【作者】 周增坡；

【导师】 张洪岩； 程维明；

【作者基本信息】 东北师范大学 ， 地图学与地理信息系统， 2009， 硕士

【摘要】 本文在综述国内外地貌学研究现状和信息提取方法的基础上,以典型地貌形态特征提取为主要研究内容。介绍了地貌形态自动/半自动提取过程中刻画地表形态特征的高程、坡度、坡向、剖面、起伏度、地表切割深度等参数,总结了地貌形态提取的三种方法:1)几何形态与地形环境相结合的地貌类型提取方法;2)专家知识支持的多尺度面向对象的地貌类型提取方法;3)遥感影像与数字高程模型相结合的地貌类型提取方法。然后使用遥感和DEM数据对吉林省靖宇县西南部的龙湾火山群和松花江哈尔滨段的阶地进行了形态特征提取。在提取火山口的过程中,根据研究区内火山口仅位于熔岩台地上的特点,首先将研究区划分为平原和山地。然后,将生成的坡度与DEM高程数据相乘,这就类似对DEM数据按照地形的高低起伏进行了加权,更能突显火山口在高程数据上的表现。根据火山口“中间低,四周高”的形态特点及在DEM上的表现,建立了火山口识别模板对火山口进行识别。最后还要根据遥感影像对提取的结果进行验证和修改。对阶地的提取是基于专家知识的面向对象地貌类型提取方法,首先将遥感影像分为水体和陆地两类。然后按照地理相关法提取出三角洲、低河漫滩、高河漫滩、阶地和平原。阶地和平原在遥感影像上表现出较高的相似性,很难进行区分,因此要进一步区分,还要使用DEM数据,对划分出的阶地和平原进行剖分。在形态特征提取的基础上,还要对各个图斑内的最大高程进行分级分类,根据地质数据对地貌的成因、次级成因和物质组成进行判别,从而最终形成能用于地貌制图的地貌数据。使用遥感和DEM进行地貌形态提取,首先要对研究区进行比较深入的了解,然后总结出提取目标在遥感影像和DEM上的表现,选择能最佳刻画地貌形态的指标进行提取。同时提取的过程中还应该充分利用空间实体的空间展布规律和组合规律,进行分类。对最终的提取结果除了要进行平滑和概括外,还要根据有经验的专家的意见进行修改。更多还原

【Abstract】 This paper summarizes the progress of geomorphology research and the geomorphological information extraction methods at home and abroad and takes the typical geomorphological feature extraction as its main study contents.This paper introduces the parameters which are used most in the process of automatic / semi-automatic extracting geomorphological features.The parameters include elevation,slope,profile,relief,cutting depth of the surface and etc.Then the paper sums up three methods for extracting geomorphological features:1) Using the combination of geometry and topography for extraction;2) Expert-driven object-oriented geomorphological features extracting;3) Combining the remote sensing and elevation data for extraction.Additionally,the paper takes the extraction of craters in Jingyu County,Jilin province and the extraction of geomorphological features in section of the Songhua River in Harbin as an example.In the process of extracting craters,according to the characteristics of the study area that the craters are located on the lava platform,the study area is divided into plain and mountain.Then,the slope data is multiplied by the elevation data,which is similar to weight the elevation in accordance with the relief.It is a way to better highlight the craters in the elevation data on the performance.Take into account of the craters’ form characteristics- intermediate low,around high-and their performance on the DEM,established a template for craters identification.Finally,the results of the extraction need verification and revision based the remote sensing image.In this paper,the expert-driven object-oriented method is used for terraces extraction.Firstly,the image in the study is classified into water body and land(not water body).And then in accordance with the relationships among geographical features,the delta,low floodplain,high floodplain,terrace and plain are extracted. Terraces and plains in the remote sensing images showed high similarity,it is difficult to distinguish,for further differentiate,it is necessary to use the DEM data for classifying plains and terraces.When the form feature extraction finished,the next step is classifying the polygons according their largest elevation,according to geological data discriminating the polygons’ main causes,secondary causes and.the material composition.Eventually,the geomorphological data for geomorphological are formed.When extracting geomorphological features form DEM and remote sensing data, first of all,the study area should be understood deeply.Then sum up the extracting targets’ performance on DEM and remote sensing data,choosing the best portrait of geomorphic indicators to extract patterns.At the same time,the process of extraction also should make full use of space entities the law of spatial distribution and composition of the law for the classification.In addition,the final results of the extraction not only should be smoothed and generalized,but also be modified in accordance with the views of experts who know well of the study area.更多还原