摘要
以CBERS-1卫星图像为信息源,在新疆石河子地区,通过三种不同分类方法的对比,得出Fuzzy-ARTMAP的分类结果精度最高。在GIS的支持下,根据Fuzzy-ARTMAP的分类结果提取了棉田信息。利用2001年和2002年不同时相的MODIS和TM图像计算得到归一化植被指数(NDVI)和NDVI差值图像,将NDVI差值图像与棉田信息叠加后并分级,得到年际间和年内棉花长势的动态分布图和相同时相棉花长势的区域差异图。
The classification accuracy of Fuzzy-ARTMAP was best among three classifiers in comparison with other two classifiers' for land use classification in arid areas using CBERS-1 image located in Shihezi Municipality, Xinjiang Province in our case study. Cotton field information in Shihezi was extracted based on the classification result of Fuzzy-ARTMAP neural network classifier and GIS analysis. And the NDVI images and NDVI difference images were calculated from multi-temporal MODIS and TM images in 2001 and 2002. Then overlaid with the distribution map of cotton field in Shihezi, the NDVI difference image was classified into five categories to provide inter-annual and intra-annual growth status maps of cotton, so as to monitor the cotton growth condition.
出处
《地理与地理信息科学》
CSSCI
CSCD
北大核心
2003年第4期101-104,共4页
Geography and Geo-Information Science
基金
国家重点基础研究发展规划项目(G2000077900)
国家科技攻关计划(2001DFBA0005)
