邻近场空间关系约束下线要素化简方法研究

Automatic cartographic generalization is a worldwide technical problem. The simplification of linear elements has always been one of the hotspots and difficulties in the field of cartographic generalization. However, current studies have shown great uncertainty in keeping the original data geometry features, consistency of the topological relation of the map and the evaluation indexes and methods on generalized results. Classical line simplification algorithms always aimed at single line, which didn’t consider the spatial relationships. So, the shortcomings were obviously, for example, the calculated results is blunt in D-P algorithm, and the local extreme value point is missing in L-O algorithm, especially it will exist abnormal curve intersection with the larger of linear curvature. So, a line simplification method under the constraint of spatial relationships in the adjacent field will be proposed. Firstly, the one-order, two-order and K-order adjacent field of linear elements will be investigated systematically. Then, the representation model for spatial relationships will be studied, and the relationships of the linear element with other objects will be derived. Secondly, the global simplification method on linear elements and the evaluation indexes and model will be established. Finally, the new method will be validated by making use of the actual data including contour, roads and water. This study is very valuable for cartographic generalization automatically both in theoretically and practically.

地图自动综合是世界性技术难题，线要素的化简一直是地图综合领域研究的热点和难点之一。然而，已有研究在保持原始数据几何结构特征与拓扑关系一致性、综合结果评定指标和评定方法等方面都表现出极大不确定性。经典化简算法多针对单独线要素进行处理，缺乏对该线要素邻近场同构、异构要素之间整体空间关系的考虑，并且存在计算结果生硬（如D-P算法）、局部极值点缺失（如L-O算法），特别是在曲度较大之处出现相交异常等问题。本项目开展一种邻近场空间关系约束下线要素化简方法研究，首先探讨线要素的K阶邻近场，其次研究同构或异构线要素空间关系表达方法；然后建立顾及邻近场空间关系的线要素全局化简方法（LGSM），建立以信息量为核心的化简结果质量综合评价模型；最后采用等高线、道路、水系实际数据，针对上述理论方法开展真实性、有效性检验。本项目的研究对地图自动综合有重要的理论和实用价值。

地图自动综合是世界性技术难题，线要素的化简一直是地图综合领域研究的热点和难点之一。现有研究在保持原始数据几何结构特征与拓扑关系一致性、综合结果评定指标等方面都表现出极大不确定性。经典化简算法多针对单独线要素进行处理，缺乏对该线要素邻近场同构、异构要素之间整体空间关系的考虑，并且存在计算结果生硬、局部极值点缺失，特别是在曲度较大之处出现相交异常等问题。本项目系统化研究了线要素邻近场、线要素与邻近场内的同构或异构实体空间关系表达及其计算检索，提出了邻近场空间关系约束下线要素化简方法以及化简结果的定量化评价指标。1）依托动态距离变换探讨线要素的K阶邻近场建立；2）研究同构或异构场景下层次嵌套结构、树状结构和格网结构等多种空间特征的邻近关系计算方法；3）提出邻近场约束下的线要素选取与化简方法，并采用等高线、道路、水系实际数据开展真实性、有效性检验；4）提出兼顾局部与全局、位置精度和形态相似的定量化评价指标及评价方式。.项目研究成果能够主动探测线要素邻近场相关实体特征及空间关系并以此为约束引导化简算法的智能化优化，避免了现有化简方法在实际应用时会有相交、相接、弯曲群化简不合理等严重问题，为实现线要素自动化简奠定基础，对地图自动综合有重要的理论和实用价值。项目研究成果累计发表论文20篇，其中SCI/EI论文16篇，获批发明专利5项，软件著作权1项。培养了博士、硕士研究生8名（毕业6名）。