Go Back

The national security/intelligence community needs improved intuitive mechanisms to access geospatial information. Geospatial analysts, for example, are often interested in finding configurations of objects in space that are similar to some baseline configuration. In response, research in spatial reasoning at the University of Maine has focused on the development of a "spatial-similarity search engine." Models developed have applications in defense and civilian operations.

This research has been supported by the Air Force, the National Imagery and Mapping Agency, the National Science Foundation, and Lockheed-Martin MD&S.

Search engine design has been investigated through a program of several parts

The first part is mathematical model that compares different geometries and determines which of these best match a target configuration. This geometric similarity model is based on spatial relations among objects. It allows an operator to tailor search and retrieval for different spatial criteria, such as the preservation of bearings and distances between elements of the baseline and observed objects.

The second component captures differences in types of objects, for example, the conceptual difference between one type of building and another, and how such differences can be assessed under different contexts. The consideration of such meaning enhances the geometric similarity search.

The third part of the spatial-similarity search engine is an advanced user interface, through which geospatial analysts sketch the configuration they want to find. This freehand drawing is translates into a representation that can be processed with the spatial- similarity search engine.

For more information:

Dr. Max J. Egenhofer

National Center for Geographic Information and Analysis

Department of Spatial Information Science and Engineering

Department of Computer Science

University of Maine

Orono, ME 04469-5711

(207) 581-2149, Fax: (207) 581-2206

E-mail: max@spatial.maine.edu

UCGIS web site: http://www.ucgis.org