利用3S技术对梅里雪山地区植被制图的精度检验分析

摘要/Abstract

摘要： 在野外考察的基础上,应用3S技术,完成了云南西北部梅里雪山地区的1:50 000的植被图.对已完成的植被图通过野外收集的GPS点进行校正,GPS样点数的多少依据统计学抽样调查的样本大小计算而得,用这些校正样点数建立混淆矩阵进行植被图精度计算,最后利用计算成数方差进行检验.混淆矩阵计算得出植被图总的判对精度即整体精度OA为84.7%,利用计算成数方差检验,结果表明大部分类型为90%以上.基于3S技术完成的植被图精度取决于区域面积大小和植被分类等级,而利用遥感技术来划分的植被等级与传统的植被分类等级不完全一致.

关键词: 植被制图, 3S, 混淆矩阵, 精度分析, 梅里雪山

Abstract: The Meili Snow Mountain (28°20'～28°33'N,98°30'～8°52'E) is a very famous mountain in Northwest Yunnan of China by its rich and well protected biodiversity and Tibetan cultural diversity.By applying 3S (RS-Remote Sensing,GIS-Geography Information System,GPS-Global Position System) technology,the 1∶50 000 vegetation map of Meili Snow Mountain area (total about 332 km²) was drawn out.The vegetation in this area was classified into 18 vegetation types except for stone,glacier and river system.The vegetation map was rectified by applying the GPS points got from the fields.The numbers of GPS points were calculated by the formula of numbers of samples in statistics.313 GPS points were used to rectify the vegetation map.The numbers were fit for the formula of numbers of samples in Statistics.The accuracy and verify of vegetation types distribution in the map was analyzed by building a Probability Error Matrix (PEM) and through the variance analysis.The results indicated that the overall accuracy (OA) of the vegetation map was 84.7%.The accuracy of vegetation map finished by 3S technology was lied on the area of the region and the grade of vegetation class first,but the grade of vegetation class classified by remote sensing technology disaccord with the traditional vegetation class system.The other factors deciding the accuracy of vegetation were the distinguish ability of Remote Sensing image,the accuracy of distinguish,and the numbers of the samples,including vegetation class experts knowledge.

Key words: Vegetation map, 3S technology, Error matrix, Accuracy analysis, Meili Snow Mountain

中图分类号:

S127
TP97

张志明, 欧晓昆, 王崇云, 吴玉成. 利用3S技术对梅里雪山地区植被制图的精度检验分析[J]. 应用生态学报, 2004, (9): 1517-1522.

ZHANG Zhiming, OU Xiaokun, WANG Chongyun, WU Yucheng. Accuracy analysis of vegetation mapping for Meili Snow Mountain area,northwest Yunnan,China[J]. Chinese Journal of Applied Ecology, 2004, (9): 1517-1522.

参考文献

[1] Fu B-J(傅伯杰).2001.Landscape Ecology Principle and Application. Beijing: Science Press. 342～ 358 (in Chinese)
[2] Gan S(甘淑),HeD-M(何大明).2001.Study on remote sensing monitoring forest cover by NDVI technique in mountain area of Lancang river basin.Sci Silv Sin(林业科学),27(4):134～136(in Chinese)
[3] Giles MF, Richard ML. 2001. Mapping tropical forest fractional cover from coarse spatial resolution remote sensing imagery. Plant Ecol, 131:143 ～ 154
[4] Guo Z-H(郭志华),Peng S-L(彭少麟).2001.Combining multitemporal NOAA-AVHRR NDVI and ground auxiliary data for estimating terrestrial net primary production in Guangdong. Acta Ecol Sin(生态学报),21(9):1444～1449(in Chinese)
[5] Hobbs RJ, Wallace JF, Campbell NA. 1989. Classification if vegetation in the Western Astralian wheatbelt using Landsat MSS data.Vegetatio, 80: 91 ～ 105
[6] Jin Z-Z(金振洲),Ou X-K(欧晓昆).2000.Vegetation of Dry-hot Valley. Kunming: Yunnan Science and Technology Press. 265～ 269(in Chinese)
[7] Kimberly RH, Susan LM. 2001. Mapping a forest mosaic. Plant Ecol, 156:105 ～ 120
[8] Li H-G(李红旮),Cui W-H(崔伟宏).2001.A study on the vegetationin Ayarwaddy Watershed,Myanma.Chin J Ecol(生态学杂志),20(supp.):34～38(in Chinese)
[9] Li J-L(李建龙),Jiang P(蒋平).1998.Advancesin study on the remote sensing technology and GPS and GIS integration systems in estimating grassland yield applications in the North of Xinjiang,China.Acta Ecol Sin(生态学报),18(5):504～510(in Chinese)
[10] Li Z-X(李芝喜),He Q(何强).1995.Remote Sensing Analysis on Forest in Yunnan Diqing. Kunming: Yunnan Science and Technology Press. 10～ 17(in Chinese)
[11] Liu C-R(刘灿然),Chen L-Z(陈灵芝).2000.Analysis ofthe patch shape with shape indices for the vegetation landscape in Beijing.Acta Ecol Sin(生态学报),20(4):559 567(in Chinese)
[12] Liu X-H(刘先华).1998.Characteristics of vegetation diversity in the Xilin River basin and their responses to climatic factors. Acta Phytoecol Sin(植物生态学报),22(5):466 472
[13] Ma K-M(马克明),Fu B-J(傅伯杰).2000.Spatial neighboring and distribution regularities of landscape types in Donglingshan mountain region,Beijing,China.Acta Ecol Sin(生态学报),20(5):748～752(in Chinese)
[14] Megan ML. 1998. Numeric classification as an aid to spectral mapping of vegetation communities. Plant Ecol, 136:133 ～ 149
[15] Niu J-M(牛建明).2001.Climate-based digital simulation on spatial distribution of vegetation-A case study in Inner Mongolia. Acta Ecol Sin(生态学报),21(7):1064～1071(in Chinese)
[16] Peng S-L(彭少麟),Guo Z-H(郭志华),Wang B-S(王伯荪).1999. Applications of RS and GIS terrestrial vegetation ecology.Chin J Ecol(生态学杂志),18(5):52～64(in Chinese)
[17] Pinder JE. 1997. The relationship between vegetation type and topography in Lassen Volcanic National Park. Plant Ecol, 131: 17～29
[18] Shirish AR, Roy PS. 1997. Satellite remote sensing for ecological analysis of forested landscape. Plant Ecol, 131:129 ～ 141
[19] Thomas RA, John AK. 2001. Spectral response and pattern of Fraser fir mortality and regeneration, Great Smoky Mountains, USA.Plant Ecol, 156: 59 ～ 74
[20] Wang B-R(王宝荣).2001a.Remote sensing image interpretation of vegetation types.Chin J Ecol(生态学杂志),20(supp.):18～20(in Chinese)
[21] Wang B-R(王宝荣).2001b.Apllication of RS technology to the study of vegetation mapping in the Yulong Snow Mountain, Lijiang,Yunnan.Chin J Ecol(生态学杂志),20(supp.):39～41(in Chinese)
[22] Wu J-G(邬建国).2000.Landscape Ecology-Pattern,Process,Scale and Hierarchy. Beijing: Higher Education Press. 190 ～ 206 (in Chinese)
[23] Xu J-C(徐继初).1990.Biological Statistics and Experimental Design. Beijing: Agriculture Press. 186 ～ 189 (in Chinese)
[24] Xiao D-N(肖笃宁). 1999. The Development of Landscape Ecology Changsha: Hunan Science and Technology Press. (in Chinese)