Single layer map viewer approach for multi-scale data extraction

Abstract

In any desktop and online mapping platform, almost all types of spatial datasets are prepared in multi-scale levels to support map ratio in each zoom class. Current practices in multi-scale data management are either by file-based, database, layer tiles or by using generalisation algorithms to support each transition zoom level of online maps. Multi-scale datasets, either 2-Dimension (2D) and 3-Dimension (3D) vectors or raster, are prepared to serve as a set of data in several accuracy levels or level of details (LoDs) for sharing purposes. However, existing solutions have several drawbacks, either at the storage cost, updating workload or visualisation (graphic and memory), and are time-consuming. For example, vector datasets lead to redundancy in geometry, attribute, topology and semantics for each LoDs and are unable to be viewed in a single viewer. On the other hand, the raster dataset has multiple resolutions from various sources and times, leading to extra storage and slower rendering performance. Thus, this study aimed to develop new approach in properly managing existing multi-scale spatial data such as vector 2D, 3D CityGML and rasters toward a single layer map viewer. The study introduced the Scale Unique Identifier (ID) for vectors to connect all respective LoDs in the attribute database, enabling cross-LoD information query. The High Definition (HD) Map Extractor tool was developed for the raster, and the Enhanced Terrain Profile (ETP) was upgraded based on works in QGIS software to support cross-resolution queries respectively for polygons and lines. The performance of each mentioned dataset was tested, especially in machine resources utilisation on memory, graphic and processor for updating workload, time taken for cross-scale query, and cost-benefits compared to the existing solutions. The experiment performed in this study improved up to 75 percent of time taken for information retrieval, cost-benefits and maintenance efficiency compared to the existing solutions. The findings would benefit data owners and providers in sharing their spatial datasets while minimising the drawbacks. The study has proven that proper construct, control and management of multi-scale spatial datasets would undoubtedly encourage and expedite data sharing among respective data owners, agencies, stakeholders and public users. The study could be extended by improving data sharing standards, implementing scale unique ID in spatial databases and single viewer for 3D city models.