王云艳,罗冷坤,王重阳

• 1:湖北工业大学电气与电子工程学院

摘要(Abstract)：

针对传统极化合成孔径雷达图像分类中特征提取不完整、特征表征性不够强、分类中干扰杂质较多等问题,该文提出了一种基于Deeplab模型的极化合成孔径雷达图像地物分类方法。实验通过在荷兰地区数据上对田野、植被、建筑区、水域、山体5类进行分类,然后在欧洲其他区域进行了算法评价。与传统的结合条件随机场的FCN-8s特征分类模型相比,该文方法能够提取更高效的底层特征,得到更高的分类精度、Kappa系数和总体精度。该方法不仅能在山体上提高10%左右的分类精度,而且能在这5类以外的类别掺杂情况,保证模型良好的鲁棒性。

关键词(KeyWords)： 极化合成孔径雷达;Deeplab网络;空洞卷积;多孔空间金字塔

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金项目（41601394）;; 湖北工业大学博士启动基金项目（BSQD2016010）

作者(Author): 王云艳,罗冷坤,王重阳

DOI: 10.16251/j.cnki.1009-2307.2020.06.016

参考文献(References)：

• [1] LEE J A,GRUNES M R,KWORK R.Classification of multi-look polarimetric SAR imagery based on complex wishart distribution[J].International Journal of Remote Sensing,1994,15(11):2299-2311.
• [2] 普亚如.基于CNN特征学习和SVM的极化SAR影像地物分类[D].西安:西安电子科技大学,2015.(PU Yaru.Classification of polarized SAR images based on CNN feature learning and SVM[D].Xi’an:Xi’an University,2015.)
• [3] 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):17-24.(ZHANG Lamei,CHEN Zeqian,ZOU Bin.Fine classification of PolSAR images based on 3D convolutional neural network[J].Infrared and laser engineering,2008,47(7):17-24.)
• [4] 陈永生,喻玲娟,谢晓春.基于全卷积神经网络的SAR图像目标分类[J].雷达科学与技术,2018,16(3):242-248.(CHEN Yongsheng,YU Lingjuan,XIE Xiaochun.SAR image target classification based on all convolutional neural network[J].Radar Science and Technology,2018,16(3):242-248.)
• [5] 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.
• [6] 刘靖,赵逢禹.高维数据降维技术及研究进展[J].电子科技,2018,31 (3):36-38,43.(LIU Jing,ZHAO Fengyu.Research progress of dimensional reduction technology on high-dimensiondate[J].Electronic Science and Technology,2018,31 (3):36-38,43.)
• [7] 喻玲娟,王亚东,谢晓春,等.基于FCNN和ICAE的SAR图像目标识别方法[J].雷达学报,2018,7(5):622-631.(YU Lingjuan,WANG Yadong,XIE Xiaochun,et al.SAR ATR based on FCNN and ICAE[J].Journal of Radars,2018,7(5):622-631.)
• [8] WU J,LENG C,WANG Y,et al.Quantized convolutional neural networks for mobile devices[C]//IEEE Conference on Computer Vision and Pattern Recognition.Las Vegas:IEEE,2016:4820-4828.
• [9] GU F,ZHANG H,WANG C.A classification method for POLSAR images using SLIC superpixel segmentation and deep convolution neural network[C]//IEEE International Geoscience and Remote Sensing Symposium.Valencia Spain:IEEE,2018:6671-6674.
• [10] HE C,HAN G,LIU X,et al.A low-rank fully convolutional network for classification based on a multi-dimensional description primitive of time series polarimetric sar images[C]//International Geoscience and Remote Sensing Symposium(IGARSS).[S.l.]:IEEE,2017:3101-3104.
• [11] LYU N,CHEN C,QIU T,et al.Deep Learning and superpixel feature extraction based on contractive autoencoder for change detection in SAR images[J/OL].IEEE Transactions on Industrial Informatics,2018[2018-10-01].https://www.x-mol.com/paper/836483.
• [12] MULLISSA A G,PERSELLO C,TOLPEKIN V,Fully convolutional networks for multi-temporal SAR image classification[C]//IEEE International Geoscience and Remote Sensing Symposium.Valencia:IEEE,2018:6635-6638.
• [13] WANG Y,WANG C,ZHANG H.Integrating H-A-α with fully convolutional networks for fully PolSAR classification[C]//International Workshop on Remote Sensing with Intelligent Processing.Shanghai:IEEE,2017:1870-1872.
• [14] CHOLLET F.Xception:deep learning with depthwise separable convolutions[C]//IEEE Conference on Computer Vision and Pattern Recognition (CVPR).[S.l.]:IEEE,2017:1800-1807.
• [15] 李倩,裴炳南,常芳芳.基于多层卷积神经网络的SAR图像分类方法[J].探测与控制学报,2018,40(3):85-90.(LI Qian,PEI Bingnan,CHANG Fangfang.SAR image classification method based on multi-layer convolutional neural network[J].Journal of Detection and Control,2008,40(3):85-90.)
• [16] 汤浩,何楚.全卷积网络结合改进的条件随机场-循环神经网络用于SAR图像场景分类[J].计算机应用,2016,36(12):3436-3441.(TANG Hao,HE Chu.SAR image scene classification with fully convolutional network and modified conditional random field-recurrent neural network[J].Journal of Computer Applications,2016,36(12):3436-3441.)
• [17] 肖艳.面向对象的PolSAR影像分类研究[D].长春:吉林大学,2017.(XIAO Yan.Research on object-oriented PolSAR image classification[D].Changchun:Jilin University,2017.)
• [18] WANG Y,HE C,LIU X L,et al.A hierarchical fully convolutional network integrated with sparse and low-rank subspace representations for PolSAR imagery classification[J].Remote Sensing,2018,10(2):342-347.
• [19] HE C,LIU X L,FENG D,et al.Hierarchical terrain classification based on multilayer bayesian network and conditional random field[J].Remote Sensing,2017,9(1):96-104.
• [20] CUé LA ROSA L E,HAPP P N,FEITOSA Q R.Dense fully convolutional networks for crop recognition from multitemporal SAR image sequences[C]//IEEE International Geoscience and Remote Sensing Symposium.Valencia,Spain:,IEEE,2018:7460-7463.