一种基于张量积扩散的非监督极化SAR图像地物分类方法

邹焕新 李美霖 马倩 孙嘉赤 曹旭 秦先祥

邹焕新, 李美霖, 马倩, 等. 一种基于张量积扩散的非监督极化SAR图像地物分类方法[J]. 雷达学报, 2019, 8(4): 436–447. doi: 10.12000/JR19057
引用本文: 邹焕新, 李美霖, 马倩, 等. 一种基于张量积扩散的非监督极化SAR图像地物分类方法[J]. 雷达学报, 2019, 8(4): 436–447. doi: 10.12000/JR19057
ZOU Huanxin, LI Meilin, MA Qian, et al. An unsupervised PolSAR image classification algorithm based on tensor product graph diffusion[J]. Journal of Radars, 2019, 8(4): 436–447. doi: 10.12000/JR19057
Citation: ZOU Huanxin, LI Meilin, MA Qian, et al. An unsupervised PolSAR image classification algorithm based on tensor product graph diffusion[J]. Journal of Radars, 2019, 8(4): 436–447. doi: 10.12000/JR19057

一种基于张量积扩散的非监督极化SAR图像地物分类方法

doi: 10.12000/JR19057
基金项目: 国家自然科学基金(61331015, 41601436)
详细信息
    作者简介:

    邹焕新(1973–),男,广东人,现任国防科技大学电子科学学院教授,硕士生导师,主要研究方向为SAR图像解译、多源信息融合、计算机视觉、图像处理、模式识别等。E-mail: hxzou2008@163.com

    李美霖(1995–),女,山西人,现为国防科技大学电子科学学院硕士,主要研究方向为极化SAR图像地物分类、模式识别等。E-mail: summit_mll@qq.com 

    马倩:马 倩(1996–),女,湖南人,现为国防科技大学电子科学学院硕士,主要研究方向为多源遥感数据变化检测。E-mail: 2233809618@qq.com

    孙嘉赤(1996–),男,山西人,现为国防科技大学电子科学学院硕士,主要研究方向为SAR图像和光学图像目标检测分类与识别。E-mail: 445219733@qq.com

    曹旭:曹 旭(1996–),男,天津人,现为国防科技大学电子科学学院硕士,主要研究方向为SAR图像和光学图像目标检测分类与识别。E-mail: 1135459767@qq.com

    秦先祥(1986–),男,广西人,现任空军工程大学信息与导航学院讲师,主要研究方向为SAR图像解译。E-mail: qinxianxiang@126.com

    通讯作者:

    邹焕新 hxzou2008@163.com

  • 中图分类号: TN957

An Unsupervised PolSAR Image Classification Algorithm Based on Tensor Product Graph Diffusion

Funds: The National Natural Science Foundation of China (61331015, 41601436)
More Information
  • 摘要: 针对相似度表达的困难性以及极化SAR图像中固有的相干斑噪声问题,该文提出了一种基于张量积(TPG)扩散的非监督极化SAR图像地物分类算法。张量积扩散一般用于光学图像的分割或检索,目前研究表明,其已可用于极化SAR(PolSAR)图像地物分类。基于张量积扩散可以稳健地度量数据点之间的测地线距离,因此能够更好地挖掘数据点之间内在的相似度信息。首先,将极化SAR图像进行分割,生成许多超像素;其次,基于超像素提取7种特征并生成一个特征向量,进而利用高斯核构建相似度矩阵;再次,基于已构建的相似度矩阵,利用张量积扩散沿着数据点的内在流形结构进行相似度的传播,实现全局的相似性度量,从而获得一个具有更强判别能力的相似度矩阵;最后,基于此相似度矩阵进行谱聚类以得到地物分类结果。该文在仿真和实测极化SAR图像上均进行了大量实验,并与4种经典算法进行对比,结果表明该方法可以有效地结合空间邻域相似度信息并取得更高的分类精度。

     

  • 图  1  本文算法框架流程图

    Figure  1.  The flowchart of the proposed method

    图  2  张量积图简易示例

    Figure  2.  An example of tensor product graph

    图  3  仿真极化SAR图像

    Figure  3.  The simulated PolSAR image

    图  4  实测极化SAR图像

    Figure  4.  The real-world PolSAR image

    图  5  仿真极化SAR图像的分类结果

    Figure  5.  Classification results of the simulated PolSAR image

    图  6  实测极化SAR图像的分类结果

    Figure  6.  Classification results of the real-world PolSAR image

    图  7  实测极化SAR图像超像素分割实验结果

    Figure  7.  The results of the superpixel segmentation for the real-world PolSAR image

    图  8  实测极化SAR图像在不同参数值Sk$\mu $时的整体精度

    Figure  8.  The OAs for the real-world PolSAR image under different parameters of S, k and $\mu $

    图  9  仿真极化SAR图像5种算法的分类结果

    Figure  9.  Classification results of five methods for the simulated PolSAR image

    图  10  实测极化SAR图像5种算法的分类结果

    Figure  10.  Classification results of five methods for the real-world PolSAR image

    表  1  OM方法基于仿真数据的5种评价度量结果

    Table  1.   The five evaluation criteria of the OM method for the simulated PolSAR image

    类别类别1类别2类别3类别4UA
    类别17822261927580.7362
    类别23975192370.9751
    类别312479228880.9843
    类别414961198790.9879
    PA0.99630.98300.99580.7622
    OA: 91.70%, K: 0.8894
    下载: 导出CSV

    表  2  PM方法基于仿真数据的5种评价度量结果

    Table  2.   The five evaluation criteria of the PM method for the simulated PolSAR image

    类别类别1类别2类别3类别4UA
    类别11030826192720.9701
    类别23975192370.9751
    类别316479228840.9843
    类别415961198780.9878
    PA0.99670.98300.99580.9434
    OA: 97.91%, K: 0.9722
    下载: 导出CSV

    表  3  两种算法基于实测极化SAR图像的整体精度和Kappa系数

    Table  3.   The OAs and Ks of two methods for the real-world PolSAR image

    算法精度
    KOA (%)
    OM算法0.619378.51
    PM算法0.809789.36
    下载: 导出CSV

    表  4  5种算法基于仿真极化SAR图像的3种评价度量结果

    Table  4.   The three evaluation criteria of five methods for the simulated PolSAR image

    算法精度
    类别1类别2类别3类别4KOA (%)
    UCSC0.94000.98570.99190.95040.954196.56
    UKWC0.50120.00000.99190.94590.619471.62
    GDWC0.50710.22420.99580.98080.647073.68
    CPWC0.50071.00000.93120.16330.611871.05
    PM0.99670.98300.99580.94340.972297.91
    下载: 导出CSV

    表  5  5种算法基于实测极化SAR图像的3种评价度量结果

    Table  5.   The three evaluation criteria of five methods for the real-world PolSAR image

    算法精度
    林地开放区1开放区2KOA (%)
    UCSC0.96380.54770.00040.594077.09
    UKWC0.96570.81540.50110.666679.81
    GDWC0.95140.60300.78300.664580.74
    CPWC0.99840.43070.00140.440765.73
    PM0.97070.74610.92260.809789.36
    下载: 导出CSV
  • [1] SHI Lei, ZHANG Lefei, ZHAO Lingli, et al. Adaptive Laplacian Eigenmap-based dimension reduction for ocean target discrimination[J]. IEEE Geoscience and Remote Sensing Letters, 2016, 13(7): 902–906. doi: 10.1109/LGRS.2016.2553046
    [2] 杨文, 钟能, 严天恒, 等. 基于黎曼流形的极化SAR图像分类[J]. 雷达学报, 2017, 6(5): 433–441. doi: 10.12000/JR17031

    YANG Wen, ZHONG Neng, YAN Tianheng, et al. Classification of polarimetric SAR images based on the Riemannian manifold[J]. Journal of Radars, 2017, 6(5): 433–441. doi: 10.12000/JR17031
    [3] LIU Wensong, YANG Jie, LI Pingxiang, et al. A novel object-based supervised classification method with active learning and random forest for PolSAR imagery[J]. Remote Sensing, 2018, 10(7): 1092. doi: 10.3390/rs10071092
    [4] SHI Lei, ZHANG Lefei, ZHAO Lingli, et al. The potential of linear discriminative Laplacian Eigenmaps dimensionality reduction in polarimetric SAR classification for agricultural areas[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2013, 86: 124–135. doi: 10.1016/j.isprsjprs.2013.09.013
    [5] WANG Shuang, LIU Kun, PEI Jingjing, et al. Unsupervised classification of fully polarimetric SAR images based on scattering power entropy and copolarized ratio[J]. IEEE Geoscience and Remote Sensing Letters, 2013, 10(3): 622–626. doi: 10.1109/LGRS.2012.2216249
    [6] LEE J S, GRUNES M R, POTTIER E, et al. Unsupervised terrain classification preserving polarimetric scattering characteristics[J]. IEEE Transactions on Geoscience and Remote Sensing, 2004, 42(4): 722–731. doi: 10.1109/TGRS.2003.819883
    [7] RATHA D, BHATTACHARYA A, and FRERY A C. Unsupervised classification of polsar data using a scattering similarity measure derived from a geodesic distance[J]. IEEE Geoscience and Remote Sensing Letters, 2018, 15(1): 151–155. doi: 10.1109/LGRS.2017.2778749
    [8] LEE J S, GRUNES M R, AINSWORTH T L, et al. Unsupervised classification using polarimetric decomposition and the complex Wishart classifier[J]. IEEE Transactions on Geoscience and Remote Sensing, 1999, 37(5): 2249–2258. doi: 10.1109/36.789621
    [9] 钟能, 杨文, 杨祥立, 等. 基于混合Wishart模型的极化SAR图像非监督分类[J]. 雷达学报, 2017, 6(5): 533–540. doi: 10.12000/JR16133

    ZHONG Neng, YANG Wen, YANG Xiangli, et al. Unsupervised classification for polarimetric synthetic aperture radar images based on Wishart mixture models[J]. Journal of Radars, 2017, 6(5): 533–540. doi: 10.12000/JR16133
    [10] WU Yonghui, JI Kefeng, YU Wenxian, et al. Region-based classification of polarimetric SAR images using Wishart MRF[J]. IEEE Geoscience and Remote Sensing Letters, 2008, 5(4): 668–672. doi: 10.1109/LGRS.2008.2002263
    [11] VON LUXBURG U. A tutorial on spectral clustering[J]. Statistics and Computing, 2007, 17(4): 395–416. doi: 10.1007/s11222-007-9033-z
    [12] YANG Yifang, WANG Yuping, XUE Xingsi, et al. A novel spectral clustering method with superpixels for image segmentation[J]. Optik, 2016, 127(1): 161–167. doi: 10.1016/j.ijleo.2015.10.053
    [13] HU Jingliang, WANG Yuanyuan, GHAMISI P, et al. Evaluation of polsar similarity measures with spectral clustering[C]. Proceedings of 2017 IEEE International Geoscience and Remote Sensing Symposium, Fort Worth, USA, 2017: 3254–3257.
    [14] LI Yonggang, ZHANG Shichao, CHENG Debo, et al. Spectral clustering based on hypergraph and self-re-presentation[J]. Multimedia Tools and Applications, 2017, 76(16): 17559–17576. doi: 10.1007/s11042-016-4131-6
    [15] YANG Xingwei, PRASAD L, and JAN LATECKI L. Affinity learning with diffusion on tensor product graph[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013, 35(1): 28–38. doi: 10.1109/TPAMI.2012.60
    [16] ZHANG Yue, ZOU Huanxin, LUO Tiancheng, et al. A fast superpixel segmentation algorithm for PolSAR images based on edge refinement and revised Wishart distance[J]. Sensors, 2016, 16(10): 1687. doi: 10.3390/s16101687
    [17] CAO Fang, HONG Wen, WU Yirong, et al. An unsupervised segmentation with an adaptive number of clusters using the SPAN/H/α/A space and the complex Wishart clustering for fully polarimetric SAR data analysis[J]. IEEE Transactions on Geoscience and Remote Sensing, 2007, 45(11): 3454–3467. doi: 10.1109/TGRS.2007.907601
    [18] ZHOU Xiaofeng, WANG Shuang, HUA Wenqiang, et al. Unsupervised classification of PolSAR data based on a novel polarization feature[C]. Proceedings of 2017 IEEE International Geoscience and Remote Sensing Symposium, Fort Worth, USA, 2017: 4594–4599.
    [19] 张月, 邹焕新, 邵宁远, 等. 基于相似度网络融合的极化SAR图像地物分类[J]. 系统工程与电子技术, 2018, 40(2): 295–302. doi: 10.3969/j.issn.1001-506X.2018.02.09

    ZHANG Yue, ZOU Huanxin, SHAO Ningyuan, et al. Terrain classification of polarimetric SAR images based on consensus similarity network fusion[J]. Systems Engineering and Electronics, 2018, 40(2): 295–302. doi: 10.3969/j.issn.1001-506X.2018.02.09
    [20] FREEMAN A and 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
    [21] CLOUDE S R and POTTIER E. A review of target decomposition theorems in radar polarimetry[J]. IEEE Transactions on Geoscience and Remote Sensing, 1996, 34(2): 498–518. doi: 10.1109/36.485127
    [22] YANG Xingwei, SZYLD D B, and JAN LATECKI L. Diffusion on a tensor product graph for semi-supervised learning and interactive image segmentation[J]. Advances in Imaging and Electron Physics, 2011, 169: 147–172. doi: 10.1016/B978-0-12-385981-5.00004-5
    [23] VAN LOAN C F. The ubiquitous Kronecker product[J]. Journal of Computational and Applied Mathematics, 2000, 123(1/2): 85–100.
    [24] QIN Xianxiang, ZOU Huanxin, ZHOU Shilin, et al. Simulation of spatially correlated PolSAR images using inverse transform method[J]. Journal of Applied Remote Sensing, 2015, 9(1): 095082. doi: 10.1117/1.JRS.9.095082
    [25] HOU Biao, WU Qian, WEN Zaidao, et al. Robust semisupervised classification for PolSAR image with noisy labels[J]. IEEE Transactions on Geoscience and Remote Sensing, 2017, 55(11): 6440–6455. doi: 10.1109/TGRS.2017.2728186
    [26] SONG Wanying, LI Ming, ZHANG Peng, et al. Unsupervised PolSAR image classification and segmentation using Dirichlet process mixture model and Markov random fields with similarity measure[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2017, 10(8): 3556–3568. doi: 10.1109/JSTARS.2017.2684301
    [27] VASILE G, TROUVÉ E, LEE J S, et al. Intensity-driven adaptive-neighborhood technique for polarimetric and interferometric SAR parameters estimation[J]. IEEE Transactions on Geoscience and Remote Sensing, 2006, 44(6): 1609–1621. doi: 10.1109/TGRS.2005.864142
  • 加载中
图(10) / 表(5)
计量
  • 文章访问数:  1258
  • HTML全文浏览量:  357
  • PDF下载量:  154
  • 被引次数: 0
出版历程
  • 收稿日期:  2019-05-14
  • 修回日期:  2019-07-19
  • 网络出版日期:  2019-07-25
  • 刊出日期:  2019-08-01

目录

    /

    返回文章
    返回