Author(s):
Jan, Sahib
Date: 2011
Persistent ID: http://hdl.handle.net/10362/8261
Origin: Repositório Institucional da UNL
Subject(s): Qualitative Spatial Reasoning; Spatial Reasoning Done Qualitatively; Application Programming Interface; Generic Qualitative Reasoner; Jointly Exhaustive and Pair wise Disjoint; Geographic Information Systems; Allen’s Time Interval Calculus; Dipole Relation Algebra; Cardinal Direction Calculus; Region Connected Calculus-8; Double Cross Calculus; Algebraic Specification Language; Relational Database Management System; Computer Assisted Mass Appraisal
Description
Dissertation submitted in partial fulfillment of the requirements for the Degree of Master of Science in Geospatial Technologies.
Qualitative spatial reasoning deals with knowledge about an infinite spatial domain using a finite set of qualitative relations without using numerical computation. Qualitative knowledge is relative knowledge where we obtain the knowledge on the basis of comparison of features with in the object domain rather then using some external scales. Reasoning is an intellectual facility by which, conclusions are drawn from premises and is present in our everyday interaction with the geographical world. The kind of reasoning that human being relies on is based on commonsense knowledge in everyday situations. During the last decades a multitude of formal calculi over spatial relations have been proposed by focusing on different aspects of space like topology, orientation and distance. Qualitative spatial reasoning engines like SparQ and GQR represents space and reasoning about the space based on qualitative spatial relations and bring qualitative reasoning closer to the geographic applications. Their relations and certain operations defined in qualitative calculi use to infer new knowledge on different aspects of space. Today GIS does not support common-sense reasoning due to limitation for how to formalize spatial inferences. It is important to focus on common sense geographic reasoning, reasoning as it is performed by human. Human perceive and represents geographic information qualitatively, the integration of reasoner with spatial application enables GIS users to represent and extract geographic information qualitatively using human understandable query language. In this thesis, I designed and developed common API framework using platform independent software like XML and JAVA that used to integrate qualitative spatial reasoning engines (SparQ) with GIS application. SparQ is set of modules that structured to provides different reasoning services. SparQ supports command line instructions and it has a specific syntax as set of commands. The developed API provides interface between GIS application and reasoning engine. It establishes connection with reasoner over TCP/IP, takes XML format queries as input from GIS application and converts into SparQ module specific syntax. Similarly it extracts given result, converts it into defined XML format and passes it to GIS application over the same TCP/IP connection. The most challenging part of thesis was SparQ syntax analysis for inputs and their outputs. Each module in Sparq takes module specific query syntax and generates results in multiple syntaxes like; error, simple result and result with comments. Reasoner supports both binary and ternary calculi. The input query syntax for binary-calculi is different for ternary-calculi in the terms of constraint-networks. Based on analysis I, identified commonalities between input query syntaxes for both binary and ternary calculi and designed XML structures for them. Similarly I generalized SparQ results into five major categories and designed XML structures. For ternary-calculi, I considered constraint-reasoning module and their specific operations and designed XML structure for both of their inputs and outputs.
