To define and develop the prototypal components of a decision
support system for fire risk studies in Mediterranean areas based
on the combined evaluation of remotely sensed data and geographic
information, and using dynamic modelling of the fire spread.
The project will define, experiment, and apply a knowledge-based
strategy for the automatic analysis of multi-source data, using
fuzzy sets as representation framework. Approximate reasoning
techniques based on fuzzy production rules will be employed to
model the multifactorial evaluation process in which results from
classification of remote sensing (Landsat-TM, HVR-SPOT, and
AVHRR-NOAA) images are integrated with other geographic
The methodology will be implemented with a general and flexible
set of tools which can provide for procedures tailored to
specific application of fire prevention in different geographic
areas. The overall project will be organised in an orderly
sequence of complementary and cooperatively research activities:
General assessment of the list of factors concurring in fire
risk assignment: selected factors will be modelled according to
requirements imposed by the conceptual model of the Decision
Support System, drawn from fuzzy set framework.
Identification of the set of data sources supporting each factor
listed, including satellite sensors, topography, and others.
Development and testing of software tools to process data to
define the identified factors and organise local database useful
in the preparation of the multifactorial evaluation process.
Development of the Decision Support System including the
modelling of the evaluation-decision scheme underlying the
formulation of fire risk judgements and the development of
general purpose software tools to implement the decision-making
Validation of the performance of the Decision Support System by
comparison with results obtained from the dynamic modelling of
the forest fire spread, and of the overall project by comparing
the mapping results with maps derived by the classification of
the burned areas from satellite images.
A feasibility study to optimise the organisation of the software
tools in a main unit located in a fire prevention office devoted
to the acquisition of multisource data, developing the fire risk
maps for the region concerned, and in remote units (forestry
administrative peripheral centres) devoted to the reception and
visualisation of the local fire risk maps.
The expected achievement is to provide an intelligent support in
the preparation of maps of risk for regions of the Mediterranean
ecosystem in which areas at different degree of risk are
classified by assigning graduated soft risk judgements.
Funding SchemeCSC - Cost-sharing contracts