基于支持向量机与图斑的高光谱分类方法研究
|
摘要
高光谱遥感作为一种新型的遥感方式在近几年的发展中已经广泛运用于军用和民,用的多个领域,然而如何从其产生的海量数据中快速而准确地挖掘出所需要的信息,目前仍然是一个待解决的问题。
简单的支持向量机只能处理二值分类问题,不能直接处理多值分类问题。而现实世界中的大部分数据都是多类数据,所以需要对简单支持向量机作进一步扩展,使之能解决多值分类问题。本文介绍了几种用于多值分类的支持向量机分类的方法,包括“一对多”、“一对一”、基于决策树以及基于有向无环图的支持向量机分类方法,并比较了它们各自的优点和缺点。通过分析各种支持向量机分类方法的不足之处,提出了一种新的支持向量机分类的方法,即将图斑理论与多值分类的支持向量机分类方法相结合。最后通过试验比较传统的支持向量机分类技术与结合图斑特性的支持向量机的分类技术,结果表明结合图斑的支持向量机在应用于高光谱影像分类问题中取得了较好的效果。
基于图斑的支持向量机方法是选择一个合适的尺度,利用光谱信息按照一定的策略将图像分割为一系列的图斑,并确保图斑内大多数像元的光谱特征相近,分别对图斑内各像元进行统计,求出各个波段的均值用该均值替换图斑内所有像元的原始亮度值。这样分类的目的是为了使各种原因产生的噪声点可以被其周围的像元同化而融合到一个图斑类别中,相当于根据其周围像元的信息对其进行了有效的恢复,不至于在分类结果上出现孤立的错分点,避免产生“椒盐现象”。试验结果表明这种方法是可行的并且分类精度和速度均较传统的支持向量机分类法有所提高。
During the recent two decades, hyperspectral remote sensing has been playing an important role in both military and civil applications. It's urgent to develop fast and accurate methods in order to discover the interested information from the huge data which were produced by hyperspectral sensors.
Simple SVM can only handle binary classification problems; can not directly handle multi-value classification. In the real world most of the data is multi-class data, so the simple SVM need for further expansion, so that it can solve the multi-value classification. This paper introduces several SVMs for multi-value classification, including "One against Rest", "One against One", Decision Tree and Directed Acyclic Graph SVMs, and compares their respective advantages and disadvantages. By analyzing the deficiencies of various SVMs, a new SVM method, namely, the theory of combining spot and SVM, is put forward. Finally, comparing the traditional SVM to SVM binding spot feature, the tests show the SVM combining of spot applied to hyperspectral image classification has achieved good results.
The principle of SVM based on spot is to choose an appropriate scale to split the image into a series of segmentation, according to certain strategy using spectral information. And this principle ensures the spectral features of the majority of patch pixel similar. This method gathers statistics of each pixel value in the spot and obtains the mean value of each band to replace the original value of all pixels in the spot. The purpose of this classification is that the pixel having the noise brought by various causes is assimilated by the surrounding pixels to merge into a single spot. In other words, under the information of its surrounding pixels recovering the value of the pixel having noise is to not appear the fault isolation in the classification map and to avoid the salt and pepper phenomenon. The results show that this method is feasible and the classification accuracy and speed is better than traditional support vector machine.
简单的支持向量机只能处理二值分类问题,不能直接处理多值分类问题。而现实世界中的大部分数据都是多类数据,所以需要对简单支持向量机作进一步扩展,使之能解决多值分类问题。本文介绍了几种用于多值分类的支持向量机分类的方法,包括“一对多”、“一对一”、基于决策树以及基于有向无环图的支持向量机分类方法,并比较了它们各自的优点和缺点。通过分析各种支持向量机分类方法的不足之处,提出了一种新的支持向量机分类的方法,即将图斑理论与多值分类的支持向量机分类方法相结合。最后通过试验比较传统的支持向量机分类技术与结合图斑特性的支持向量机的分类技术,结果表明结合图斑的支持向量机在应用于高光谱影像分类问题中取得了较好的效果。
基于图斑的支持向量机方法是选择一个合适的尺度,利用光谱信息按照一定的策略将图像分割为一系列的图斑,并确保图斑内大多数像元的光谱特征相近,分别对图斑内各像元进行统计,求出各个波段的均值用该均值替换图斑内所有像元的原始亮度值。这样分类的目的是为了使各种原因产生的噪声点可以被其周围的像元同化而融合到一个图斑类别中,相当于根据其周围像元的信息对其进行了有效的恢复,不至于在分类结果上出现孤立的错分点,避免产生“椒盐现象”。试验结果表明这种方法是可行的并且分类精度和速度均较传统的支持向量机分类法有所提高。
During the recent two decades, hyperspectral remote sensing has been playing an important role in both military and civil applications. It's urgent to develop fast and accurate methods in order to discover the interested information from the huge data which were produced by hyperspectral sensors.
Simple SVM can only handle binary classification problems; can not directly handle multi-value classification. In the real world most of the data is multi-class data, so the simple SVM need for further expansion, so that it can solve the multi-value classification. This paper introduces several SVMs for multi-value classification, including "One against Rest", "One against One", Decision Tree and Directed Acyclic Graph SVMs, and compares their respective advantages and disadvantages. By analyzing the deficiencies of various SVMs, a new SVM method, namely, the theory of combining spot and SVM, is put forward. Finally, comparing the traditional SVM to SVM binding spot feature, the tests show the SVM combining of spot applied to hyperspectral image classification has achieved good results.
The principle of SVM based on spot is to choose an appropriate scale to split the image into a series of segmentation, according to certain strategy using spectral information. And this principle ensures the spectral features of the majority of patch pixel similar. This method gathers statistics of each pixel value in the spot and obtains the mean value of each band to replace the original value of all pixels in the spot. The purpose of this classification is that the pixel having the noise brought by various causes is assimilated by the surrounding pixels to merge into a single spot. In other words, under the information of its surrounding pixels recovering the value of the pixel having noise is to not appear the fault isolation in the classification map and to avoid the salt and pepper phenomenon. The results show that this method is feasible and the classification accuracy and speed is better than traditional support vector machine.
引文
[1]张良培,张立福,等.高光谱遥感[M] 湖北：武汉大学出版社,2005.
[2]浦瑞良.高光谱遥感及其应用北京：高等教育出版社[M],2003
[3]李石华,等.遥感图像分类方法研究综述[J].国土资源遥感,2005,64(2)：1-6
[4]杜凤兰,田庆久,夏学齐.遥感图像分类方法评析与展望[J].遥感技术与应用,2004,19(6)：521-525
[5]王一达,沈熙玲,谢炯.遥感图像分类方法综述[J].遥感信息,2006,5：67-71
[6]Wu Hao,Kuang Gangyao, Yu Wenxian. An Unsupervised Classification Method for Hyperspectral Image Combining PCA and Oaussian Mixture Model. Proceedings of SPIE-The International Society for Optical Engineering.2003,5286(2):729-734
[7]Wang Jing, Chang Chein-I. A uniform projection-based unsupervised classification for hyperspectral imagery International Geoscience and Remote Sensing Symposium(1GARSS).2004,5:3066-3068
[8]Shah Chintan A. Arora Manoj K, Varshney Pramod K. Unsupervised classification of hyperspectral data:An ICA mixture model based approach. International Journal of Remote Sensing.2004,25(2):481-487
[9]钱乐祥等编著.遥感数字影像处理与地理特征提取.北京：科学出版社2004：171-172
[10]Chang Chein-Ⅰ, Chiang Shao—Shan. Anomaly detection mad classification for hyperspectral imagery. IEEE Transactions on Geoscience and Remote Sensing.2002, 40(6):1314-1325
[11]Rand Robea S. Spectral/spatial annealing of hyperspectral imagery initialized by a supervised classification method. Proceedings of SHE-TheInternational Society for Optical Engineering.2002,4725:242-253
[12]Meyer Alan, PaglieroniDavid, Astaneh Cyrus. K-Means Re-Clustering. Algorithmic options with quantifiable performance comparisons. Proceedings of SPIE-The International Society for Optical Engineering.2003,5001:84-92
[13]蒋宗礼.人工神经网络导论[M],北京：高等教育出版社,2001,39-52
[14]Safavian S R, Landgrebe D. A Survey of Decision Tree Classifier Methodology [J]. IEEE Trans. Syst. Man Cybern,1991,21:660-674.
[15]Friedl M A, Brodeley C E. Decision Tree Classification of Land Cover from Remotely Sensed Data [J].RemoteSens.Environ,1997,61:399-409.
[16]孙家炳,舒宁,关泽群.遥感原理、方法和应用[M].北京：测绘出版社,1997.
[17]萧嵘,孙晨,王继成,等.一种具有容噪性能的SVM多值分类器[J].计算机研究与发展,2000,37(9)：1071-1075
[18]明冬萍,骆剑承,周成虎,等.高分辨率遥感影像信息提取及块状基元特征提取研究[J].数据采集与处理,2005,20(1)：34-39.
[19]李小娟,宫兆宁,等.ENVI遥感影像处理教程[M].北京：中国环境科学出版社.2007
[20]张瑞丰.精通MATLAB6.5[M].北京：中国水利水电出版社,2004
[21]Mayoraz E, Alpaydin E. Support vector machines for multi—class classification[C], Proceedings of International Workshop on Artificial Neural Networks,1999,2:838-842
[22]Platt J C, Cristianini N, Shawe-Taylor J. Large margin DAGs for multiclass classification[J], Advances in Neural Information Processing Systems. Cambridge, MA, MIT Press,2000,23:547-553
[23]Hsu C W, Lin C J. A simple decomposition method for support vector machines[J]. Machine Learning,2002,46:291-314
[24]FrancescoM asulli, Gi o rgi o V alent ini. Comparing Decomposition Methods for Classification. Fourth International Conference on Know ledge2 Based Intelligent Engineering, System s & Allied Technologies,2000,788-792
[25]Dietterich T G, Bakiri G. Solving Multiclass L earning Problems via Error Correcting Output Codes. Journal of Artificial Intelligence Research,1995, (2):263-286
[26]杨路明,李丽.一种加速大规模SVM训练的新思路[J].微机发展,2004,14(12)：136-138.
[27]郑春红,郑贵文,焦李成.基于FSVM的雷达多类目标识别[J].系统工程与电子技术,2003,25(11)：1358-1361.
[28]李昆仑,黄厚宽,田盛丰.模糊多SVM模型[J].电子学报,2004,32(5)：830-832.
[29]安金龙,王正欧,马振平.一种新的支持向量机多类分类方法[J].信息与控制,2004,33(1)：262-267
[30]张春华.支持向量机中最优化问题的研究[D].中国农业大学,2004
[31]王锐,胡容兵,王志勇.基于支持向量机的武器装备批量生产经济性分析[J].海军航空工程学院学报,2006,21(6)：687-689
[32]曹克强,胡良谋,张春山,等.支持向量机在电液伺服系统辨识建模巾的应用[J].空军工程大学学报(自然科学版),2007,8(3)：43-45
[33]张华国,黄韦艮,周长宝.应用IKONOS卫星遥感图像监测南麂列岛土地覆盖状况[J].遥感技术与应用,2003,18(5)：306-312
[34]Jutten C, Herault J. Blind Separation of Sources,Part 1:an adaptive algorithm based on neuromimetic architecture [J]. Signal Processing,1991,24(1):1-10.
[35]Vapnik. The nature of statistical learning theory.NewYork:Springer-Verlag,1995
[36]Weston J. and Watkins C. Multi-class support vector machines. In M.Verleysen, editor, Proceeding of ESANN99, Brussels,1999, D. Factor Press
[37]Sebald D J, Buehlew J A. Support vector machines and the multiple hypothesis test problem.IEEE Trans on Signal Processing,2001,49(11):2865-2872
[38]安金龙,王正欧.一种新的支持向量机多类分类方法.信息与控制,2004,6(3)：262-267
[39]卢增样,李衍达.交互支持向量机学习算法及其应用.清华大学学报(自然科学版),1999,39(7)：93-97
[40]D. Tax, R.Duin. Data domain description by support vector. In Porc. ESANN, M. Verleysen, editor Brussels:D. Factor Press,1999:251-256
[41]CHUN FU LIN, SHENG DE WANG. Fuzzy Support Vector Machines[J]. IEEE Transaction on Neural Network,2002,13(2):464-471
[42]刘志刚,李德仁,秦前清,等.支持向量机在多类分类问题中的推广[J].计算机工程与应用,2004,40(7)：10-13
[43]Inoue T, Shigeo A. Fuzzy support vector machines for pattern Classification [C]//IJCNN, 2001.
[44]郑志洵,杨建刚.大规模训练数据的支持向量机学习新方法[J].计算机工程与设计,2006,27(13)：2425-2427.
[45]赵西安,李德仁.二维对称小波与多尺度影像边缘特征提取[J].测绘学报,2003,32(4)：313-319
[46]张志涌.掌握和精通MATLAB[M].北京航天航空大学出版社,2003.1-10
[47]Alberotanza L,Brando V E, Ravagnan G and Zandonella A. Hyperspectral aerial images-a valuable tool for submerged vegetation recognition in the Orbetello Lagoons, Italy. International Journal of Remote Sensing,1999,20(3):523-533.
[48]Bruzzone L and Prieto D F.A technique for the selection of kernel-function parameters in RBF neural networks for classification of remote-sensing images. IEEE Transactions on Geoscience and Remote Sensing,1999,37(3):1179-1184.
[49]Buddenbaum H, Schlerf M, and Hill J. Classification of coniferous tree species and age classes using hyperspectral data and geostatistical methods. International Journal of Remote Sensing,2005,26(24):5453-5465.
[50]Ursula C B, Hofmann P, Willhauck G,et al. Multi-resolution object-oriented fuzzy analysis of remote sensing data for GIS-ready information [J]. Journal of Photogrammetry &Remote Sensing,2004,58:239-258
[51]陈秋晓,骆剑承,周成虎,等.基于多特征的遥感影像分类方法[J].遥感学报,2004,8(3)：239-244.
[2]浦瑞良.高光谱遥感及其应用北京：高等教育出版社[M],2003
[3]李石华,等.遥感图像分类方法研究综述[J].国土资源遥感,2005,64(2)：1-6
[4]杜凤兰,田庆久,夏学齐.遥感图像分类方法评析与展望[J].遥感技术与应用,2004,19(6)：521-525
[5]王一达,沈熙玲,谢炯.遥感图像分类方法综述[J].遥感信息,2006,5：67-71
[6]Wu Hao,Kuang Gangyao, Yu Wenxian. An Unsupervised Classification Method for Hyperspectral Image Combining PCA and Oaussian Mixture Model. Proceedings of SPIE-The International Society for Optical Engineering.2003,5286(2):729-734
[7]Wang Jing, Chang Chein-I. A uniform projection-based unsupervised classification for hyperspectral imagery International Geoscience and Remote Sensing Symposium(1GARSS).2004,5:3066-3068
[8]Shah Chintan A. Arora Manoj K, Varshney Pramod K. Unsupervised classification of hyperspectral data:An ICA mixture model based approach. International Journal of Remote Sensing.2004,25(2):481-487
[9]钱乐祥等编著.遥感数字影像处理与地理特征提取.北京：科学出版社2004：171-172
[10]Chang Chein-Ⅰ, Chiang Shao—Shan. Anomaly detection mad classification for hyperspectral imagery. IEEE Transactions on Geoscience and Remote Sensing.2002, 40(6):1314-1325
[11]Rand Robea S. Spectral/spatial annealing of hyperspectral imagery initialized by a supervised classification method. Proceedings of SHE-TheInternational Society for Optical Engineering.2002,4725:242-253
[12]Meyer Alan, PaglieroniDavid, Astaneh Cyrus. K-Means Re-Clustering. Algorithmic options with quantifiable performance comparisons. Proceedings of SPIE-The International Society for Optical Engineering.2003,5001:84-92
[13]蒋宗礼.人工神经网络导论[M],北京：高等教育出版社,2001,39-52
[14]Safavian S R, Landgrebe D. A Survey of Decision Tree Classifier Methodology [J]. IEEE Trans. Syst. Man Cybern,1991,21:660-674.
[15]Friedl M A, Brodeley C E. Decision Tree Classification of Land Cover from Remotely Sensed Data [J].RemoteSens.Environ,1997,61:399-409.
[16]孙家炳,舒宁,关泽群.遥感原理、方法和应用[M].北京：测绘出版社,1997.
[17]萧嵘,孙晨,王继成,等.一种具有容噪性能的SVM多值分类器[J].计算机研究与发展,2000,37(9)：1071-1075
[18]明冬萍,骆剑承,周成虎,等.高分辨率遥感影像信息提取及块状基元特征提取研究[J].数据采集与处理,2005,20(1)：34-39.
[19]李小娟,宫兆宁,等.ENVI遥感影像处理教程[M].北京：中国环境科学出版社.2007
[20]张瑞丰.精通MATLAB6.5[M].北京：中国水利水电出版社,2004
[21]Mayoraz E, Alpaydin E. Support vector machines for multi—class classification[C], Proceedings of International Workshop on Artificial Neural Networks,1999,2:838-842
[22]Platt J C, Cristianini N, Shawe-Taylor J. Large margin DAGs for multiclass classification[J], Advances in Neural Information Processing Systems. Cambridge, MA, MIT Press,2000,23:547-553
[23]Hsu C W, Lin C J. A simple decomposition method for support vector machines[J]. Machine Learning,2002,46:291-314
[24]FrancescoM asulli, Gi o rgi o V alent ini. Comparing Decomposition Methods for Classification. Fourth International Conference on Know ledge2 Based Intelligent Engineering, System s & Allied Technologies,2000,788-792
[25]Dietterich T G, Bakiri G. Solving Multiclass L earning Problems via Error Correcting Output Codes. Journal of Artificial Intelligence Research,1995, (2):263-286
[26]杨路明,李丽.一种加速大规模SVM训练的新思路[J].微机发展,2004,14(12)：136-138.
[27]郑春红,郑贵文,焦李成.基于FSVM的雷达多类目标识别[J].系统工程与电子技术,2003,25(11)：1358-1361.
[28]李昆仑,黄厚宽,田盛丰.模糊多SVM模型[J].电子学报,2004,32(5)：830-832.
[29]安金龙,王正欧,马振平.一种新的支持向量机多类分类方法[J].信息与控制,2004,33(1)：262-267
[30]张春华.支持向量机中最优化问题的研究[D].中国农业大学,2004
[31]王锐,胡容兵,王志勇.基于支持向量机的武器装备批量生产经济性分析[J].海军航空工程学院学报,2006,21(6)：687-689
[32]曹克强,胡良谋,张春山,等.支持向量机在电液伺服系统辨识建模巾的应用[J].空军工程大学学报(自然科学版),2007,8(3)：43-45
[33]张华国,黄韦艮,周长宝.应用IKONOS卫星遥感图像监测南麂列岛土地覆盖状况[J].遥感技术与应用,2003,18(5)：306-312
[34]Jutten C, Herault J. Blind Separation of Sources,Part 1:an adaptive algorithm based on neuromimetic architecture [J]. Signal Processing,1991,24(1):1-10.
[35]Vapnik. The nature of statistical learning theory.NewYork:Springer-Verlag,1995
[36]Weston J. and Watkins C. Multi-class support vector machines. In M.Verleysen, editor, Proceeding of ESANN99, Brussels,1999, D. Factor Press
[37]Sebald D J, Buehlew J A. Support vector machines and the multiple hypothesis test problem.IEEE Trans on Signal Processing,2001,49(11):2865-2872
[38]安金龙,王正欧.一种新的支持向量机多类分类方法.信息与控制,2004,6(3)：262-267
[39]卢增样,李衍达.交互支持向量机学习算法及其应用.清华大学学报(自然科学版),1999,39(7)：93-97
[40]D. Tax, R.Duin. Data domain description by support vector. In Porc. ESANN, M. Verleysen, editor Brussels:D. Factor Press,1999:251-256
[41]CHUN FU LIN, SHENG DE WANG. Fuzzy Support Vector Machines[J]. IEEE Transaction on Neural Network,2002,13(2):464-471
[42]刘志刚,李德仁,秦前清,等.支持向量机在多类分类问题中的推广[J].计算机工程与应用,2004,40(7)：10-13
[43]Inoue T, Shigeo A. Fuzzy support vector machines for pattern Classification [C]//IJCNN, 2001.
[44]郑志洵,杨建刚.大规模训练数据的支持向量机学习新方法[J].计算机工程与设计,2006,27(13)：2425-2427.
[45]赵西安,李德仁.二维对称小波与多尺度影像边缘特征提取[J].测绘学报,2003,32(4)：313-319
[46]张志涌.掌握和精通MATLAB[M].北京航天航空大学出版社,2003.1-10
[47]Alberotanza L,Brando V E, Ravagnan G and Zandonella A. Hyperspectral aerial images-a valuable tool for submerged vegetation recognition in the Orbetello Lagoons, Italy. International Journal of Remote Sensing,1999,20(3):523-533.
[48]Bruzzone L and Prieto D F.A technique for the selection of kernel-function parameters in RBF neural networks for classification of remote-sensing images. IEEE Transactions on Geoscience and Remote Sensing,1999,37(3):1179-1184.
[49]Buddenbaum H, Schlerf M, and Hill J. Classification of coniferous tree species and age classes using hyperspectral data and geostatistical methods. International Journal of Remote Sensing,2005,26(24):5453-5465.
[50]Ursula C B, Hofmann P, Willhauck G,et al. Multi-resolution object-oriented fuzzy analysis of remote sensing data for GIS-ready information [J]. Journal of Photogrammetry &Remote Sensing,2004,58:239-258
[51]陈秋晓,骆剑承,周成虎,等.基于多特征的遥感影像分类方法[J].遥感学报,2004,8(3)：239-244.