72.39	96.32
Forest	94.08	93.57	87.49	96.03	99.89	99.98	98.81	99.19	99.85	99.05
Water	91.77	99.30	91.82	93.66	99.50	100.00	98.69	99.07	99.95	98.06
Building	0.00	37.69	85.44	73.61	90.51	96.95	76.84	95.78	96.22	81.26
OA	70.86	83.78	84.47	92.53	99.24	99.84	96.34	97.37	93.42	96.36
Kappa	0.681 7	0.823 1	0.830 7	0.918 6	0.991 8	0.995 9	0.960 2	0.971 4	0.928 5	0.960 4

极化SAR图像分类方法	代表性算法	优点	不足
基于DBN的极化SAR 图像分类方法	WRBM^[40], WBRBM^[41]	图像细节信息清晰,边缘保持得好;既包含了信念网络的优点,也保留了极化SAR的数据特性	没有考虑图像的空间信息;斑点噪声较明显;增加了网络的复杂度,训练时间成本增加
基于SAE的极化SAR 图像分类方法	SSAE-SL^[48], SAE-BWMRF^[45]	考虑极化SAR图像像素之间的空间信息	没有考虑极化SAR数据的统计分布信息和物理散射机制
	WAE^[54], MAE^[56]	图像细节信息清晰,边缘保持得好;既继承自编码网络的优点,也保留了极化SAR的数据特性	没有考虑图像的空间信息;斑点噪声较明显;增加了网络的复杂度,训练时间成本增加
	基于CNN的极化SAR 图像分类方法	3D-CNN^[80], CV-CNN^[91], CV-3D-CNN^[95],	图像区域一致性好;考虑了空间信息、特征之间的关系以及极化SAR数据的复数特性	图像细节信息不清晰;增加了网络的复杂度,训练时间成本增加
CV-FCN^[96], PCLNet^[87], KGraph-CNN^[88], CNNSP^[89]		图像区域一致性好;解决小样本问题	图像细节信息不清晰;增加了网络的复杂度,训练时间成本增加