收稿日期: 2004-11-29
修回日期: 2005-02-28
网络出版日期: 2005-11-15
基金资助
中国科学院知识创新工程重大项目(KZCX1-SW-01);国家高技术研究发展计划项目863计划(2003AA131170)资助
Evaluation of Various Classifiers on Regional Land Cover Classification in Huabei Area
Received date: 2004-11-29
Revised date: 2005-02-28
Online published: 2005-11-15
应用MODIS 250m分辨率遥感影像对中国华北地区分别采用最大似然法、Parzen窗、CART决策树、BP神经网络F、uzzy ARTMAP神经网络等5种分类方法进行区域尺度上土地覆盖制图的比较试验.结果表明:(1)Parzen窗法分类性能最优,CART和BP其次,Fuzzy ARTMAP表现较差.(2)CART决策树具有较好鲁棒性,但缺点是样本代价较大;BP神经网络分类器能达到较高精度,但缺点是需较高质量的样本、网络结构参数难以确定,造成其稳健性较差;FuzzyARTMAP则未能表现出理想结果.(3)训练样本数量差异造成:最大似然法的分类精度差异值低于5%;Parzen窗法和Fuzzy ARTMAP差异为5%～10%;CART和BP差异在10%以上。
刘勇洪 , 牛铮 , 徐永明 , 李向军 . 多种分类器在华北地区土地覆盖遥感分类中的性能评价[J]. 中国科学院大学学报, 2005 , 22(6) : 724 -732 . DOI: 10.7523/j.issn.2095-6134.2005.6.011
Five classification methods which are MLC (Maximum Likelihood Classifier),Parzen window,CART decision tree,BP neural network and Fuzzy ARTMAP neural network are selected to map the land cover of Huabei Area in China using MODIS 250m data.The results show that Parzen window performs best in the five classifiers.And CART and BP have satisfactory accuracy whereas Fuzzy ARTMAP has unexpected bad accuracy in comparison with MLC.CART decision tree has better flexibility and robustness.However,it pursues high accuracy at the cost of the sample size.BP neural network has high accuracy but requires high-quality samples and it is hard to define its net structure parameters.The results also show the classification accuracy difference caused by the size of training samples on MLC,Parzen window and Fuzzy ARTMAP,CART and BP are below 5%,5%～10% and above 10%,respectively.
Key words： MODIS 250m; land cover classification; MLC; Parzen window; CART; BP; Fuzzy ARTMAP
[1] DeFries RS,Townshed JGR.NDVI-derived land cover classification at a globe scale.International Journal of Remote Sensing,1994,15 (17):3567～3586.
[2] Loveland TR,Reed BC,Brown JF,et al.Development of a global land cover characteristics database and IGBP DISCover from 1 km AVHRR data.International Journal of Remote Sensing,2000,21:1303～1330.
[3] Zhang YH,Zhao GX.Classification methods of land use/cover based on remote sensing technologies.Journal of China Agricultural Resources and Regional Planning,2002,23 (3):21～25 (in Chinese with English abstract).
[4] Hanson MC,Dubayah R,Defries RS.Classification trees:an alternative to traditional land cover classifers.International Journal of Remote Sensing,1996,17:1075～1081.
[5] DeFries RS,Hansen MC,Townshend JRG,et al.Global land cover classifications at 8 km spatial resolution:the use of training data derived from landsat imagery in decision tree classifiers.International Journal of Remote Sensing,1998,19:3141～3168.
[6] Hansen MC,DeFries RS,Townshend JRG,et al.Global land cover classification at 1km resolution using a decision tree classifier.International Journal of Remote Sensing,2000,21:1331～1364.
[7] Friedl MA,Brodley CE,Strahler AH.Maximizing land cover classification accuracies produced by decision trees at continental to global scales.IEEE Transactions on Geoscience and Remote Sensing,1999,37 (2):969～977.
[8] Friedl MA,Mclver DK,Hodges JCF,et al.Global land cover mapping from MODIS:algorithms and early results.Remote Sensing of Environment,2002,83:287～302.
[9] Murthy CS,Raju PV,Bardrinath KVS.Classification of wheat crop with multi-temperoral images:performance of maximum likelihood and artificial neural networks.International Journal of Remote Sensing,2003,24 (23):4871～4890.
[10] Kavzoglu T,Mather PM.The use of backpropagating artificial neural networkd in land cover classification.International Journal of Remote Sensing,2003,24 (23):4907～4938.
[11] Gopal S,Woodcock CE,Strahler AH.Fuzzy neural network classification of global land cover from a 1°AVHRR data set.Remote Sensing of Environment,1999,67:230～243.
[12] Borak JS,Strahler AH.Feature selection and land cover classification of a MODIS2like data set for a semiarid environment.International Journal of Remote Sensing,1999,20:919～938.
[13] Wang J,Dong GR,Li WJ,et al.Primary study on the multi-layer remote sensing information extraction of desertion of desertification land types by using decision tree technology.Jouranl of Desert Research,2000,20 (3):243～247 (in Chinese with English abstract).
[14] Zhang F,Xing Z,Kou N.Research on rice fine classification using hyper-spectral remote sensing imagines.Journal of Wuhan University of Technology,2002,24 (10):36～39 (in Chinese with English abstract).
[15] Zhao P,Feng XZ,Lin GF.The decision tree algorithm of automatically extracting residential information from SPOT images.Journal of Remote Sensing,2003,7 (4):309～315 (in Chinese with English abstract).
[16] Sun DF,Ji CY,Lin P.Landcover classification of remote sensing imagery using self-organizing neural network.Journal of Remote Sensing,1999,3 (2):139～143 (in Chinese with English abstract).
[17] Chen YM.Application of neural network to classification of remote sensing image.Journal of Geomatics,2002,27 (3):6～8 (in Chinese with English abstract).
[18] Liu ZJ,Wang CY,Yan H,et al.High resolution land cover image classification and evaluation based on fuzzy ARTMAP neural network.Journal of Image and Graphics,2003,8 (2):151～154 (in Chinese with English abstract).
[19] Wang JB.Chinese Terrestrial Net Ecosystem Productive Model Applied Remote Sensing Data:[Ph.D.Thesis].Hangzhou:Zhejiang University,2004 (in Chinese with English abstract).
[20] Richard OD,Peter EH,David GS.Pattern Classification.Beijing:China Machine Press,CITIC Publishing House,2003 (in Chinese).
[21] Mao JX,Wang YN.Remote sensing image classification based on neural network.Measurement and Control Technology,2001,20 (5):29～30 (in Chinese with English abstract).
[22] Li M,Yan CH,Liu GH.Fuzzy ARTMAP with adaptive vigilance parameter for each cluster.Journal of Image and Graphics,1999,4 (5):377～382 (in Chinese with English abstract).
[23] Liu YH,Niu Z.Regional land cover image classification and accuracy evaluation using MODIS data.Remote Sensing Technology and Application,2004,19 (4):217～224 (in Chinese with English abstract).
[24] Wang ZX,Liu C,HUETE A.From AVHRR-NDVI to MODIS-EVI:advances in vegetation index research.Journal of Ecology,2003,23 (5):979～987 (in Chinese with English abstract).
[25] Zhao YS,et al.Application Principal and Methods of Remote Sensing.Beijing:Science Press,2003 (in Chinese).
[26] Friedl MA,McIver DK,Brodley CE.Integration of domain knowledge in the formof ancillary map into supervised classification of remotely sensed data.Proceedings of the the International Geoscience and Remote Sensing Symposium ( IGARSS),2002,2:1038～1040.
/
| 〈 |
|
〉 |
