摘要
随着多波段,多分辨率遥感数据的增加和应用质量的不断提高,需要开发更加有效的遥感数据分类算法,当前图像处理软件中的统计分类方法是以参数假设为约束条件,需要数据分布服从高斯分布。但多数的遥感数据的分布并不服从高斯分布的假设条件,因而导致分类精度不高的原因之一。因此,本文以非线性理论为指导,探索了以SOFM神经网络分类方法,改进的BP神经网络和模糊数学方法的遥感数据分类方法的实现。
本文以广州三水市Landsat TM5,4,3三波段遥感影像和SPOT遥感影像为数据源,主要研究了神经网络和模糊算法两种方法在遥感图像分类中的应用。根据所选区域的实际情况,将土地覆盖/利用类型分为五类:水体(包括河流,湖),旱地,裸露地(包括道路,居民地,桥梁和未利用地)、林地和水田5类覆盖物进行分类试验研究。在基于神经网络分类实验中,我们采用了自组织特征映射神经网络(SOFM)算法和改进的BP算法方法对图像进行了分类,并与传统的最大似然法比较达到了很好的效果。在基于模糊网络方法的分离实验中,我们分别采用了K-均值算法(一种硬划分聚类方法)、模糊C-均值算法(等轴空间聚类)对图像进行了分类,并进行了对比分析,结果显示模糊方法精度较好。
With the increase of multi-band and multi-resolution remote sensing data, more effective algorithms are needed to improve the classification precision of the new remote sensing data. Traditional supervised classification methods require the assumption that data distribution obeys Gauss normal distribution, however, actual remote sensing data distributions do not always obey Gauss normal distribution, which is one of the reasons of low precise classification by using traditional methods. Therefore, the perspective of the dissertation is to select some nonlinear method and to achieve combined results. The basic elements used in combined methods are SOFM Neural Network Classification, BP Neural Network Classification and Fuzzy Mathematics. The algorithms are realized by programing.
This paper takes the Landsat TM5,4,3 and SPOT Remote Sensing image of Guangdong Sanshui as source data. Its major tasks are to research Neural Network and Fuzzy Mathematics using in Remote Sensing image classification. According to the actual situation of the selected areas, we divide the land cover/use into five classes: artificial land(cities and towns,traffic land,mine project land), paddy field, arid land, woodland(forest,lawn, virescence land) ,water. In the experiment based on Neural Network Classification s, we use SOFM Neural Network and BP Neural Network Classification, and have a good result from comparing the result with the result of Maximum likelihood classification. In the experiment based on Fuzzy Mathematics,We analyze and compare the result of fuzzy classification with the result of Hard C-Means Clustering Method ,we can see that it has a good result and higher precision to use fuzzy technique classify remote sensing image.
引文
[1] 吕国楷,洪启旺,郝允充, 等. 遥感概论[M].北京: 高等教育出版社,1997.186～ 189
[2] 章孝灿,黄智才,赵元洪.遥感数字图像处理[M].浙江: 浙江大学出版社, 1997.187～ 196
[3] Baatz M , Schabpe A. Object-Oriented and Multi-Scale Image Analysis in Semantic Networks[A ]. In: Proc. of the 2nd International Symposium on Operationalization of Remote Sensing August 16th- 20th 1999. Enschede ITC.
[4] Tou J T, Gonzalez R C. Pattern Recognition Principles, Addison-Wesley Publishing Company[M]. Reading, Massachusetts, 1974.
[5] Richards J A. Remote Sensing Digital Image Analysis, Springer Verlag[M]. Berlin, 1994. 340.
[6] Kruse F A , Lefkoff A B, Boardman J B, etal. The Spectral Image Processing System (SIPS) – Interactive Visualization and Analysis of Imaging spectrometer Data[J]. Remote Sensing of Environment, 1993, 44: 145～ 163
[7] Mazer A S, MartinM , LeeM , et al. Image Processing Software for Imaging Spectrometry Analysis[J]. Remote Sensing of Environment, 1988, 24 (1) : 201～ 210
[8] Carter P H. Texture Discrimination U sing Wavelets [J]. In SP IE Application of Digital Image Processing X IV , 1991, 15(67) : 432～ 438
[9] Laine A ,Fan J. Texture Classification by Wavelet Packet Signatures [J]. IEEE Trans on Pattern Anal, And Machine Intell,1993, (11) : 1186～ 1191
[10] 朱长青. 小波滤波器及遥感影像分类 [J]. 解放军测绘学院学报, 1997, 14 (1) : 32～36.
[11] 朱长青,杨启和,朱文忠. 基于小波变换特征的遥感地貌影像纹理分析和分类 [J]. 测绘学报, 1996,13 (4) : 17～ 23.
[12] 骆剑承等.遥感地学智能图解模型支持下的土地覆盖土地利用分类.自然资源学报, 2001,16(2):179～183.
[13] Tatem A J , L ew is H G, A tkinson P, et al. Super-resolution Target Identification from Remotely Sensed Images U sing a Hopfield N eural N etwork [J]. IEEE T ransactions on Geoscience and Remote Sensing, 2001, 39 (4) : 781～ 796
[14] Gorte B. Probabilistic Segmentation of Remotely Sensed Images [A]. In: ITC Publication Series [C]. 1998. 63.
[15] Pierce E, Ulaby F T, Sarabandi K, etc . Knowledge Based Classification of Polari metric SAR Images [J] . IEEE Transactions on Geosciences and Remote Sensing, 1994, 32(5) :1081～1086.
[16] 孙家柄,舒宁,关泽群.遥感原理、方法和应用[M],北京:测绘出版社,1997.302～ 338
[17] 梅安新.遥感导论[M].北京:高等教育出版社,2001,1～3
[18] Congalton R G. A Review of Assessing the Accuracy of Classifications of Remotely Sensed Data [J] . Remote Sensing Environ,1991, (37) : 35～46.
[19] 王野乔.遥感及多源地理数据分类中的人工神经网络模型.地理科学,1997,17(2):105～111
[20] 张兆礼,赵春晖.现代图像处理技术及MATLAB实现[M].北京:人民邮电出版社,2001.
[21] 张银辉,赵庚星.土地利用/土地覆盖遥感分类方法的研究综述[J].2002,23(3), 21～25
[22] 陈述彭,赵英时.遥感地学分析.北京:测绘出版社,1990
[23] 周成虎,等.遥感图像地学理解与分析.北京:科学出版社,2001
[24] Boardman J W. Inversion of Imaging Spectrometry Data Using Singular Value Decomposition [J]. Proceedings IGARSS89,12th Canadian Symposium on Remote Sensing, 1989,(4) :2069～2072
[25] 郭 晶,杨章玉.MATLAB 6.5 辅助神经网络分析与设计[M].北京:电子工业出版社,2003.
[26] Boardman J W , Kruse F A , Green R O. Mapping Target Signatures Via Partial unmixing of AVIRIS Data: in Summaries, Fifth JPL Airborne Earth Science Workshop [ R] . JPL Publication, 1995. 23～26
[27] 肖平, 李德仁. 基于人工神经元网络技术的土地利用/覆盖变化探测 [ J] . 武汉大学学报·信息科学版, 2002, 27 (6) : 586-590
[28] 吴简彤,王建华.神经网络技术及其应用[M],哈尔滨:哈尔滨工程大学出版社1998,10～20
[29] 楼顺天,施阳.基于 MATLAB 的系统分析与设计-神经网络[M] .西安:西安电子科技大学出版社 ,1998.
[30] 何明一.神经计算·原理·语言·设计·应用[M] .西安:西安电子科技大学出版社 ,1992.
[31] Aboelghar M , Tateishi R, Tsolmon R. Land-cover Sub-pixel Classification Using Linear Mixture Model on Landsat ETM +Data in EGYPT [ S] . V 8. 3 User’s Guide, version 1. 4. 2
[32] Benham T, Wright G, Dunham R. NOAA AVHRR Sub-pixel Snow Classification[R]. Erdas U K Users Group Meeting, Cambridge . September, 1998.
[33] The IMAGINE Sub-pixel Modeler . Applied Analysis Inc .(1998) IMAGINESub-pixel Classifier TM [S] . ERDAS- V 8.3 User’s Guide, version 1. 4. 2.
[34] NOAA AVHRR Sub-pixel Snow Classification [A] . Erdas U K Users Group Meeting, Cambridge [C] . September, 1998.
[35] Martinec J. Snowmelt-Run off Model for Stream Flow Forecasts [J] . Nordic Hydrol ,1975, (3) : 145～154
[36] 刘庆生, 刘高焕, 刘素红. 黄河口遥感图像光谱混合分解 [J].武汉大学学报, 2001, (3) : 266～269
[37] 潘东晓.遥感图像的神经网络分类法[J ] .国土资源遥感 ,1996 ,129 (3) :49～55
[38] 李祚泳.用 B-P 神经网络实现多波段遥感图像的监督分类[J].红外与毫米波学报,1998,17(2):153- 156
[39] Zadeh L. Fuzzy sets[J]. Information and Control, 1965,8:338-353
[40] Mandani. E. H. Application of Fuzzy Algorithms for Simple Dynamic Plant. Proc. IEEE, 1974, 121:1585-1588
[41] Maiers.J, Sherif.Y. The Application of Fuzzy Control System. Automat.1972,11 (2) : 136-142
[42] King P. j, et. The Application of Fuzzy Control System to Industrial Processes. In IFAC world Congress. Mit, Boston, 1 975, 13 (3):2 35-242
[43] Mandani. E. H. An Experiment in Linguistic Synthesis with a Fuzzy Logic Controller.Int. J.Man Math Studies, 1975, 70: 1-13
[44] Dubois D et. Fuzzy Logics and the Generalized Modus Pones Revisited. Lybem. Syst,1984 , 15: 3-4
[45] Takagi T, Sugeno M. Fuzzy Identification of System and its Applications to Modeling and Control. IEEE Trans on SMC, 1985, 15 (1):116.132
[46] J.C Bezdek, A Convergence Theorem for the Fuzzy ISODATA Clustering Algorthm, IEEE Transaction on Pattern Analysis and Machine Intelligence,l(2),1980.
[47] N.R.Paletal, Generalized Clustering Networks and Kohonen’s Selforganization, IEEE Transaction on Neural Network,4(4),1993,pp.549-557
[48] E.C.K.Tsaoetal, Fuzzy Kohonen Clustering Networks, Pattern Recognition, 27(5),1994,pp.757-767
[49] K.S.Asultaetal, A Global Algorithm for the Fuzzy Clustering Problem,Pattern Recognition,26(9),1993,pp.1357-1367
[50] 高新波等,基于统计检验指导的聚类分析问题,电子科学学刊,2001,22(1).