SEIDAM: a flexible and interoperable metadata-driven system for intelligent forest monitoring

The Advanced Forest Technologies Group at the Pacific Forestry Centre is continuing to develop a System of Experts for Intelligent Data Management (SEIDAM). SEIDAM manages large amounts of remotely sensed and GIS data and processes information for intelligent forest management and inventory updates. SEIDAM uses artificial intelligence (planning, case-based reasoning, software agents and machine learning) with previously captured domain expertise. SEIDAM uses a Prolog expert system shell called RESHELL. In order to manage and process natural resource information, SEIDAM relies on metadata that describes GIS data, field data and heterogeneous, multi-temporal remotely sensed imagery. The authors discuss improvements to SEIDAM. The system is presently composed of a multitude of software agents that currently reside on a LAN. These agents are controlled by SEIDAM's main expert system and are synchronous in nature. By redesigning the interfaces between SEIDAM agents and the central system, SEIDAM will be able to operate in a distributed asynchronous manner across the Internet by taking advantage of new interchange protocols. For this initial implementation, the authors are concentrating on a suite of agents for automated analysis of AirSAR and AVIRIS data, beginning with the automated management of the hyperspectral and AirSAR meta data.