基于DEM的地形特征点簇研究
文献类型：学位论文
| 作者 | 罗明良 |
| 学位类别 | 博士 |
| 答辩日期 | 2008 |
| 授予单位 | 中国科学院水利部成都山地灾害与环境研究所 |
| 授予地点 | 成都 |
| 导师 | 汤国安 |
| 关键词 | DEM 地形特征点簇 高差阈值 空间结构 |
| 其他题名 | Research on Terrain Feature Point Cluster based on DEMs |
| 学位专业 | 自然地理学 |
| 中文摘要 | 地形表面由一系列的特征要素如点、线、面构成的，他们决定着地形地貌的几何形态和基本走势。其形态及组合特征可作为地球内部构造运动、地表过程以及气候演变等相互作用的标志性载体，探究地形形态形成机理、发展规律、地域分异特征已成为地理、地质、地貌、测绘等学科的核心命题之一。因而自人类采用图形、晕渲及等高线等方式描述地形、认识自然的过程中一直处于相当重要的地位，其中山顶点、鞍部点等特征点具有更强的地理控制、导向及标识意义。DEM的出现，为全面、深入研究特征点及其组合奠定了坚实的基础。 地形特征点簇是山顶点、鞍部点及其连接的径流节点构成的具有距离、方位及拓扑约束的有机结构体。地形特征点簇负载0维点要素信息、1维线连通与约束信息、2维面空间形态与幅度信息、多层面相互关联与控制信息，综合组成包括点、簇、群、层在内的多角度及层面描述和认识地形的基础框架。DEM的出现为地形特征点簇研究提供了丰富的数据源，其质量和精度控制保证了数据的可用性和科学研究的适宜性。在DEM的基础上，借助现代GIS工具实施快速、高效的地形特征点簇已成为可能。 本研究主要内容和结论包括如下方面： 总结前人研究，本研究认为地形特征研究需要改变当前研究单一类别地形特征的研究等思路：实施集成的地形特征研究，研究其在不同地形地貌条件下的差异性与变化，凸显地形特征点簇研究的必要性及深入地形地貌认识的紧迫性。借鉴结构的数学抽象思路，将地形特征点簇空间结构的研究归结为组成、拓扑及等级三个层面研究内容。 首次提出了地形特征点簇概念模型，给出地形特征点簇的组成要素及其内涵界定、形式化描述方法，结合其多层面解译方法给出基于线状连通图的组织方式；讨论了地形特征点簇的属性特征及其分类。 提出了一体化地形特征点簇要素提取方案及形态、模拟相结合的山脊线提取方案。在DEM基础上，进行点簇的提取研究，包括结合形态与模拟的山脊线提取方案、基于山脊线和窗口分析提取鞍部点方法以及一体化提取点簇的原理和方法，最终借助山脊线辅助实现从点到点簇的组织与构建。 分析了在点簇提取过程中涉及到的不确定性问题，主要探讨了山顶与鞍部高差阈值的选取、与山顶起伏及山顶控制面积关联关系以及不同格网DEM对其提取的影响等不确定性因素。 构建了地形特征点簇的分析框架。在点簇提取的基础上，研究了点簇的空间结构统计属性特征及其关联特征，结果表明点簇组织方式能够很好地刻画陕西省山区及黄土梁峁区地形特征，根据山顶起伏与其控制面积的相关分析，给出面积与起伏半对数比例关系的参数，该参数能较好区分山区与丘陵；初步研究了点簇空间分布形态特征及拓扑连通特征，发现点簇上述空间结构特征与地形具有较好对应性。 研究结果表明，地形特征点簇具有对地形的概括与抽象能力，在描述地形特征及其空间结构上表现出较大潜力与优势，应当注重这一研究在提取与分析方法上的充实和完善、理论认识上的提高与深入以及实际应用的快速跟进。 |
| 英文摘要 | The surface of terrain can be interpreted as a collection of points, lines and patches, which depict the structure and shape of landform and geomorphology, always named terrain feature elements. Since the characteristics of morphology and their combination identity the tectonic movement, surface process and climate elevation, the research of terrain shape on formation mechanism, development law and spatial distribution have been the core proposition in fields of geography, geology, geomorphology, surveying and mapping. It has been a long time since people recognize the terrain features from pictures, contour lines and rendered pictures. Among the features the mountain peaks, saddles have shown much stronger controlling, guiding and marking meanings in understand the terrain features. With the aid of Digital Elevation Models (DEMs), the research of terrain feature points can be promoted into a comprehensive and further stage. Terrain feature point cluster, abbreviated to TFPC, is a graphical portraying of terrain, which is formed by mountain or hill peaks, saddles and channel junctions according to some constrained rules, including distance, orientation and topology. Terrain feature point cluster can be interpreted in many hierarchies, such as points in 0 dimension, connected lines in 1 dimension and irregular patches in 2 dimensions, and even the interdependence and interaction among the points, lines and patches. Terrain feature point cluster then can be organized into points, point clusters, point groups and layers under different scale or demands of application, which forms the framework of comprehend and depicting terrain from different viewpoints and hierarchies. And now the high quality and accuracy of widely used data source Digital Elevation Models ensures the extraction of terrain feature point cluster quickly and effectively. There are the key points of this research as follows: When reviewed the current state of research in terrain features, it can be learnt that the tendency and method should be transformed into integrated terrain features research and paying more attention to the relationship between terrain features, topography and geomorphology, instead of the tendency of only concentrating on the shape and generalization research, and instead of the methodology of only perform the analysis one kind terrain features. It is the research motivation of terrain feature point cluster. According to the mathematical abstract of structure, the spatial structure of terrain feature point cluster can be implemented in three ways by research of constitute structure, topography structure and hierarchy structure. A concept model on terrain feature point cluster has been build up on what is the definition of terrain feature point cluster, and what are their constitute elements and that how it can be formalized depicted. After that there are some discussions about how to organize the feature points, in the thesis the rules of organization is recommended based on connecting points into lines according to some constrain conditions. At last, the classification of terrain feature point cluster is introduced with the aim of compare with other data models and trying to make it easy to understand what on earth terrain feature point cluster is. The extraction of terrain feature point cluster based on DEMs including the following key points. First a method combined morphologic analysis and hydrology simulation is introduced to extract ridgelines. And then a local window analysis is used to extract mountain or hill saddles based on ridgelines. After that a new incorporated method is introduced to extract terrain feature points based on dissecting the surface of terrain into different patches. Then the mountain or hill peaks will locate on the highest points in the patches and the saddles appear where the border of patches intersecting with ridgelines, and that the channel junction is the points where the border of patches intersecting with themselves. With the aided of ridgeline, the extracted terrain feature points can be organized and connect into lines. At last, some uncertainties are discussed about how to determine the elevation difference of peaks and saddles in the extraction of terrain feature point cluster, the elevation difference of peaks and discharge port , relationship between spatial extent of the peaks territory, say and as well. The research of spatial structure on TFPC is mainly concentrated on the research of statistical attribution and their spatial distribution. The result shows that TFPC can delineate the relief in mountainous area and Loess Ridge and Hillocks very well. And there is a semi-logarithm quantitative relationship between and which can distinguish the mountainous regions and hill regions to some extent. And the topology index and spatial distributed index of TFPC also show the difference in the typical areas according to their geomorphologic features individually. All the result shows that the TFPC enable the abstract and generalization of terrain features with the main geomorphologic or topographical features remained. And the conclusion is that the TFPC has show great potentials in depicting the features of topography and geomorphology. And the work of next step is to enrich and optimize the TFPC extraction, to enhance explanation theoretically, and to put it into application as soon as possible. |
| 语种 | 中文 |
| 公开日期 | 2010-10-14 |
| 分类号 | TP3;P22 |
| 源URL | [http://ir.imde.ac.cn/handle/131551/2185]  |
| 专题 | 成都山地灾害与环境研究所_成都山地所知识仓储（2009年以前） |
| 推荐引用方式 GB/T 7714 | 罗明良. 基于DEM的地形特征点簇研究[D]. 成都. 中国科学院水利部成都山地灾害与环境研究所. 2008. |
入库方式： OAI收割
来源：成都山地灾害与环境研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。