改进的OIF和SVM结合的高光谱遥感影像分类Hyperspectral remote sensing image classification based on improved OIF and SVM algorithm
张磊,邵振峰
摘要(Abstract)：
文章提出了一种结合改进的最佳指数法(OIF)和支持向量机(SVM)进行高光谱遥感影像分类新方法。利用本文提出的稳定系数进行波段初选择,根据相关系数选择波段组合生成新影像,并对新影像进行OIF计算,得到OIF值最大的波段组合为最佳波段组合;构建SVM分类器,对最佳波段组合分类;最后将分类结果与其他监督分类方法比较,并在相同核函数下与PCA和SVM结合的方法进行精度比较分析。实验结果表明,本文方法能够有效提取最佳波段组合,在SVM算法下获得较高分类精度。
关键词(KeyWords)： 高光谱;最佳指数法;支持向量机;遥感影像分类
基金项目(Foundation): 国家973计划重点项目(2010CB731801);; 国家自然科学基金项目(61172174);; 重大科技专项(2012YQ16018505,2013BAH42F03);; 国家海洋局数字海洋科学技术重点实验室开放基金资助(KLDO201307)
作者(Author): 张磊,邵振峰
DOI: 10.16251/j.cnki.1009-2307.2014.11.027
参考文献(References)：
- [1]Du Q,Yang H.Similarity-based Unsupervised Band Selection for Hyperspectral Image Analysis[J].IEEE Geoscience and Remote Sensing Letters,2008,5(4):564-568.
- [2]Du H,Qi H,Wang X,et al.Band Selection Using Independent Component Analysis for Hyperspectral Image[C]//Proceedings of 32nd Applied Imagery Pattern Recognition Workshop(AIPR03),2003,:93-98
- [3]Wang L,Gu Y,Zhang Y.Band Selection Method Based on Combination of Support Vector Machines and Subspatial Partition[J].Systems Engineering and Electronics,2005,27(6):974-977.
- [4]Chavez P S,Berlin G L,Sowers L B.Statistical Method for Selecting Landsat MSS Ratios[J].Journal of Applied Photographic Engineering,1982(8):23-30.
- [5]WU Di,HE Yong,FENG Shui juan,et al.Applications of Infrared Spectra Technique Based on LS-Support Vector Machines to the Non-destructive Measurement of Fat Content in Milk Powder[J].Journal of Infrared and Millimeter Waves,2008,27(3):180-184.
- [6]ZHAO Chun hui,Qi Bin,Zhang Yi.Hyperspectral Image Classification Based on Variational Relevance vector machine[J].ACTA OPTICA SINICA,2012,32(8):0828004.
- [7]Sun D,Li Y,Wang Q.A Unified Model for Remotely Estimating Chlorophyll a in Lake Taihu,China,Based on SVM and In Situ Hyperspectral Data[C]//IEEE Transactions on Geoscience and Remote Sensing,2009,47(8):2957-2965.
- [8]Vapnik V N.An overview of statistical learning theory[C]//IEEE Transactions on Neural Networks,1999,10(5):988-999.
- [9]Liu Xuesong,Ge Liang,Wang Bin.An Unsupervised Band Selection Algorithm for Hyperspectral Imagery Based on Maximal Information[J].Infrared Millim.Waves,2012,31(2):166-176.