【作者】 简灿良；

【导师】 龚健雅；

【作者基本信息】 武汉大学 ， 摄影测量与遥感， 2013， 博士

【摘要】 比例尺和现势性(或时态)是地图数据的两个最基本的特征。不同比例尺的地图数据表达了地球空间现象或实体在不同层次的形态、结构和细节。不同时态的数据表达了地球空间现象或实体的变化过程、趋势或规律。因而,尺度效应或现实变化都可能导致同一地理实体在不同比例尺、不同时态的地图中存在几何、属性、拓扑等方面的差异,这种差异信息可能包含了真实变化、误差(或容许变化)和不一致性等。然而,真实变化和容许变化是不需要处理的,而不一致性必须加以处理,以提升空间数据质量和空间分析结果的准确性。不一致性是国际公认的空间数据质量指标之一,可以区分为空间不一致性、属性不一致性、拓扑不一致性、语义不一致性等类型。现有的不一致性研究主要局限于相同类型目标之间,特别是线目标,研究尺度多限于相同或相近比例尺,而对于多比例尺地图数据间的不一致性探测处理问题则涉及较少,这主要是由于尺度引起的不一致性类型更为复杂,并且有些不一致性是允许的,而有些又是不允许的,从而导致多比例尺地图数据不一致性的探测、判断和处理要比单一比例尺地图数据的不一致性探测处理复杂得多,因此迫切需要发展新的技术方法来探测和处理多比例尺地图数据不一致性问题。为此,本论文从分析归纳多比例尺地图数据变化分类、描述着手,通过对多比例尺地图数据变化信息的检测,获取包括地理实体真实变化、地理实体随比例尺变换的容许变化和地图数据不一致性等在内的复合信息。本文主要研究多比例尺地图数据的空间不一致性,并将其划分为四类,即几何不一致性、拓扑不一致性、度量不一致性和方向不一致性,并对每种类型不一致性进行了描述。进而,从拓扑关系、方向关系和距离关系三个方面论述了单类型空间关系的描述,并用于探测这些单一类型的不一致性。在此基础上,针对多种类型并存的空间不一致性,本论文进一步采用分解与组合策略对复杂空间关系进行层次分析,通过基本类型空间关系的组合对基于拓扑链的空间关系进行集成描述,从而指导多比例尺地图数据不一致性探测与处理。最后,在对多比例尺地图数据不一致性进行类型分析和探测、处理方法设计的基础上,结合河流、等高线、境界和行政区域等线目标和面目标进行实验分析,验证本文所提理论和方法的可行性,为改善空间数据质量、提升基础地理信息服务水平提供一定的参考和借鉴。本文的主要研究内容和贡献包括：1.从空间数据质量的角度阐述了空间数据不一致性研究的重要性,详细回顾了当前国际GIS界围绕地图空间数据不一致性问题取得的研究进展,分析总结了当前研究及需解决的主要问题和发展多比例尺地图数据不一致性探测处理方法的迫切需求。2.将多比例尺地图空间目标的变化类型按照、线和面目标三种基本目标类型分别进行归纳分类并进行了形式化描述,进而,提出采用4交差拓扑关系模型来探测同名目标变化的方法,有效地将真实变化从制图综合引起的几何差异信息中分离出来,有效地区分了地理实体真实变化和制图综合操作影响。3.构建了多比例尺地图数据不一致性的分类体系,完善了空间数据不一致性分类框架。针对多比例尺地图数据,详细而全面地总结归纳了不一致性的来源,将空间不一致性分为几何不一致性、拓扑不一致性、度量不一致性和方向不一致性四类,并分别进行了形式化描述,提升了现有空间数据不一致性研究的系统性。4.在描述空间目标基本关系的基础上,探讨了单类型空间关系不一致性的探测方法。进而,运用分解和组合策略对复杂空间关系进行层次分析,提岀了基于拓扑链的复杂空间关系集成描述方法,并将其应用于复杂空间关系不一致性的探测。通过河流与等高线等空间不一致性探测实验,验证了所提的空间不一致性探测方法的正确性和有效性。5.分析了当前空间数据不一致性处理处理方法的不足和局限性,在结点捕捉的基础上提岀了空间点、线段和折线目标的不一致性处理模型和方法,并导出了折线目标不一致性处理结果的误差估计公式;基于拓扑链的复杂空间关系集成描述方法,以等高线与河流为例论述了集成空间关系不一致性处理流程,通过与现有方法比较验证了本文所提方法在精度方面的优势。更多还原

【Abstract】 Scale and time are two basic characteristics of map data. Due to different representation details of spatial objects in different scale maps, the same geographic entity may have different geometry, attribute and topology in multi-scale maps, which contain real change, error (or allowable variation) and inconsistency, etc. To improve the quality of spatial data and accuracy of spatial analysis results, inconsistency must be treated rather than real variation and allowable variation.Inconsistency is one of the internationally recognized spatial data quality indices, which can be divided into spatial inconsistency, property inconsistency and semantic inconsistency. Current researches are mostly presented for the same type of spatial objects with same or similar scales, especially for line objects. However, there is little research on inconsistency detection in multi-scale maps. Novel methods must be developed to detect and handle the inconsistency problems among multi-scale maps. For this purpose, existing methods of classification and representation of multi-scale map data variation are summarized systematically. Through the detection of multi-scale map data variation, the compound information including real variation of geographic entity, allowable variation of geographic entity changing with scale and inconsistency of map data are obtained. This thesis mainly focus on the spatial inconsistency of multi-scale map data, which is divided into four classes, including geometry inconsistency, topological inconsistency, metric inconsistency and direction inconsistency. The representation of individual spatial relation and the inconsistency in detection method are discussed from topological relation, direction relation and distance relation. Further, the idea of decomposition and combination is used to hierarchically analyze the complex spatial relation, and the spatial relation based on topology chain is represented integratedly through combination of basic spatial relation types. Based on the method desiged for type analysis, detection and processing of multi-scale map data inconsistency, the experimental results of analysis of line objects and region objects, such as rivers, contour lines, boundaries and administrative regions, have testified the feasibility of the proposed method, which will provide some referenced value for improving spatial data quality and promoting basic geographic information service level.The main contents and contributions of this thesis are presented as follows.1. The significance of inconsistency is elaborated from the aspect of spatial data quality. A detailed review of the progress on spatial data inconsistency in the fields of geographical information science (GIS) is made. Main issues involved in current methods are pointed out, and the urgency of developing inconsistency detection methods for multi-scale map data is analyzed.2. According to the basic spatial object types, the classification and formal representation of variation type of multi-scale map spatial objects are discussed respectively. A method adopting4-intersection-and-difference model to detect the variation of homonym objects is prosposed, which can effectively separate real variation from geometric diffence information caused by map generation and thus lay the foundation for analyzing the sources, classification and respresentation of multi-scale map data inconsistency.3. The main sources of multi-scale map data inconsistency are systematically studied from spatial data quality, thematic properties, storage and representation, and analysis and manipulation, and classified into geometry inconsistency, topological inconsistency, metric inconsistency and direction inconsistency. The four types of representations are given to provide the basis for detecting and processing the multi-scale map data inconsistency.4. Based on the representation of basic relation between spatial objects, the detection method of the inconsistency of the same type of spatial relations is discussed. Further, the complex spatial relation is hierarchically analyzed using the decomposition and combination strategy. An integrated representation method of complex spatial relation based on topology chain is proposed and applied to detect the complex spatial relation inconsistency. The experiments of spatial inconsistency detection among rivers and contour lines have verified the validity and effectiveness of the method.5. The shortage and limitation of existing spatial data inconsistency methods are analyzed. An inconsistency processing model and method are presented for point objects, line objects and poly-line objects based on node capture. An error estimation formula is derived for the processing results of poly-line object inconsistency. Based on integrated representation method of complex spatial relation on topology chain, the integrated spatial relation inconsistency processing procedure is discussed taking the rivers and contour lines as examples. A comparison with existing method verified the accuracy advantage of the method proposed in this thesis.更多还原