基于Srtm-DEM与遥感的长白山基本地貌类型提取方法

Extraction of Landform Information in Changbai Mountains Based on Srtm-DEM and TM Data

基于中国1∶100万数字地貌制图工作,以吉林省长白山区为试验区,对基于Srtm-DEM与遥感提取基本地貌类型方法进行了探讨。研究表明,基于Srtm-DEM派生的各种模型对基本地貌类型自动提取具有很大的实效性:其中坡度3.5°是实现该区平地和山地宏观地貌单元自动提取的最佳断点,高程模型及起伏度模型相结合,可初步实现基本地貌类型的自动提取,其结果可很好的反应区域地貌宏观规律;在此基础上,基于30 m空间分辨率的遥感影像,综合运用区域地质、土地利用、植被、土壤等多地理要素信息及制图综合知识,以及地貌单元的完整性特征,以山脊线、坡折线和山麓线为地貌单元的标志线,基于地貌知识的专家修正,可获得研究区完整的基本类型地貌单元,实现了基于Srtm-DEM与遥感的基本地貌类型的智能化提取;最后,以研究区内1∶50万地貌图为准,对提取结果进行定性、定量评价分析表明,其提取总体精度达76%,面积相差较小;该方法在一定程度上改进了传统手工和野外调查为主的提取方法,实现了地貌信息的数字化、定位化及定量化,为全国1∶100万数字地貌信息提取与集成奠定了基础。

Basic landform types are the most fundamental geomorphologcal units that only inferior to Plain and Mountain. Which controls the distribution and change of other ecological and environmental factors to some extent and play the important role for extraction of more detailed landform information in the whole geomorphological system. In this article, based on the project of Digital Geomorphology of 1：1,000,000 in China, a method was designed for the extraction of basic landform types from the 90 m Srtm-DEM and TM imagery and which was demonstrated for the Changbai Mountains of Jilin Province. The results shows that the various models from Srtm-DEM are effective for the extraction of basic landform types. Primary plain and mountain information can be automatically extracted from slope models. Basic geomorphology types sketch map can be obtained from height and relief models by overlaying in ArcGIS. Then, the more precise basic landform types boundary lines were obtained and improved based above obtained Plain and Mountain boundary and expert＇s knowledge from TM imageries which taking into account the relationship between landform and geology, vegetation, landuse, and soil. Therefore, the method of automatic/semi-automatic extraction of geomorphology information based on Srtm-DEM and TM imagery was primary realized. By contrasting to 1：500,000 old landform maps, the overall kappa coefficients of accurate assessment was more than 80%. The result shows that the method is effective and which can improve the work efficiency and reduce the labor intensity of the researchers. The storage form of results not only realize digitalization and quantilization of landform information, but also which can be updated.