An Object Model of Direction and Its Implications

  • Authors:

  • Shashi Shekhar;Xuan Liu;Sanjay Chawla

  • Affiliations:

  • Computer Science Department, University of Minnesota, EE/CS 4-192, 200 Union St. SE., Minneapolis, MN 55455 shekhar@cs.umn.edu http://www.cs.umn.edu/research/shashi-group;Computer Science Department, University of Minnesota, EE/CS 4-192, 200 Union St. SE., Minneapolis, MN 55455 xliu@cs.umn.edu http://www.cs.umn.edu/research/shashi-group;Computer Science Department, University of Minnesota, EE/CS 4-192, 200 Union St. SE., Minneapolis, MN 55455 chawla@cs.umn.edu http://www.cs.umn.edu/research/shashi-group

  • Venue:
  • Geoinformatica
  • Year:
  • 1999

Quantified Score

Hi-index 0.00

Visualization

Abstract

Direction is an important spatial concept that is used in many fields such as geographic information systems(GIS) and image interpretation. It is also frequently used as a selection condition in spatial queries. Previous work has modeled direction as a relational predicate between spatial objects. Conversely, in this paper, we model direction as a new kind of spatial object using the concepts of vectors, points and angles. The basic approach is to model direction as a unit vector. This novel view of direction has several obvious advantages: Being modeled as a spatial object, a direction object can have its own attributes and operation set. Secondly, new spatial data types such as oriented spatial objects and open spatial objects can be defined at the abstract object level. Finally, the object view of direction makes direction reasoning easy and also reduces the need for a large number of inference rules. These features are important in spatial query processing and optimization. The applicability of the direction model is demonstrated by geographic query examples.