摘要
以北京市为研究区域,分析了该区域的TM(Thematic Mapper;专题制图仪)卫星影像特征,探讨了水域、农田、林地、草地、城市用地以及云和云影在TM的7个波段上的光谱可分性,提出了NDCI(Normalized Difference Cloud Index;归一化云指数),分析建立了基于NDCI、NDVI(NormalizedDifference Vegetation Index;归一化植被指数)、NDBI(Normalized Difference Built-up Index;归一化建筑指数)、MNDWI(Modified Normalized Difference Water Index;改进型归一化水体指数)和坡度数据的简单决策树模型,对研究区的几类主要地物、云和云影的信息进行了提取,并对结果进行了精度评价。在GIS支持下计算了水域、林地、草地和农田的面积,计算了北京市2005年第3季度的水体密度指数和植被覆盖指数。结果表明:该方法的总体提取效果较好,在分类过程中阈值的选取简单、有效,分类结果能够满足计算水体密度指数和植被覆盖指数的要求,从而将遥感技术运用到生态质量气象评价中去,并取得了较为满意的结果。
In this study,the characteristics of TM images for Beijing study area are analyzed to find the spectral separability of TM＇ s bands for different objects, such as water, farmland, forest, grassland, concrete,cloud and shadow. An Normalized Difference Cloud Index （NDCI） is proposed to construct a simple decision tree mode with the Normalized Difference Vegetation Index （NDVI）, Normalized Difference Built-up Index （NDBI）, Modified Normalized Difference Water body Index（MNDWI） and slope. This model is then used to derive the information of ground objects, cloud and shadow, and the precision of results is evaluated subsequently. The areas of water body, forest, grassland and farmland are calculated under the support of GIS technology. Finally, the indexes of water body density and vegetation-cover in the autumn of 2005 in Beijing area are calculated. Results indicate that the selection of thresholds for this model is very simple and effective and the precision of model can satisfy the requirement to calculate the indexes of water body density and vegetation-cover. The results also indicate that the remote sensing technology can be used effectively in evaluating the quality of ecological meteorology.
出处
《南京气象学院学报》
CSCD
北大核心
2007年第5期610-616,共7页
Journal of Nanjing Institute of Meteorology
基金
北京市自然科学基金资助项目(8052010)
关键词
TM卫星影像
生态气象
水体密度指数
植被覆盖指数
遥感
决策树
TM images
ecological meteorology
water body density index
vegetation-cover index
Remote sensing
decision tree
