基于空间邻域信息的高光谱遥感影像半监督协同训练

doi:10.11873/j.issn.1004-0323.2016.6.1122

遥感技术与应用

2016, Vol. 31

Issue (6): 1122-1130 DOI: 10.11873/j.issn.1004-0323.2016.6.1122

数据与图像处理

基于空间邻域信息的高光谱遥感影像半监督协同训练

朱济帅¹,尹作霞²,谭琨¹,王雪¹,李二珠³,杜培军³

(1.中国矿业大学江苏省资源环境信息工程重点实验室,江苏徐州 221116;
2.济南市城市规划咨询服务中心,山东济南 250099;
3.卫星测绘技术与应用国家测绘地理信息局重点实验室,南京大学,江苏南京 210023)

Semi-supervised Tri_training Hyperspectral Image Classification Approach based on Spatial Neighborhood Information

Zhu Jishuai¹,Yin Zuoxia²,Tan Kun¹,Wang Xue¹,Li Erzhu²,Du Peijun³

(1.Jiangsu Key Laboratory of Resources and Environment Information Engineering,
China University of Mining and Technology,Xuzhou 221116,China;
2.Jinan City Planning and Consulting Service Center,Jinan 250099,China;
3.Key laboratory for Satellite Mapping Technology and Applications of State Administration
of Surveying,Mapping and Geoinformation of China,Nanjing 210003,China)

全文: PDF(40877 KB)

摘要：

针对tri_training协同训练算法在小样本的高光谱遥感影像半监督分类过程中,存在增选样本的误标记问题,提出一种基于空间邻域信息的半监督协同训练分类算法tri_training_SNI(tri_training based on Spatial Neighborhood Information)。首先利用分类器度量方法不一致度量和新提出的不一致精度度量从MLR(Multinomial Logistic Regression)、KNN(k\|Nearest Neighbor)、ELM(Extreme Learning Machine)和RF(Random Forest)4个分类器中选择3分类性能差异性最大的3个分类器;然后在样本选择过程中,采用选择出来的3个分类器,在两个分类器分类结果相同的基础上,加入初始训练样本的8邻域信息进行未标记样本的二次筛选和标签的确定,提高了半监督学习的样本选择精度。通过对AVIRIS和ROSIS两景高光谱遥感影像进行分类实验,结果表明与传统的tri_training协同算法相比,该算法在分类精度方面有明显提高。

关键词： 空间邻域信息(SNI); 协同训练; 半监督; 高光谱遥感影像分类

Abstract:

In the process of hyperspectral image classification using the tri_training algorithm,the labels of unlabeled samples have error labels when the amount of initial training samples is small.In this paper,we propose a novel tri_training based on spatial neighborhood information(tri_training_SNI) to solve the problem for the tri_training algorithm.Firstly,we choose three basic classifiers from MLR(Multinomial Logistic Regression),KNN(k\|Nearest Neighbor),ELM(Extreme Learning Machine) and RF(Random Forest) classifier based on disagreement measure and disagreement\|accuracy.These classifiers are redefined using unlabeled samples in the tri_training_SNI process.In detail,in each round of tri_training_SNI,unlabeled samples are labeled for a classifier by the following two steps.Step 1:the first selection of unlabeled samples is constructed under certain conditions that the other two classifiers have the same labels.Step 2:spatial Neighborhood Information of initial training samples based on 8\|neighborhood is applied in this proposed approach to construct the secondary selection of unlabeled samples and the labels of unlabeled samples.Then the final classification results are produced via majority voting by the classification results of three classifiers.Experiments on two real hyperspectral data indicate that the proposed approach can effectively improve classification performance.

Key words: Spatial Neighborhood Information(SNI);Tri_training;Semi\ supervised;Hyperspectral remote sensing image classification

收稿日期: 2015-12-02 出版日期: 2016-12-30

TP 391.4

基金资助:

国家自然科学基金项目(41471356),卫星测绘技术与应用测绘地理信息局重点实验室项目(KLAMTA-201410)。

通讯作者: 谭琨（1981-），男，湖南衡阳人,博士，教授，主要从事高光谱遥感研究。Email:tankuncu@gmail.com。

作者简介: 朱济帅(1991-),男,河南新乡人,硕士研究生,主要从事模式识别研究。Email:js_zhucumt@163.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
	朱济帅
	尹作霞
	谭琨
	王雪
	李二珠
	杜培军

引用本文:

朱济帅,尹作霞,谭琨,王雪,李二珠,杜培军. 基于空间邻域信息的高光谱遥感影像半监督协同训练[J]. 遥感技术与应用, 2016, 31(6): 1122-1130.

Zhu Jishuai,Yin Zuoxia,Tan Kun,Wang Xue,Li Erzhu,Du Peijun. Semi-supervised Tri_training Hyperspectral Image Classification Approach based on Spatial Neighborhood Information. Remote Sensing Technology and Application, 2016, 31(6): 1122-1130.

链接本文:

http://www.rsta.ac.cn/CN/10.11873/j.issn.1004-0323.2016.6.1122 或 http://www.rsta.ac.cn/CN/Y2016/V31/I6/1122

［1］Landgrebe D.Hyperspectral Image Data Analysis［J］.IEEE Signal Processing Magazine,2002,19(1):1728.

［2］TanKun,Du Peijun.Hyperspectral Remote Sensing Image Classification based on Support Vector Meachine［J］,Journal of Infrared and Millimeter Waves,2008,27(2):123128.［谭琨,杜培军.基于支持向量机的高光谱遥感图像分类［J］.红外与毫米波学报,2008,28(2):123128.］

［3］Bazi Y,Alajlan N,Melgani F,et al.Differential Evolution Extreme Learning Machine for the Classification of Hyperspectral Images［J］.IEEE Geoscience and Remote Sensing Letters,2014,11(6):10661070.

［4］Pal M.Random Forest Classifier for Remote Sensing Classification［J］.International Journal of Remote Sensing,2005,26(1):217222.

［5］Zhang Huawei,Wang Mingwen,Gan Lixin.Automatic Text Classification Model based on Random Forest［J］.山东大学学报(理学版),2006,41(3):59.［张华伟,王明文,甘丽新.基于随机森林的文本分类模型研究［J］.山东大学学报(理学版),2006,41(3):59.］

［6］Li J,BioucasDias J M,Plaza A.Spectralspatial Hyperspectral Image Segmentation Using Subspace Multinomial Logistic Regression and Markov Random Fields［J］.IEEE Transactions on Geoscience and Remote Sensing,2012,50(3):809823.

［7］Zhu X,Goldberg A B.Introduction to Semisupervised Learning［J］.Synthesis Lectures on Artificial Intelligence and Machine Learning,2009,3(1):1130.

［8］Zhang Lei,Shao Zhenfeng,Zhou Xiran,et al.Semisupervised Collaborative Classification for Hyperspectral Remote Sensing Image with Combination of Cluster Feature and SVM［J］.Acta Geodaetica et Cartographica Sinica,2014,43(8):855861.［张磊,邵振峰,周熙然,等.聚类特征和SVM 组合的高光谱影像半监督协同分类［J］.测绘学报,2014,43(8):855861.］［9］Tan K,Zhou S,Du Q.Semisupervised Discriminant Analysis for Hyperspectral Imagery With Blocksparse Graph［J］.IEEE Geoscience and Remote Sensing Letters,2015,12(8):15.

［10］Tuia D,Ratle F,Pacifici F,et al.Active Learning Methods for Remote Sensing Image Classification［J］.IEEE Transactions on Geoscience and Remote Sensing,2009,48(7):22182232.

［11］Huang R,He W.Using Tritraining to Exploit Spectral and Spatial Information for Hyperspectral Data Classification［C］//International Conference on IEEE Computer Vision in Remote Sensing (CVRS),2012:3033.

［12］Tan K,Li E,Du Q,et al.An Efficient SemiSupervised Classification Approach for Hyperspectral Imagery［J］.ISPRS Journal of Photogrammetry & Remote Sensing,2014,97:3645.

［13］Blum A,Mitchell T.Combining Labeled and Unlabeled Data with Cotraining［C］//Proceedings of the Eleventh Annual Conference on Computational Learning Theory.ACM,Madison,Wisconsin,1998:92100.

［14］Goldman S A,Zhou Y.Enhancing Supervised Learning with Unlabeled Data［C］//Proceedings of the Seventeenth International Conference on Machine Learning.Morgan Kaufmann Publishers Inc.,Palo Alto,California,2000:327334.

［15］Zhou Y,Goldman S.Democratic Colearning.Tools with Artificial Intelligence［C］//Proceedings of the 16th IEEE International Conference on Tools with Artificial Intelligence.Boca Raton,Florida,2004:594202.

［16］Zhou Z H,Li M.Tritraining:Exploiting Unlabeled Data using Three Classifiers［J］.IEEE Transactions on Knowledge and Data Engineering,2005,17(11):15291541.

［17］Kuncheva L I,Whitaker C J.Ten Measures of Diversity in Classifier Ensembles:Limits for Two Classifiers［J］.Proc of Iee Workshop on Intelligent Sensor Processing,2001,(50):10/11010.

［18］Li Shijin,Tao Jian,Wan Dingsheng,et al.Contentbased Remote Sensing Image Retrieval Using Cotraining of Multiple Classifiers［J］.Journal of Remote Sensing,2010,14(3):493506.［李士进,陶剑,万定生,等.多分类器实例协同训练遥感图像检索［J］.遥感学报,2010,14(3):493506.］

［19］Dopido I,Li J,Marpu P R,et al.Semisupervised Selflearning for Hyperspectral Image Classification［J］.IEEE Transactions on Geoscience and Remote Sensing,2013,51(7):40324044.

[1]	袁春琦,徐佳,程圆娥,陈媛媛,许康. 基于协同训练与集成学习的极化SAR图像半监督分类[J]. 遥感技术与应用, 2017, 32(2): 380-385.
[2]	李婷,程博,尤淑撑. 基于半监督鉴别分析的高分辨率SAR影像建筑区提取[J]. 遥感技术与应用, 2016, 31(2): 239-246.
[3]	钱进,罗鼎. 基于成对约束半监督降维的高光谱遥感影像特征提取[J]. 遥感技术与应用, 2014, 29(4): 681-688.

Viewed

Full text

Abstract

Cited

Shared

Discussed