面向地貌形态模式识别的地形空间关系研究

The intrinsic spatial relationship information contained in various landform elements themselves is not only the derivation source of landform information, but also the essential manifestation for regional geomorphological characteristics. The realization of their quantitative analytics is an effective way to break through the barrier of the current short-sighted analytical techniques, and to establish the DEM based geomorphologic form pattern recognition methods with macroscopic cognitive capabilities. Therefore, starting from the progressive perspectives “surface point group->local slopes->regional unit->macro structure”, this project will deeply analyse the spatial relationships between different landform elements themselves and among them through the expression forms, influence factors, basic types and change characteristics, and establish the scientific conceptions and their expression ways; will research for the efficient automated methods for extracting the basic elements of spatial relationships, the algorithms for recognising spatial relationships and for detecting the changes of spatial relationships, which are all mainly based on DEM data and aided by other information sources, and establish the analytic models for landform spatial relationships; will study the regional characteristics of similarities and discrepancies in landform space through multi-region sample analytics and experiments, taking the Loess Plateau as the sample area, due to its prominent diversity in landform types, in order to build up the basic criteria for geomorphologic form pattern recognition based on the landform spatial relationships. The research in this project will not only be the improvement and enrichment to the theories and the methods in the DEM digital landform analytics, but also the innovation and development to the theories and methods of GIS spatial relationships, and hence, it is fully possible to achieve the breakthrough in the results of the scientific research.

地形内蕴空间关系信息，不仅是派生地形信息的本源，而且是区域地貌特征的体现，对其实现量化分析，是突破当前DEM近视眼分析技术，建立具有宏观认知能力的DEM地貌形态模式识别方法的有效途径。对此，本项目将从“地面点群->局地坡面->区域单元->宏观结构”的层进式视角，深入剖析不同地形要素自身及其相互之间空间关系的表现形式、影响因素、基本类型与变化特征，建立地形空间关系的科学概念和表达方式；研究以DEM数据为主，辅助以其它信息源的地形空间关系基本要素高效提取方法、地形空间关系识别算法和地形空间关系变化检测算法，建立地形空间关系分析模式；通过地貌类型多样的黄土高原多样区分析实验，研究地形空间关系的区域相似性和差异性特征，建立基于地形空间关系的地貌形态模式识别基本准则。项目研究不仅是对DEM数字地形分析理论与方法的完善与充实，同时也是GIS空间关系理论与方法的创新与发展，有望取得突破性研究成果。

地形内蕴空间关系信息是区域地貌特征的体现，也是派生地形信息的本源，现有DEM的构建与应用分析缺乏对这一信息的充分关注。对此，本课题从“地面点群->局地坡面->区域单元->宏观结构”的层进式视角，一是研究了不同地形要素自身及其相互之间空间关系的表现形式、影响因素、基本类型与变化特征；二是研究了以DEM数据为主，辅助以其它信息源的地形空间关系基本要素高效提取方法、地形空间关系识别算法和地形空间关系变化检测算法，初步建立地形空间关系分析模式；三是通过黄土高原多样区分析实验，探索了地形空间关系的区域相似性和差异性特征及基于地形空间关系的地貌形态模式识别基本准则。同时，还重点研究了现有散点DEM、等高线DEM、格网DEM对地形要素及其空间关系描述的局限性，初步构建了基于四元表结构的体模式DEM数据模型，有望实现地下地上一体化的DEM模型构建。以及顾及地形要素空间关系正确性的城市DEM、平原河网DEM、高标准农田DEM的构建问题，并成功地将研究成果应用于多项社会服务项目中，取得了良好社会经济效益。整合本课题及前期研究成果，课题团队在2018年研究提出了以高精细DEM为核心的全信息实景地理环境技术与方法，并重点进行关键技术和示范应用攻关。研究成果2019年获安徽省科技进步奖二等奖和首届安徽高校成果转移转化大赛全省第二名（二等奖），同时获批实景地理环境安徽省重点实验室。