A Study On Multi-scale Vector Map Objects Matching Method And Its Application

Scale and real-time (or temporal) are two most basic characteristics of geospatial data. Multi-scale maps are multiple representations of geospatial phenomena or entities at different levels. Multi-temporal data can be used to represent change process, trends or laws of geospatial phenomena or entities. As the stage named "Primitive accumulation" of fundamental geographic data ends, information share is becoming more and more frequent, which makes it urgent to keep spatial data latest. This situation indicates a need to develop a methodology for geospatial data updating.Currently, many scholars from international societies of geographic information science (GIScience) and cartography have presented a promising program for the updating of multi-scale maps. It indicates to use the updated larger-scale map to update the smaller-scale map. This includes four steps: detecting change from large-scale maps, automatic generalization of change, updates propogating and inconsistency handling. One can find from the four steps that the key is object matching, that is, to find the correspondences beteeen two different scales of maps. Indeed, object matching is also the core technology of spatial data collection, integration and updating.By literature of the methods of object matching are mostly presented for line objects with same or similar scales. In order to match spatial objects among multi-scale maps, a novel matching method is indeed expected. For this purpose, existing methods of object matching are firstly summarized systematically, and some problems are pointed out. Further, the issue of matching between spatial objects from multi-scale maps are highlighted, which is the core of this thesis. To solve this issue, a methodology is proposed, which is comprised of the construction of spatial index (also regarded as rough match), hierarchical matching methods, quantitative assessment of matching result and change detection (that is, application of object matching). The main contents and contributions of this thesis are presented as follows.1. A crucial role of object matching technique is described in the process of map updating. A detailed review of the progress on matching methods of spatial objects in the fields of geographical information science (GIS) and computer science is made. Main issues involved in current methods are pointed out, and futher the urgency of developing matching methods for objects from multi-scale maps are analyzed.2. Based on hierarchy theory, hierarchical methods of the representations of geometric location, semantics and spatial relations are systematically studied, which not only improve efficiency in complex system managing and operating, also found a basis for the proposal of hierarchical spatial index.3. With combination of hierarchical theory and principles of urban morphology, a hierarchical spatial index is constructed by partitioning GIS space irregularly, which overcomes the shortcomings of index record redundancy produced by regular grid index. This index results in quick and efficent retrieval and management for spatial objects. Indeed, this index can be viewed as an extension for existing classification framework of spatial index, and plays a role of "rough matching" for area objects.4. Classification and applicable condition of existing matching methods are discussed, and also categories of object matching methods between multi-scale maps are investigated. Further, various modes and related strategies for matching objects of different types are analyzd. A unified geometric matching solution for spatial objects from multi-scale vector maps is established.5. Transmission of geospatial information in the process of spatial data processing is studied based on multi-scale map object matching and information theory principles. Then a quantitative method for calculating the similarity between matched objects is developed from the perspective of geographic information transmission.6. Existing methods for map change detection and their applicable circumstances are summarized. The content, purpose and role of change detection between multi-scale maps are discussed. Then, driving factors of discrepancies between corresponding objects in multi-scale maps are analyzed. Change type and its identification method for spatial objects are also studied. Finally, a methodology for the change detection based on multi-scale object matching is proposed.