dealing with geospatial information in the semantic ?· dealing with geospatial information in the...
Post on 30-Jun-2018
Embed Size (px)
Dealing with Geospatial Information in the Semantic Web
Damian ODea, Sean Geoghegan and Chris Ekins Command and Control Division
Defence Science and Technology Organisation Edinburgh, SA 5111
Australia Damian.Odea, Sean.Geoghegan, Chris.Ekins@dsto.defence.gov.au
Abstract Geospatial information plays a critical role to the military user. Geospatial Web Services promise to make such information more readily available. Searching for the right Geospatial information, however, is a difficult task. The Semantic Web promises to facilitate this process by improving the capability to search for information by better expressing the context and meaning of the search query. Meshing the two approaches to create a Geospatial Semantic Web is an idea that is gaining prominence in both areas of Geospatial Information Science and Semantic Web Services. The Defence Science and Technology Organisation is also working to develop a system to explore and develop the Geospatial Semantic Web as a concept. We report our initial thoughts about the feasibility of such a program.
1 Introduction Geospatial information is critical to the military user. Accessing the correct information is a complex task that often requires that the user understand more about the geospatial domain than their training provides. Enabling a query process that allows effective retrieval of the required information is a positive step for a geospatial intelligence officer. Intelligence analysts often need to deal with geospatial information in the course of their duties, such as ascertaining where events may occur and what facilities or logistically important environmental elements are present in an area of concern. The concept of Network Centric Warfare requires the availability of information from sensor-to-shooter, and so knowing where to look and where to target is of critical importance to the military user of the future. While the Semantic Web (SW) activities and products address many of the requirements for capturing semantics and expressing ontologies, there is a fundamental lack in regard to capturing even the basics of geospatial information. There is a recognised need for an improvement in the capture of geospatial concepts to better facilitate querying of geospatial databases (Egenhofer, 2002). Furthermore, the nature of geospatial processing is highly mathematical, and thus unsuited for the logical formalisms behind the Semantic Web languages. Such processing needs to be performed externally, but the result of such processing needs to be semantically marked up and made available to classifiers for immediate reasoning.
Copyright 2005, Commonwealth of Australia. This paper appeared at the Australasian Ontology Workshop (AOW 2005), Sydney, Australia. Conferences in Research and Practice in Information Technology (CRPIT), Vol. 58. T. Meyer, M. Orgun, Eds. Reproduction for academic, not-for profit purposes permitted provided this text is included.
2 Current efforts in the Geospatial Semantic Web Numerous efforts are currently active within both the Geospatial Information Science and Semantic Web world. Some of these will be reviewed in brief in this section. The University Consortium for Geospatial Information Science identified the Geospatial Semantic Web as a short-term research priority in 2002. Additionally the development of Spatial Ontologies has been identified as a long-term Research Challenge (Mark et al., 2000) by the consortium. The key areas of interest to the participants involve retrieval of, and interoperability of, spatial information through semantic mechanisms, the development of semantic ontologies for both the geospatial and ecological cases a distinction drawn by Fonseca, et al. (Fonseca et al., 2002a) in 2002 - and development of Ontology-Driven Geospatial Information Systems (ODGIS) (Fonseca et al., 2002b). The Wisconsin Land Information System (WLIS) described by (Wiegand et al., 2003) is a means of providing full-fledged DBMS style querying across multiple distinct web-accessible geospatial information systems (GIS). This is achieved through a technique of ontology mapping from a global ontology (encoded as an XML document) to the diverse schema of various heterogeneous web-enabled GISs. The technique, as described in (Cruz et al., 2002), is really a solution to database schema mapping and overcomes the issues of syntactic heterogeneity in GIS data models, while relying upon the expert driver of an Agreement Manager to identify the semantic homogeneity of the concepts encoded in the databases. The query language used in the WLIS is XQuery, and a query processor forms sub-queries for each distinct data source, based upon the agreement (schema mapping) developed through the Agreement Manager above. The global ontology used for the basis of the mappings is a taxonomic classification of land use concepts, and does not contain any relationships between these concepts. At the W3C Workshop on Rule Language for Interoperability (April 2005), several use cases were presented to drive the effort of developing a standard web-based language for expressing rules. In use-case 6.1 Chen, et al. presented the need for the proposed language to express and process geospatial inference rules (Chen et al., 2005). Rules are needed for geospatial processing to deal with situations where reasoning moves from the geometry level (spatial representation of real world features) to the topology level (rules defining how spatial features relate to one another). For example, at the geometry level one may define a relationship between two spatial features to capture the relationship that the first object contains the second, or that the first crosses the second. A topology instead defines a rule that governs how a group of spatial objects behaves, e.g. no spatial feature in the group may cross another (ensuring boundaries remain contiguous instead of overlapping).
Additional requirements include enabling data interoperability by using rules to convert from one geometric representation to another or converting units of measure between different schemes. Another W3C Workshop, Frameworks for Semantics of Web Services (June 2005) saw a presentation from Lieberman, et al. (Lieberman et al., 2005) discussing the evolution of semantics for Geospatial Web Services. This presentation focused primarily on the Geospatial Semantic Web Interoperability Experiment described below. The emphasis of the paper was on the requirements for developing the geospatial semantic web further, as although the current technology goes part of the way to meeting the needs of this project there are still remaining aspects that are not available. These include a means to enable the processing of geospatial relationships and mechanisms to be able to describe the service content of the web services supporting the GSW. The Open Geospatial Consortium standards provide mechanisms for geospatial web services to describe the capabilities of the service itself, but require greater detail to support self-description of the dynamic content of many such services. The Geospatial Semantic Web Interoperability Experiment (GSW-IE) is a project being conducted by the Open Geospatial Consortium (OGC) in 2005. The intention is to develop a means of expressing spatial queries in a semantic manner (i.e. with an ontology) and provide web-based services to fulfil these queries. The program involves participants from Northrop-Grumman, BBN Technologies, and the National Geospatial-intelligence Agency (NGA) as well as several universities, and the Defence Science and Technology Organisation (DSTO) has Observer status on the project. The experiment requires the development of several ontologies, including a geospatial feature ontology for expressing the query, a service ontology to describe the web service function of a Web Feature Service Filter Encoding (WFS/FE) (Vretanos, 2005), and domain ontologies to define the specific domain of interest for the experiment. The base geospatial ontology will closely match the key conceptualisations captured in the Geospatial Markup Language (GML) specification (Cox et al., 2004) as well as geospatial metadata standards such as FGDC (Federal Geographic Data Committee, 1998). GML is an XML grammar encoded in XML Schema supporting the transport of geospatial information across the Web. As such it is a purely syntactic language and lacks explicit semantic information about the objects, relationships and functions described by documents that conform to the schema. The key concepts of GML are laid down by the OGC Abstract Specification1, which acts as a roadmap for the development of the implementation specifications produced by the OGC. The service ontology expresses the contract fulfilled by the WFS interface to support the response to geospatial feature requests. It is described in OWL-S, a language intended to describe Semantic Web Services.
3 Towards a Geospatial Knowledge Infrastructure Within DSTO there is an ongoing work program to develop a demonstration of a Geospatial Knowledge Infrastructure (GKI). The GKI is intended to provide a geospatial or intelligence analyst with numerous capabilities not available today. These include the ability to access data from diverse sources using a single web-based application, supported by semantic querying and browsing. As such the GKI plays an information integration role. Furthermore an analyst is
empowered to make assertions into the knowledge base to explore the entailments of such assertions in a sandboxed model, and to commit useful knowledge back to the communal knowledge base. The intention of the prototype is to explore the applicability of semantic web technologies and the automatic generation of geospatial knowledge in enabling shared geospatial intelligence at a strategic level. Exploration of semantic web technologies include the suit