Hyperspectral Image Segmentation Using a New Bayesian Approach With Active Learning

被引:357
作者
Li, Jun [1 ]
Bioucas-Dias, Jose M. [2 ]
Plaza, Antonio [1 ]
机构
[1] Univ Extremadura, Hyperspectral Comp Lab HyperComp, Dept Technol Comp & Commun, E-10071 Caceres, Spain
[2] Univ Tecn Lisboa, Inst Telecomunicacoes, Inst Super Tecn, P-1049001 Lisbon, Portugal
来源
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING | 2011年 / 49卷 / 10期
关键词
Active learning; graph cuts; hyperspectral image segmentation; ill-posed problems; integer optimization; mutual information (MI); sparse multinomial logistic regression (MLR); MULTINOMIAL LOGISTIC-REGRESSION; SEMISUPERVISED CLASSIFICATION; SVM; ALGORITHMS;
D O I
10.1109/TGRS.2011.2128330
中图分类号
P3 [地球物理学]; P59 [地球化学];
学科分类号
0708 ; 070902 ;
摘要
This paper introduces a new supervised Bayesian approach to hyperspectral image segmentation with active learning, which consists of two main steps. First, we use a multinomial logistic regression (MLR) model to learn the class posterior probability distributions. This is done by using a recently introduced logistic regression via splitting and augmented Lagrangian algorithm. Second, we use the information acquired in the previous step to segment the hyperspectral image using a multilevel logistic prior that encodes the spatial information. In order to reduce the cost of acquiring large training sets, active learning is performed based on the MLR posterior probabilities. Another contribution of this paper is the introduction of a new active sampling approach, called modified breaking ties, which is able to provide an unbiased sampling. Furthermore, we have implemented our proposed method in an efficient way. For instance, in order to obtain the time-consuming maximum a posteriori segmentation, we use the a-expansion min-cut-based integer optimization algorithm. The state-of-the-art performance of the proposed approach is illustrated using both simulated and real hyperspectral data sets in a number of experimental comparisons with recently introduced hyperspectral image analysis methods.
引用
收藏
页码:3947 / 3960
页数:14
相关论文
共 53 条
[1]   Fast Image Recovery Using Variable Splitting and Constrained Optimization [J].
Afonso, Manya V. ;
Bioucas-Dias, Jose M. ;
Figueiredo, Mario A. T. .
IEEE TRANSACTIONS ON IMAGE PROCESSING, 2010, 19 (09) :2345-2356
[2]  
[Anonymous], 2003, WILEY HOBOKEN
[3]  
[Anonymous], 1973, Pattern Classification and Scene Analysis
[4]  
[Anonymous], MIT PRESS SERIES
[5]  
[Anonymous], 1995, Markov Random Field Modeling in Computer Vision
[6]  
Bagon S., 2006, MATLAB WRAPPER GRAPH
[7]  
BESAG J, 1974, J ROY STAT SOC B MET, V36, P192
[8]  
Bioucas-Dias J., 2009, Logistic regression via variable splitting and augmented lagrangian tools
[9]   Hyperspectral subspace identification [J].
Bioucas-Dias, Jose M. ;
Nascimento, Jose M. P. .
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2008, 46 (08) :2435-2445
[10]  
Bishop C.M., 2006, J ELECTRON IMAGING, V16, P049901, DOI DOI 10.1117/1.2819119