高光谱图像拥有更为丰富的特征信息，近年来被广泛应用在国防建设、资源勘测、城市开发、精准农林等领域。高光谱图像地物分类是遥感图像分类的重要研究课题，目前使用卷积神经网络解决高光谱图像分类问题是主流的研究方向，但如何精准且迅速的得到有效的分类结果，仍然是地物分类问题中的难点。深入探讨高光谱图像数据的自身特性，研究高光谱遥感图像特征提取方法和基于卷积神经网络的分类算法，并在公开的高光谱遥感数据集中对模型进行分析评估。主要成果如下：

(1)针对高光谱数据结构维度较高、样本信息丰富导致信息冗余和地物分类精度不足等问题，提出了一种融合多维度卷积神经网络的高光谱地物分类算法SENet-residual。算法分为多特征权重激励模块和多尺度宽残差网络模块两部分，多特征权重激励模块通过特征选拔、激励、分配权重的方式收集丰富的特征信息，并做特征融合。多尺度宽残差块由不同尺度的卷积操作并入残差网络中所得，多尺度网络结构通过由多个串连的多尺度残差块组合而成。最终对结果加权融合，求得的平均值用来区分高光谱地物类型。实验结果表明，利用SENet-residual融合不同尺度特征可以有效提升平均分类精度，权重激励也在一定程度改善特征提取质量。

(2)针对高光谱遥感图像空间特征与频谱特征存在差异性而导致高光谱地物分类的特征提取不合理、分类精度不稳定和训练时间长等问题，提出了基于3D密集全卷积（3D-DSFCN）的高光谱图像分类算法。算法通过密集模块中的3D卷积核分别提取光谱特征和空间特征，采用特征映射模块替换传统网络中的池化层和全连接层，最后通过softmax分类器进行分类。实验结果表明：基于3D-DSFCN的HSI分类方法提高了地物分类的准确率、增强了低频标签的分类稳定性。

Hyperspectral images have more abundant characteristic information, and have been widely used in national defense construction, resource survey, urban development, precision agriculture and forestry and other fields in recent years. Hyperspectral image classification is an important research topic of remote sensing image classification. At present, it is the mainstream research direction to use convolution neural network to solve hyperspectral image classification problem. However, how to get effective classification results accurately and quickly is still the difficulty in the classification of ground objects. In this paper, the characteristics of hyperspectral image data are deeply discussed, the feature extraction method of hyperspectral remote sensing image and the classification algorithm based on convolution neural network are studied, and the model is analyzed and evaluated in the open hyperspectral remote sensing data set. The main achievements of this paper are as follows:

(1) Aiming at the problems of high dimension of hyperspectral data structure, abundant sample information, resulting in redundant information and insufficient accuracy of ground feature classification, this paper proposes a hyperspectral ground feature classification algorithm, SENet-residual, which combines multi-dimensional convolutional neural networks. The algorithm is divided into two parts: multi-feature weight excitation module and multi-scale wide residual network module. The multi-feature weight excitation module collects rich feature information through feature selection, excitation and weight distribution, and performs feature fusion. Multi-scale wide residual blocks are obtained by merging convolution operations of different scales into the residual network, and the multi-scale network structure is formed by combining a plurality of serially connected multi-scale residual blocks. The final results are weighted and fused, and the average value obtained is used to distinguish hyperspectral ground objects. The experimental results show that using SENet-residual to fuse features of different scales can effectively improve the average classification accuracy, and the weight incentive can also improve the quality of feature extraction to some extent.

(2) To solve the problems of unreasonable feature extraction, unstable classification accuracy and long training time caused by the difference between spatial features and spectral features of hyperspectral remote sensing images, a hyperspectral image classification algorithm based on 3D dense total convolution (3D-DSFCN) is proposed. The algorithm extracts spectral features and spatial features by 3D convolution kernel in dense module, replaces the pool layer and full connection layer in traditional network by feature mapping module, and finally classifies them by softmax classifier. The experimental results show that the HSI classification method based on 3D-DSFCN improves the accuracy of ground object classification and enhances the classification stability of low-frequency tags.

[1] 张 丽,李国庆,朱岚巍,郭华东.海南省遥感大数据服务平台建设与应用示范[J].遥感学报,2019,23(02):327-335.

[2] 龚文引,蔡之华,杨鸣.智能算法在高光谱遥感数据处理中的应用[M].武汉:中国地质大学出版社,2014.

[3] 赵春晖.高光谱遥感图像处理方法及应用[M].北京:电子工业出版社, 2016.

[4] Feilong C,Wenhui G. Deep hybrid dilated residual networks for hyperspectral image classification[J].Neurocomputing,2020,384.

[5] 杨斌,王斌.光谱解混技术及其应用研究进展[J].激光与光电子学进展,2021,58(16):76-103.

[6] Ying F,Yongrong Zh,Lin Zh,Yinqiang Zh,Hua H. Simultaneous hyperspectral image super-resolution and geometric alignment with a hybrid camera system[J]. Neurocomputing,2020,384.

[7] Raffaele V,Siewert H,Dario C,Cyril R. A spatial constraint to model and extract texture components in Multivariate Curve Resolution of near-infrared hyperspectral images[J]. Analytica Chimica Acta,2020,1095.

[8] 张淳民,穆廷魁,颜廷昱,陈泽宇.高光谱遥感技术发展与展望[J].航天返回与遥感,2018,39(03):104-114.

[9] 杨润宇,贾亦雄,徐鹏,谢晓振.截断核范数和全变差正则化高光谱图像复原[J].中国图象图形学报,2019,24(10):1801-1812.

[10] 郭文慧,曹飞龙.基于 3D 多尺度特征融合残差网络的高光谱图像分类[J].模式识别与人工智能,2019,32(10):882-891.

[11] Vapnik, V.N. Statistical Learning Theory. John Wiley&Sons, New York, 1998.

[12] Melgani F., Bruzzone L.Classification of hyperspectral remote sensing images with support vector machines[J]. IEEE Transactions on Geoscience and Remote Sensing, 2004,42(8):1778-1790.

[13] Camps-Valls G, Gomez-Chova L, Calpe-Maravilla J. Robust support vector method for hyperspectral data classification and knowledge discovery[J]. IEEE Transactions on Geoscience and Remote Sensing,2004,42(7):

[14] Bruzzone L, Chi M, Marconcini M. A novel transductive SVM for semisupervised classification of remote sensing images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2006, 44(11): 3363-3373.

[15] Demir, B, Ertiirk S. Hyperspectral image classification using relevance vector machines[J].IEEE Geoscience and Remote Sensing Letters,2007,4(4):586-590.

[16] HintonG. Learning multiple layers of representation[J]. Trends in Cognitive Sciences,2007, 11(10):428.

[17] Yedidia J S, Freeman W T, Weiss Y. Constructing free energy approximations and generalized belief propagation algorithms[J]. IEEE Transaction on Information Theory,2005,(51)7: 2282-2312.

[18] Zhang M M,Li W,Du Q,Gao L R,Zhang B. Feature Extraction for Classification of Hyperspectral and LiDAR Data Using Patch-to-Patch CNN.[J]. IEEE transactions on cybernetics,2018.

[19] Hinton G, Salakhutdinov R. Reducing the dimensionality of data with neural networks[J]. Science, 2006,313(5786): 504-507.

[20] Salakhutdinov R, Hinton G. Deep Boltzmann Machines[J]. Journal of Machine Learning Research,2009,5(2):1967-2006.

[21] Chen Y, Lin Z, Zhao X, Wang G. Deep Learning-Based Classification of Hyperspectral Data[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7(6): 2094-2107.

[22] Yang J, Zhao Y, Chan C W. Hyperspectral image classification using two-channel deep convolutional neural network[C]// Geoscience and Remote Sensing Symposium.IEEE, 2016:5079-5082.

[23] Li Y, Zhang H, Shen Q. Spectral-Spatial Classification of Hyperspectral Imagery with 3D Convolutional Neural Network [J]. Remote Sensing, 2017, 9(1):67-85.

[24] Santara A, Mani K, Hatwar P, Singly A. BASS Net: Band-Adaptive Spectral-Spatial Feature Learning Neural Network for Hyperspectral Image Classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2017, 55(9):5293-5301.

[25] Zilong, Zhong, Jonathan, et al. Spectral–Spatial Residual Network for Hyperspectral Image Classification: A 3-D Deep Learning Framework[J]. IEEE Transactions on Geoscience and Remote Sensing，2018，56(2)：847 - 858.

[26] Wang W, Dou S, Jiang Z, et al. A Fast Dense Spectral–Spatial Convolution Network Framework for Hyperspectral Images Classification[J]. Remote sensing (Basel, Switzerland),2018, 10(7): 1068.

[27] 魏飞鸣,李小文,顾行发,余涛,孙源. 基于形状参数的遥感图像“同谱异物”目标区分[C]//第十四届全国图象图形学学术会议论文集.,2008:477-481.

[28] 黄琦. 基于极限随机树的遥感影像的建筑物屋顶提取[D].江西理工大学,2021.DOI:10.27176/d.cnki.gnfyc.2021.000169.

[29] 杜培军,夏俊士,薛朝辉,谭琨,苏红军,鲍蕊.高光谱遥感影像分类研究进展[J].遥感学报,2016,20(02):236-256.

[30] 高奎亮,余旭初,宋治杭,张晋,刘冰,孙一帆.联合空谱信息的高光谱影像深度胶囊网络分类[J].遥感学报,2021,25(06):1257-1269.

[31] 刘奕,李建华,张一瑫,孟涛.基于特征属性信息熵的网络异常流量检测方法[J].信息网络安全,2021,21(02):78-86.

[32] 张学工.关于统计学习理论与支持向量机[J].自动化学报,2000(01):36-46.DOI:10.16383/j.aas.2000.01.005.

[33] 王涛,李艾华,蔡艳平,王声才.基于核的学习机研究综述[J].计算机应用研究,2010,27(06):2011-2015.

[34] 张号逵,李映,姜晔楠.深度学习在高光谱图像分类领域的研究现状与展望[J].自动化学报,2018,44(06):961-977.DOI:10.16383/j.aas.2018.c170190.

[35] 冯志伟.神经网络、深度学习与自然语言处理[J].上海师范大学学报(哲学社会科学版),2021,50(02):110-122.DOI:10.13852/J.CNKI.JSHNU.2021.02.012.

[36] Wiesel T N, Hubel D H. Spatial and chromatic interactions in the lateral geniculate body of the rhesus monkey[J]. Journal of Neurophysiology, 1966, 29(6):1115-1156.

[37] 张健,保文星.生成式对抗网络的高光谱遥感图像分类方法研究[J].遥感学报,2022,26(02):416-430.

[38] 陈超,齐峰.卷积神经网络的发展及其在计算机视觉领域中的应用综述[J].计算机科学,2019,46(03):63-73.

[39] 周婷,邹俊,崔玉环,王笑宇,谢传流,夏萍.基于主成分分析和支持向量机的遥感影像水体提取方法及应用[J/OL].水资源保护:1-15[2022-04-09].http://kns.cnki.net/kcms/detail/32.1356.TV.20220407.1836.006.html

[40] 李宝芸,范玉刚,杨明莉.基于LFDA和GA-ELM的高光谱图像地物识别方法研究[J].遥感技术与应用,2021,36(03):587-593.

[41] Hu J，Shen L，Sun G．Squeeze-and-excitation networks［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Ｒecognition．［s．l．］:［s．n．］，2018:7132-7141．

[42] 于慧伶,霍镜宇,张怡卓.融合多层特征SENet和多尺度宽残差的高光谱图像地物分类[J].实验室研究与探索,2020,39(07):28-34+44.

[43] 戚银城,金超熊,赵振兵,丁洁涛,吕斌.最优知识传递宽残差网络输电线路螺栓缺陷图像分类[J].中国图象图形学报,2021,26(11):2571-2581.

[44] Li J Y, Zhang H Y, Zhang L P. Efficient superpixel-level multitask joint sparse representation for hyperspectral image classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2015, 53(10):5338-5351.

[45] Su H J, Zhao B, Du Q. Kernel collaborative representation with local correlation features for hyperspectral image classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(2): 1230-1241.