Canadian Forest Service Publications
Processing of spatial and temporal queries by AI-guided remote sensing analysis. 1994. Thomson, A.J.; Goodenough, D.G.; Charlebois, D. Proceedings from AI Research in Environmental Systems (AIRIES) Workshop, November 15-17, 1994, Biloxi, Mississippi. [s.n.], [s.1.].
Issued by: Pacific Forestry Centre
Catalog ID: 21331
CFS Availability: Not available through the CFS (click for more information).
An intelligent system (SEIDAM: System of Experts for Intelligent Data Mangement) is being developed to respond to spatial and temporal queries or product requests by integrating remote sensing data from satellites and aircraft with raster and vector Geographic Information Systems (GIS). Expert systems are used to select the appropriate mix of sensors, data processing methods and GIS to provide the answers. Machine learning and case-based reasoning are used to develop plans for handling complex queries. Queries pass through five stages: query parsing, processor selection, processing, processor output and output translation. At each stage, metaknowledge of processor requirements and capabilities as well as metaknowledge about the data itself are used to guidethe analysys. Metaknowledge in SEIDAM is stored in prolog frames which are object-oriented structures with multiple inheritance. The potential of natural language query processing is discussed, along with issues such as ellipses (the ability to build on previous queries), and fuzzy queries (including terms such as “near”). Examples of queries and their processing requirements are discussed.