Querying GIS with Animated Spatial Sketches

We present the design of the visual query system VISCO that offers a sketch-based query language for defining approximate spatial constellations of objects. VISCO smoothly integrates geometrical and topological querying with deductive spatial reasoning. It is based on a strong physical metaphor visualizing semantics of query elements. Approximate queries rely on combined topological and geometrical constraints enhanced with relaxations and “don’t cares” that are visualized through live animations. Keywords— visual query systems, visual parsing, deductive GIS, constraints. 1 Motivation and Introduction The need to develop new interface paradigms for interacting with spatial information systems or databases, especially geographical information systems (GIS), has already been noted elsewhere [1, 2]. With respect to HCI an interface should be convenient, easy-to-use, and actively supporting the user. A strong metaphor can motivate users and guide them through the interaction with a system. In response to these considerations we present the visual query system VISCO (Vivid Spatial Constellations) that provides a sketch-based query language for defining approximate spatial constellations of objects. Our query language can express geometric as well as topological constraints. The user is actively supported by animations of queries that visualize examples of approximate constellations. The query language elements are visualized with the help of a naive physics metaphor utilizing rubber bands, (cross)beams, swivel joints, nails, marbles, etc. The meaning of VISCO’s language elements is immediately graspable from the physical properties of their visualizations, e.g. a rubber band may be stretched, shrunk and wrapped around in contrast to a (rigid) beam, a marble can roll around and change its position in contrast to a nail. VISCO offers several novel features that correspond to issues mentioned in a recent survey on visual query systems for databases. Catarci et al. [3] conclude this excellent survey with a list of most significant issues for the design of next generation visual query systems. VISCO’s features incorporate solutions for several of these issues. • Animation is an essential part of VISCO and illustrates possible variations in user sketches. • VISCO deals with spatial data types such as points, segments, polylines, polygons and their possible spatial relationships. • A formal semantics for VISCO is based on a space box (SBox). This SBox [4] combines a qualitative spatial calculus with a description logic (DL) (see also [5, 6] for an introduction to DL) and is grounded on quantitative reasoning (i.e. computational geometry). This logic is an extension of a framework that has been successfully applied to the specification of visual languages [7, 5, 6]. VISCO’s query language can be mapped to expressions of this SBox language. • Data is encoded with the help of the SBox that offers well-defined mechanisms for inferring implicit (spatial) knowledge (e.g. from stored maps). • DL (as basic part of the SBox) is well suited for expressing incomplete or indefinite knowledge and for dealing with metaknowledge. For instance, DL systems automatically compute the subsumption relationship between DL expressions. The resulting taxonomy can easily be utilized for query optimization. • VISCO offerspowerful tools forapproximate questioning that may be used for formulating queries about approximate spatial constellations of database objects. In contrast to other relevant work [2] that focuses on topological descriptions we adopt a bottom-up approach and parse the sketches and their geometry as drawn by the user. VISCO takes the geometry of query sketches seriously but supports the annotation of meta information which can be used to specify almost pure topological queries. The user may add meta information to a sketched query. This meta information specifies relaxations, additional constraints or “don’t cares” that define the interpretation of the query. The visibility of user-defined relaxations and “don’t cares” is a major advantage of our approach. In our opinion this explicit meta information (which has to be supplied by the user) is important since drawings are always in a sense “overspecified” and their (relaxed) interpretation strongly depends on the application domain. The effects of relaxations are visualized through animations. 2 VISCO: Vivid Spatial Constellations The interpretation of a user’s spatial query is a critical part for any spatial query system. The user’s intuition about the interpretation should match with the system’s implemented algorithms. For instance, the concept of a right angle has a strong significance in a CAD system but must be relaxed in