论文部分内容阅读
本文提出了一种基于知识规则的土地利用/土地覆被分类的新方法。知识规则是基于专家经验建立起来的,反映研究区内不同分类系统下各类别的地理分布特性与地理分布交叉可能性。基于黑河流域90m分辨率DEM、2009年逐月1km分辨率NDVI,参考美国地质调查局(USGS)1km分辨率土地利用/土地覆被数据在欧亚大陆上各类别的聚类中心,应用在上、中、下游分别建立的知识规则,以知识规则结合最近距离的USGS类别聚类的方法,制作了一套与USGS全球土地覆被分类标准一致的、可以用于大气模式以及陆面过程模式的黑河流域土地覆被类型分布数据。本方法分类结果与以往研究采用的类别映射方法的分类结果及实际地物影像进行对比,表明知识规则下的分类结果更能准确表达流域地表覆盖特征,对冰雪、冻土类别和沙地荒漠类别的表现更优。
This paper presents a new method of land use / land cover classification based on knowledge rules. Knowledge rules are based on expert experience and reflect the possibility of geographic distribution and geographical distribution of different categories under different classification systems in the study area. Based on the 90m resolution DEM in the Heihe River Basin and the monthly 1km resolution NDVI in 2009, reference is made to the cluster centers for various types of land use / land cover data on the Eurasian continent at 1km resolution of the United States Geological Survey (USGS) , Middle and lower reaches of the establishment of knowledge rules, knowledge rules combined with the nearest USGS clustering method to produce a consistent with the USGS Global Land Cover Classification criteria, can be used for atmospheric mode and land surface process model Distribution of Land Cover Types in Heihe River Basin. The classification result of this method is compared with the classification results of the class mapping methods used in previous studies and the actual object images. It shows that the classification results under the knowledge rules can express the surface coverage characteristics of the river basin more accurately, Better performance.