Qualitative matching of spatial information

  • Authors:

  • Jan Oliver Wallgrün;Diedrich Wolter;Kai-Florian Richter

  • Affiliations:

  • University of Bremen, Bremen, Germany;University of Bremen, Bremen, Germany;University of Bremen, Bremen, Germany

  • Venue:
  • Proceedings of the 18th SIGSPATIAL International Conference on Advances in Geographic Information Systems
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

Next to authoritative spatial representations stemming from surveying and cartography efforts, there have always been spatial representations produced by laypeople which often come in a non-georeferenced form, such as sketch maps or verbal descriptions. With the advent of volunteered geographic information the amount and accessibility of such "sketched" information increased drastically. This results in issues of ambiguity (not knowing what is depicted) and trust (not knowing whether the provided information is correct). To process this kind of information, matching approaches for establishing the correct correspondences between multiple representations are needed. As typically only qualitative relations are preserved in sketched information, performing the matching on a qualitative level has been suggested, but efficient solutions that are able to handle the involved combinatorial explosion of matching hypotheses are still lacking. We address this problem by developing a matching approach that exploits qualitative spatial reasoning to prune the search space while performing a heuristic search through the tree of possible matching hypotheses. The developed approach is general in that it can be employed for different tasks and problem domains, such as data integration and retrieval. In a case-study we apply it to the task of matching a sketch map to a geo-referenced data set.