Objective To develop a knowledge based decision support system which bringsto bear ground and spaceborne multisensor data, together withexpert knowledge about desertification processes, on the problemof desertification risk assessment after forest fires, withparticular relevance to the Mediterranean countries of theCommunity.The basic approch is to develop a system using a modern GIS shellfor the integration of various moduli. Input to the system willbe the major parameters that govern the natural forestregeneration, and soil erosion, remotely sensed data and fieldobservations. The output of the system will consist of a set ofinferences and maps showing the degree of risk ofdesertification associated with each area.A rule base will be incorporated into the system, constructedfrom modelling the physical processes that contribute to soildegradation.This rule database used with the ground and remotely sensed datawill give the probabilities for soil erosion anddesertificaion, and of natural regeneration. The ground dataused will refer to 1 : 50 000 topographic maps and will includevegetation cover, soil, erosion and land use data. Further, datafrom existing GIS will include topographical, geological andelevation data. A database will be constructed to incorporate theabove mentioned data .The relevance of the remotely sensed data (ERS--1 SAR, SPOTpanchromatic, Landsat Thematic Mapper)will be studied by investigating the statistical correlationsbetween reflectance and backscattering values with specificstates of soil erosion and natural regeneration. A library offunctions will be build forimage segmentation and relevant feature extraction. Segmentationtechniques combining edge and region information will beinvestigated. Multiband texture analysis will be used . Theinvocation of the appropriate algorithms will be knowledgecontrolled, using information about the time of imageacquisition. The image classificationwill be done using Bayesian decision theory. Two approaches willbe investigated: either combining all sources of information atthe data level, or interpreting the data from each source firstand combining them at the symbolic level afterwards.Finally, the performance of the system will be evaluated bycomparing its output risk maps with conventional maps, to assesspositional and labelling accuracy as well as cost effectiveness. Fields of science natural sciencescomputer and information sciencesdatabasesnatural sciencesearth and related environmental sciencesphysical geographycartographygeographic information systems Programme(s) FP3-ENV 1C - Specific research and technological development programme (EEC) in the field of the environment, 1990-1994 Topic(s) 040103 - Wildfires Call for proposal Data not available Funding Scheme CSC - Cost-sharing contracts Coordinator NATIONAL AGRICULTURAL RESEARCH FOUNDATION Address Terma alkmanos street 11528 Athens Greece See on map EU contribution € 0,00 Participants (2) Sort alphabetically Sort by EU Contribution Expand all Collapse all JOANNEUM RESEARCH FORSCHUNGSGESELLSCHAFT MBH Austria EU contribution € 0,00 Address Wastiangasse 6 8010 Graz See on map National Technical University of Athens Greece EU contribution € 0,00 Address 9,iroon polytechniou avenue 15780 Athens See on map
