基于复杂度聚类的自适应遥感场景分类

doi:10.19678/j.issn.1000-3428.0056209

摘要/Abstract

摘要： 遥感图像场景分类任务较普通图像分类任务的特征范围更广且分布更复杂，难以实现精准分类。针对遥感图像特征分布与神经网络结构存在一定适应性关系的情况，提出一种利用复杂度适配聚类的自适应神经网络遥感场景分类模型。构建含有颜色矩、灰度共生矩阵、信息熵、信息增益、线占比等多重特征的遥感图像复杂度评价矩阵，通过计算图像相似性得到不同复杂度的图像子集，采用层次聚类方式将图像复杂度分为高、中、低等级，并分别使用DenseNet、CapsNet和SENet神经网络对复杂度适配的图像子集进行训练，最终获得自适应遥感场景分类模型。实验结果表明，与DenseNet、CapsNet、SENet等模型相比，该模型能更有针对性地提取不同复杂度的图像特征，具有更高的遥感场景分类准确率。

关键词: 遥感图像, 场景分类, 图像复杂度, 自适应神经网络, 深度学习

Abstract: Compared with common image classification tasks,the classification of remote sensing images has a wider feature range and more complex distribution,which makes it difficult to achieve accurate classification.In view of the adaptive relationship between the feature distribution of remote sensing images and the structure of neural network,this paper proposes a remote sensing scene classification model using adaptive neural network based on complexity-adaptive clustering.The complexity evaluation matrix of remote sensing images is constructed,which includes multiple features including color moment,gray level co-occurrence matrix,information entropy,information gain and line ratio.Image subsets with different complexity degrees are obtained by calculating image similarity.The image complexity is divided into high,medium and low levels by using hierarchical clustering.Then the complexity-adaptive image subsets are trained by using DenseNet,CapsNet and SENet to obtain the adaptive remote sensing scene classification model.Experimental results show that compared with DenseNet,CapsNet,SENet and other models,this model has better performance in extracting image features with different complexity degrees,and has higher accuracy of remote sensing scene classification.

Key words: remote sensing image, scene classification, image complexity, adaptive neural network, deep learning

中图分类号:

TP751

梁文韬, 康雁, 李浩, 李晋源, 宁浩宇. 基于复杂度聚类的自适应遥感场景分类[J]. 计算机工程, 2020, 46(12): 254-261,269.

LIANG Wentao, KANG Yan, LI Hao, LI Jinyuan, NING Haoyu. Adaptive Remote Sensing Scene Classification Based on Complexity Clustering[J]. Computer Engineering, 2020, 46(12): 254-261,269.

https://www.ecice06.com/CN/Y2020/V46/I12/254

图/表 12

20201216142558

20201216142601

20201216142605

20201216142609

20201216142613

20201216142616

20201216142620

20201216142623

20201216142632

20201216142636

20201216142640

20201216142643

参考文献

[1] LI Erzhu,XIA Junshi,DU Peijun,et al.Integrating multilayer features of convolutional neural networks for remote sensing scene classification[J].IEEE Transactions on Geoscience and Remote Sensing,2017,55(10):5653-5665.
[2] CHENG Gong,HAN Junwei,LU Xiaoqiang.Remote sensing image scene classification:benchmark and state of the art[EB/OL][2019-08-23].https://www.research gate.net/publication/314152791_Remote_Sensing_Image_Scene_Classification_Benchmark_and_State_of_the_Art.
[3] ZOU Qin,NI Lihao,ZHANG Tong,et al.Deep learning based feature selection for remote sensing scene classification[J].IEEE Geoscience and Remote Sensing Letters,2015,12(11):2321-2325.
[4] 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.
[5] YE Lihua,WANG Lei,SUN Yaxin,et al.Parallel multi-stage features fusion of deep convolutional neural networks for aerial scene classification[J].Remote Sensing Letters,2018,9(3):294-303.
[6] CAO Linlin,LI Haitao,HAN Yanshun,et al.Application of convolutional neural network in high-scoring remote sensing image classification[J].Surveying and Mapping Science,2016,41(9):170-175.(in Chinese) 曹林林,李海涛,韩颜顺,等.卷积神经网络在高分遥感影像分类中的应用[J].测绘科学,2016,41(9):170-175.
[7] HUANG Hong,HE Kai,ZHENG Xinlei,et al.Spatial-spectral joint feature extraction of hyperspectral images based on deep learning[J].Laser and Optoelectronics Progress,2017,54(10):174-182.(in Chinese) 黄鸿,何凯,郑新磊,等.基于深度学习的高光谱图像空-谱联合特征提取[J].激光与光电子学进展,2017,54(10):174-182.
[8] TANG Xiaoqing,LIU Yazhou,CHEN Junlong.Improved classification method of remote sensing image based on sparse representation[J].Computer Engineering,2016,42(3):254-258.(in Chinese) 唐晓晴,刘亚洲,陈骏龙.基于稀疏表示的遥感图像分类方法改进[J].计算机工程,2016,42(3):254-258.
[9] LIU Xiaodan,YUE Shuang.Forest vegetation texture segmentation of remote sensing image based on visual attention[J].Computer Engineering,2018,44(4):274-280.(in Chinese) 刘小丹,岳爽.基于视觉注意的遥感图像森林植被纹理分割[J].计算机工程,2018,44(4):274-280.
[10] LI Haifeng,PENG Jian,TAO Chao,et al.What do we learn by semantic scene understanding for remote sensing imagery in CNN framework?[EB/OL].[2019-08-23].https://www.researchgate.net/publication/317040292_What_do_We_Learn_by_Semantic_Scene_Understanding_for_Remote_Sensing_imagery_in_CNN_framework.
[11] SONG Xibin,DAI Yuchao,QIN Xueying.Deep depth super-resolution:learning depth super-resolution using deep convolutional neural network[C]//Proceedings of ACCV’16.Berlin,Germany:Springer,2017:360-376.
[12] TSAMPIKOS K.Depth-adaptive methodologies for 3D image caregorization[EB/OL].[2019-08-23].https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.669124.
[13] RAGHU M,POOLE B,KLEINBERG J,et al.On the expressive power of deep neural networks[EB/OL].[2019-08-23].https://www.researchgate.net/publication/304018151_On_the_expressive_power_of_deep_neural_networks.
[14] XIE S N,GIRSHICK R,DOLLAR P,et al.Aggregated residual transformations for deep neural networks[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2017:21-28.
[15] LI L L,WANG L G,TEIXEIRA F L,et al.DeepNIS:deep neural network for nonlinear electromagnetic inverse scattering[J].IEEE Transactions on Antennas and Propagation,2019,67(3):1819-1825.
[16] LIU Yi.Review of infrared image complexity evaluation method[EB/OL].[2019-08-23].http://en.cnki.com.cn/Article_en/CJFDTotal-HKBQ201403017.htm.
[17] GAO Zhenyu,YANG Xiaomei,GONG Jianming,et al.Research on image complexity description methods[EB/OL].[2019-08-23].http://en.cnki.com.cn/Article_en/CJFDTotal-ZGTB201001023.htm.
[18] CIOCCA G,CORCHS S,GASPARINI F,et al.Does color influence image complexity perception?[C]//Proceedings of CCIW’15.New York,USA:ACM Press,2015:139-148.
[19] SONG Qinghuan,CHEN Zhongbi,SUN Shaohua,et al.A scene recognition method based on image complexity[C]//Proceedings of International Symposium on Advanced Optical Manufacturing and Testing Technologies.Washington D.C.,USA:IEEE Press,2014:20-26.
[20] CHEN Yanqin,DUAN Jin,ZHU Yong,et al.Research on the image complexity based on neural network[C]//Proceedings of 2015 International Conference on Machine Learning and Cybernetics.Washington D.C.,USA:IEEE Press,2015:33-38.
[21] MURAMATSU C.Overview on subjective similarity of images for content-based medical image retrieval[J].Radiological Physics and Technology,2018,11(2):109-124.
[22] SUCIATI N,HERUMURTI D,WIJAYA A Y.Feature extraction using gray-level co-occurrence matrix of wavelet coefficients and texture matching for batik motif recognition[EB/OL].[2019-08-23].https://www.researchgate.net/publication/313471714_Feature_extraction_using_gray-level_co-occurrence_matrix_of_wavelet_coefficients_and_texture_matching_for_batik_motif_recognition.
[23] CHEN Yanqin,DUAN Jin,ZHU Yong,et al.Research on the image complexity based on texture features[EB/OL].[2019-08-23].https://www.researchgate.net/publication/282458233_Research_on_the_image_complexity_based_on_texture_features.
[24] HUANG G,LIU Z,VAN D M L,et al.Densely connected convolutional networks[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2017:51-57.
[25] OYAMA T,YAMANAKA T.Fully convolutional DenseNet for saliency-map prediction[C]//Proceedings of 2017 IAPR Asian Conference on Pattern Recognition.Washington D.C.,USA:IEEE Press,2017:25-32.
[26] SABOUR S,FROSST N,HINTON G E.Dynamic routing between capsules[EB/OL].[2019-08-23].https://www.researchgate.net/publication/320627198_Dynamic_Routing_Between_Capsules.
[27] LI X Y,KIRINGA I,YEAP T,et al.Exploring deep anomaly detection methods based on capsule net[EB/OL].[2019-08-23].https://www.researchgate.net/publication/334478921_Exploring_Deep_Anomaly_Detection_Methods_Based_on_Capsule_Net.
[28] HU J,SHEN L,ALBANIE S,et al.Squeeze-and-excitation networks[C]//Proceedings of 2018 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2018:42-49.
[29] AN Gaoyun,ZHOU Wen,WU Yuxuan,et al.Squeeze-and-excitation on spatial and temporal deep feature space for action recognition[C]//Proceedings of IEEE International Conference on Signal Processing.Washington D.C.,USA:IEEE Press,2018:52-27.
[30] LIU Yanfei,ZHONG Yanfei,FEI Feng,et al.Scene semantic classification based on random-scale stretched convolutional neural network for high-spatial resolution remote sensing imagery[C]//Proceedings of 2016 IEEE International Geoscience and Remote Sensing Symposium.Washington D.C.,USA:IEEE Press,2016:23-29.
[31] YU Yunlong,LIU Fuxian.A two-stream deep fusion framework for high-resolution aerial scene classification[EB/OL].[2019-08-23].https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5822919/.

选择文件类型/文献管理软件名称

选择包含的内容