Development of Web Processing Services for the automated generalisation of isolated polygon groups (WPS-PolyGen)

用于纸质和网络演示的地形图和专题地图的制作越来越多地以自动化程序为基础。根据应用和质量要求，它们由人工、交互的前处理和后处理提供支持。拟议项目的目的是进一步开发自动综合程序，特别是处理分组制图对象的方法。因此，有必要考虑结构关系，这是Bertin(1974)提到的最复杂的概化任务。第一批克服单一对象孤立概化的方法存在于多边形马赛克内的聚集、道路和水网的细化以及建筑群的概化。对于代表湖泊、岛屿和森林等自然对象的孤立多边形组的地图表示的推导，需要进行深入的研究和开发。在拟议的项目中，必须处理以下主要任务。首先，必须开发使用空间索引结构(如R树)进行分区和冲突检测的技术。在第二步中，必须对孤立的多边形组进行泛化算法的设计和开发，重点是类型化。起始点可以是利用分层、聚集和基于图形的聚类方法。其目的是开发一种迭代方法，该方法考虑到随着距离和大小阈值的增加而缩放各自的分辨率来生成特征表示。第三个主要任务是在概化过程中对多边形之间的结构关系进行建模和考虑。为此目的，丰富的数据结构可以被实现为水平关系对象，并且必须定义额外的地图约束以保持各自的组属性的同质化。作为上述任务的补充，该项目的目标是为孤立的多边形组实现进一步的概化运算符作为Web处理服务。研究问题将涉及为个人和组对象开发的泛化运算符的编排和组合。仍然占主导地位的个别对象的孤立概括是可以克服的。通过发展数据丰富方面的关系模型化方法和综合操作员对结构关系的考虑，可以在自动化地图制作方面达到一个新的质量水平。改进的概化Web服务将有助于减少人工处理并增加地图的现实性。

The production of topographic and thematic maps for paper and web presentations are based increasingly on automated procedures. They are supported by manual, interactive pre-and post-processing depending on the application and quality requirements. The aim of the proposed project is to develop automated generalisation procedures further, in particular approaches for processing grouped cartographic objects. Necessary therefore is the consideration of structural relations, which is the generalisation task with the greatest complexity as mentioned by Bertin (1974).First approaches that overcome an isolated generalisation of single objects exist for the aggregation within polygon mosaics, for the thinning of roads and water networks as well as for the generalisation of building groups. Advanced research and development is required for the derivation of cartographic representations of isolated polygon groups which stand for natural objects such as lakes, islands and forests.In the proposed project, the following main tasks has to be processed. First, techniques must be developed for the partitioning and conflict detection using spatial index structures such as the R-tree. In the second step, the design and development of generalisation algorithms has to be carried out for isolated polygon groups with focus on the typification. Starting point can be the utilisation of hierarchical, agglomerative and graph-based clustering methods. The aim is to develop an iterative method, which generates characteristic representations in consideration of scale respective resolution with increasing distance and size thresholds. The third main task is on the modelling and consideration of structural relationships between polygons in the generalisation process. For this purpose, enriched data structures can be introduced implemented as horizontal relational objects and additional cartographic constraints has to be defined to maintain respective homogenise the group properties.Complementary to the above mentioned tasks the projects aims on the implementation of further generalisation operators for isolated polygon groups as Web Processing Service. Research questions will deal with the orchestration and combination of the generalisation operators which have been developed for individual and group objects. The still predominant isolated generalisation of individual objects could be overcome. A new qualitative level can be achieved in the automated map production through the development of approaches for modelling relations in terms of data enrichment and a consideration of the structural relationships by generalisation operators. Refined generalisation web services would help to reduce manual processing and to increase the actuality of maps.