Database support for multiresolution terrain visualization

  • Authors:
  • Kai Xu

  • Affiliations:
  • School of Information Technology and Electrical Engineering, The University of Queensland, Brisbane, Queensland 4072, Australia

  • Venue:
  • ADC '03 Proceedings of the 14th Australasian database conference - Volume 17
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

Using Multiresolution Terrain Model (MTM) is a common approach to improve the performance of visualizing large terrain data. With the constant increase in the size of terrain data, it is becoming less feasible to have all the data in main memory during visualization. The data exchange between main memory and secondary storage becomes a bottle-neck of terrain visualization, especially for operations like selective refinement. However, this problem has received little attention so far in the context of multiresolution terrain visualization. The need to access the terrain database at multiple levels of detail (LOD) is a fundamental requirement to support secondary-storage multiresolution terrain visualization. However, conventional spatial database access methods are typically based on single resolution only. A new indexing structure, LOD-quadtree, is proposed to support selective refinement on tree-structured MTMs. The new indexing method manages the terrain data in an x-y-error three-dimensional space. In this way, it incorporates the LOD information into the spatial index and at the same time provides efficient locality support. The new indexing method improves visualization performance by accessing the terrain database efficiently at different levels of detail and restricting the data retrieval within specified area. As the LOD-quadtree has no specific requirements for MTM hierarchy structure, it can be used to support different types of tree-structured MTM without modifying them.