A Method of Semi-supervised Classification for Hyperspectral Images Based on Spatial Information and Genetic Optimization

doi:10.13474/j.cnki.11-2246.2018.0308

Abstract

Abstract: Among many dimensionality reduction methods,because of the simplicity of spectral methods,it has been widely applied in recent years.In the spectral method,the construction of graph and the selection of similarity function are the key factors affecting the dimensionality reduction effect.A semi-supervised hyperspectral image classification method is proposed based on spatial nearest neighbor information,genetic algorithm,spectral method and a small number of label sample points.The algorithm firstly constructs a new similarity function by considering the spatial information of the sample points and the class information of a small number of labeled sample points.Then K-nearest neighbor method and genetic algorithm are used to obtain the optimal graph.Based on the optimized graph,the spectral method is used to reduce the dimension of the data.Final adoption the local average pseudo-nearest neighbor method is applied to cluster analysis.In this paper,two classical hyperspectral images SalinasA data sets and Botswana data sets are tested.The results show that the proposed method can achieve high classification accuracy.

Key words: spectral clustering, genetic algorithm, spatial information, hyperspectral image, semi-supervised classification

CLC Number:

P237

HU Dongcui, XIE Fuding, YANG Jun, ZHANG Yong. A Method of Semi-supervised Classification for Hyperspectral Images Based on Spatial Information and Genetic Optimization[J]. 测绘通报, 2018, 0(10): 22-26.

References

[1] 田彦平,陶超,邹峥嵘,等.主动学习与图的半监督相结合的高光谱影像分类[J].测绘学报,2015,44(8):919-926.
[2] 丁胜,袁修孝,陈黎.粒子群优化算法用于高光谱遥感影像分类的自动波段选择[J].测绘学报,2010,39(3):257-263.
[3] SCHAEFFER S E.Survey:Graph Clustering[J].Computer Science Review,2007,1(1):27-64.
[4] CHRYSOULI C,TEFAS A.Spectral Clustering and Semi-supervised Learning Using Evolving Similarity Graphs[J].Applied Soft Computing,2015,34(C):625-637.
[5] 杨钊霞,邹峥嵘,陶超,等.空-谱信息与稀疏表示相结合的高光谱遥感影像分类[J].测绘学报,2015,44(7):775-781.
[6] LUXBURG U.A Tutorial on Spectral Clustering[J].Statistics and Computing,2007,17(4):395-416.
[7] GOU J,DU L,ZHANG Y,et al.A New Distance-weighted K-nearest Neighbor Classifier[J].Journal of Information and Computational Science,2012,9(6):1429-1436.
[8] GOU J,ZHAN Y,RAO Y,et al.Improved Pseudo Nearest Neighbor Classification[J].Knowledge-based Systems,2014,70(C):361-375.
[9] SABA F,VALADANZOU M J,MOKHTARZADE M.The Optimazation of Multi Resolution Segmentation of Remotely Sensed Data Using Genetic Alghorithm[J].ISPRS-international Archives of the Photogrammetry,Remote Sensing and Spatial Information Sciences,2013,40(1):345-349.
[10] 何芳,贾维敏,王标标,等.基于加权空-谱联合的遥感图像聚类[C]//国家安全地球物理丛书(十二)——地球物理与信息感知.鞍山:[s.n.],2016:5.
[11] BELKIN M,NIYOGI P,SINDHWANI V.Manifold Regularization:A Geometric Framework for Learning from Labeled and Unlabeled Examples[J].Journal of Machine Learning Research,2006,7(1):2399-2434.
[12] CAMPS-VALLS G,MARSHEVA T V B,Zhou D.Semi-supervised Graph-based Hyperspectral Image Classification[J].IEEE Transactions on Geoscience and Remote Sensing,2007,45(10):3044-3054.
[13] 邬文慧.空谱联合高光谱遥感图像半监督分类[D].西安:西安电子科技大学,2014.
[14] XIE J,GUO W,XIE W,et al.K-means Clustering Algorithm Based on Optimal Initial Centers Related to Pattern Distribution of Samples in Space[J].Applications Research of Computers,2012,29(3):888-892.
[15] KANNAN S R,RAMATHILAGAM S,CHUNG P C.Effective Fuzzy C-means Clustering Algorithms for Data Clustering Problems[J].Expert Systems with Applications,2012,39(7):6292-6300.
[16] 李志敏,郝盼超,黄鸿,等.半监督复合核图聚类在高光谱图像中的应用[J].光电工程,2016,44(4):33-39.