Please wait a minute...
img

官方微信

遥感技术与应用  2019, Vol. 34 Issue (4): 712-719    DOI: 10.11873/j.issn.1004-0323.2019.4.0712
CNN 专栏     
基于频带特征融合的GL-CNN遥感图像场景分类
崔先亮(),陈立福(),邢学敏,袁志辉
长沙理工大学电气与信息工程学院,湖南 长沙 410114
Remote Sensing Image Scene Classification based on Frequency Band Feature Fusion and GL-CNN
Xianliang Cui(),Lifu Chen(),Xuemin Xing,Zhihui Yuan
School of Electrical and Information Engineering, Changsha University of Science and Technology, Changsha, 410114,China
 全文: PDF(9133 KB)   HTML
摘要:

高分辨率卫星遥感图像场景信息的分类对影像分析和解译具有重要意义,传统的高分辨卫星遥感图像场景分类方法主要依赖于人工提取的中、低层特征且不能很好的利用图像丰富的场景信息,针对这一问题,提出一种基于频带特征融合与GL-CNN(Guided Learning Convolutional Neural Network,指导学习卷积神经网络)的分类方法。首先通过NSWT(Non-Subsampled Wavelet Transform,非下采样小波变换)提取出图像的高低频子带,将高频子带进行频带特征融合得到融合高频子带,然后联合频谱角向能量分布曲线的平稳区间分析实现融合高频子带与低频子带的样本融合,最后指导卷积神经网络自动提取图像的高低频子带包含的高层特征来实现场景分类。通过对UCM_LandUse 21类数据进行试验表明,本文方法的分类正确率达到94.52%,相比以往算法有显著提高。

关键词: 非下采样小波变换频带特征融合指导学习样本融合场景分类    
Abstract:

The classification of high-resolution satellite remote sensing image scene information is of great significance for image analysis and interpretation. The traditional high-resolution satellite remote sensing image scene classification method mainly relies on the artificially extracted middle and low-level features and can not make good use of image-rich scenes. In response to this problem, a classification method based on band feature fusion and GL-CNN (Guided Learning Convolutional Neural Network) is proposed. Firstly, the high-low frequency sub-band of the image is extracted by NSWT (Non-Subsampled Wavelet Transform), and then the high-frequency sub-band is fused to obtain the fused high-frequency sub-band, and then the angular energy distribution curve is combined. The stationary interval analysis realizes the fusion of the fusion high-frequency sub-band and the low-frequency sub-band, and finally guides the convolutional neural network to automatically extract the high-level features contained in the high-low frequency sub-band of the image to realize the scene classification. Experiments on UCM_LandUse 21 data show that the classification accuracy of this method reaches 94.52%, which is significantly improved compared with previous algorithms.

Key words: Non-subsampled wavelet transform    Band feature fusion    Guided learning    Sample fusion    Scene classification
收稿日期: 2018-05-23 出版日期: 2019-10-16
ZTFLH:  TP753  
基金资助: 国家自然科学基金青年基金项目“基于机载双天线InSAR系统三维地形实时获取算法研究”(41201468);湖南省教育厅项目“高分辨率SAR图像复杂背景下高精度鲁棒的道路提取算法研究”(16B004);国家自然科学基金青年基金项目“顾及流变参数的时序InSAR公路形变模型研究”(41701536);国家自然科学基金青年基金项目“面向复杂地形的多通道干涉SAR高精度DEM稳健反演研究”(61701047)
通讯作者: 陈立福     E-mail: 421428120@qq.com;lifu_chen@139.com
作者简介: 崔先亮(1994-),男,湖南常德人,硕士研究生,主要从事图像处理、深度学习、遥感图像解译研究。E?mail:421428120@qq.com
服务  
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章  
崔先亮
陈立福
邢学敏
袁志辉

引用本文:

崔先亮,陈立福,邢学敏,袁志辉. 基于频带特征融合的GL-CNN遥感图像场景分类[J]. 遥感技术与应用, 2019, 34(4): 712-719.

Xianliang Cui,Lifu Chen,Xuemin Xing,Zhihui Yuan. Remote Sensing Image Scene Classification based on Frequency Band Feature Fusion and GL-CNN. Remote Sensing Technology and Application, 2019, 34(4): 712-719.

链接本文:

http://www.rsta.ac.cn/CN/10.11873/j.issn.1004-0323.2019.4.0712        http://www.rsta.ac.cn/CN/Y2019/V34/I4/712

图1  NSWT分解结果图
图2  PCA图像融合流程图
图3  高频子带特征融合效果图
图4  本文构建的卷积神经网络结构及参数图
图5  本文算法整体流程图
图6  21类样本分类正确率统计
图7  3类样本频谱与角向能量分布曲线示例
图8  样本融合效果示例
图9  样本融合后的分类混淆矩阵
原图像未融合频带融合
正确率/%87.478.692.6
表1  融合前后高频子带分类正确率比较

类别

子类

融合高频子带
灌木丛密集住宅移动住房停车场中密住宅
正确率/%1009510010095
平均值/%98

类别

子类

低频子带
高速公路十字路口天桥跑道稀疏住宅
正确率/%95909510095
平均值/%95
表2  高低频分量显著的10类样本分类正确率分析
分类算法分类正确率/%
MNCC[18]88.26
ConvNET[19]89.79
SVM[20]78.57
CCM-BOVW[21]86.64
MS-DCNN[10]91.34
PCA-CNN[1]92.86
本文算法94.52
表3  不同算法的分类正确率对比
1 HeXiaofei,ZouZhengrong,TaoChao,et al.Combined Saliency with Multi-Convolutional Neural Network for High Resolution Remote Sensing Scene Classification[J].Acta Geodaetica et Cartographica Sinica,2016,45(9):1073-1080.
1 何小飞,邹峥嵘,陶超,等.联合显著性和多层卷积神经网络的高分影像场景分类[J].测绘学报,2016,45(9):1073-1080.
2 FanMin,HanQi,WangFen,et al. Scene Image Categorization Algorithm based on Multi-level Features Representation[J].Journal of Jilin University(Engineering and Technology Edition),2017,46(6):1909-1915.
2 范敏,韩琪,王芬,等.基于多层次特征表示的场景图像分类算法[J].吉林大学学报(工学版),2017,46(6):1909-1915.
3 LiuDawei,HanLing,HanXiaoyong.High Spatial Resolution Rmote Sensing Image Classification based on Deep Learning[J].Acta Optica Sinica,2016,36(4):0428001.
3 刘大伟,韩玲,韩晓勇.基于深度学习的高分辨率遥感影像分类究[J].光学学报,2016,36(4):0428001.
4 OlivaA,TorralbaA. Modeling the Shape of the Scene:A Holistic Representation of the Spatial Envelope[J].International Journal of Computer Vision,2001,42(3):145-175.
5 SivicJ,ZissermanA.Video Google:A Text Retrieval Approach to Object Matching in Videos[C]//Proceedingsof the Ninth IEEE International Conference on Computer Vision.Nice,2003,2:1470-1477.
6 XiaoX Z,DevisTuia , LiC M,et al.Deep Learning in Remote Sensing:A Review[J].IEEE Geoscience and Remote Sensing Magazine,2017:1-60.
7 CastelluccioM, PoggiG, SansoneC,et al.Verdoliva.Landuse Classification in Remote Sensing Images by Convolutional Neuralnetworks[J]. arXiv:1508.00092,2015.
8 NogueiraK, PenattiO, SantosJ. Towards Better Exploiting Convolutional Neural Networks for Remote Sensing Scene Classification[J]. Pattern Recognition,2016,61:539-556.
9 LuusF, SalmonB,BerghF,et al. Multiview Deep Learning for Land-use Classification[J].IEEE Geoscience and Remote Sensing Letters, 2015,12(12):2448-2452.
10 XuSuhui,MuXiaodong,ZhaoPeng,et al.Scene Classification of Remote Sensing Image based on Multi-scale Feature and Deep NeuralNetwork[J].Acta Geodaetica et Cartographica Sinica,2016,45(7):834-840.
10 许夙晖,慕晓东,赵鹏,等.利用多尺度特征与深度网络对遥感影像进行场景分类[J].测绘学报,2016,45(7):834-840.
11 YaermaimaitiYilihamu,XieLirong,KongJun.Remote Sensing Image Fusion based on PCA Transform and Wavelet Transform[J].Infrared and Laser Engineering, 2014,43(7):2335-2340.
11 伊力哈木·亚尔买买提,谢丽蓉,孔军.基于PCA变换与小波变换的遥感图像融合方法[J].红外与激光工程,2014,43(7):2335-2340.
12 JiFeng,LiZeren,ChangXia,et al.Remote Sensing Image Fusion Method based on PCA and NSCT Transform[J].Journal of Graphics, 2017,38(2):247-252.
12 纪峰,李泽仁,常霞等.基于PCA和NSCT变换的遥感图像融合方法[J].图学学报,2017,38(2):247-252.
13 BengioY,CourvilleA,VincentP.Representation Learning:A Review and New Perspectives[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2013,35(8):1798-1828.
14 ZhaoJuanping,GuoWeiwei,LiuBin,et al.Convolutional Neural Network-based SAR Image Classification with Noisy Labels[J].Journal of Radars, 2017,6(5):514-523.
14 赵娟萍,郭炜炜,柳彬,等.基于概率转移卷积神经网络的含噪标记SAR图像分类[J].雷达学报,2017,6(5):514-523.
15 WangSiyu,GaoXin,SunHao,et al.An Aircraft Detection Method based on Convolutional Neural Networks in High-Resolution SAR Images[J] Journal of Radars, 2017,6(2):195-203.
15 王思雨,高鑫,孙皓,等.基于卷积神经网络的高分辨率SAR图像飞机目标检测方法[J].雷达学报,2017,6(2):195-203.
16 XuFeng, WangHaipeng,JinYaqiu.Deep Learning as Applied in SAR Target Recognition and Terrain Classification[J]. Journal of Radars, 2017,6(2):136-148.
16 徐丰,王海鹏,金亚秋.深度学习在SAR目标识别与地物分类中的应用[J].雷达学报,2017,6(2):136-148.
17 YangY, NewsamS. Bag-of-visual-words and Spatial Exensions for Land-use Classification[C]⫽ACM SIGSPATIAL Internati-onal Conference on Advances in Geographic Information Systems (ACM GIS), 2010.
18 LiuY, FuZ Y,ZhengF B. Scene Classification of High Resolution Remote Sensing Image based on Multimedia Neural Cognitive Computing[J].Systems Engineering and Electronics,2015,37(11):2623-2633.
19 NogueiraK,MirandaW O,SantosjA D.Improving Spatial Feature Representation from Aerial Scenes by Using Convolutional Network[C] // Proceedings of the 2015 28th SIBGRAPI Conference on Graphics,Patterns and Images.Salvador:IEEE,2015:289-296.
20 ZhengX W, SunX, FuK,et al.Automatic Annotation of S-atellite Images via Multifeature Joint Sparse Coding with Spatial Relation Constraint[J].IEEE Geoscience and Remote Sensing Letters,2013,10(4):652-656.
21 ZhaoL J, TangP, HuoL Z. Land-use Scene Classificaion Using a Concentric Circle-structured Multiscale Bag-of-visual-words Mode[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing,2014,7(12):4613-4620.
No related articles found!