This work proposes a new thermal image processing method for computing the rate of spread (ROS) of forest fires. It is based on an application for linear flame fronts that are generated on flat surfaces with known dimensions. In the first step of the method, the correspondence between the points of the thermal image obtained and the real plane is calculated by means of a direct linear transformation (DLT). Subsequently, the position of the flame front is determined by applying a threshold-value-searching criterion within the temperature matrix of the target surface. The design principles for the implementation of this method in the laboratory are described in depth, as well as the considerations that must be taken into account if the method is to be correctly extrapolated to more complex experimental scenarios, such as prescribed burnings or real forest fire emergencies.

This research is supported by the Spanish Ministry of Education and Science under projects DPI 2004-5414, TIC 2003-09291, and a Juan de la Cierva Post doctoral Fellow ship to JAC. It is also supported by the Department of Universities, Research and the Information Society (DURSI) of the Catalan government, the European Social Fund and the Polytechnic University of Catalonia (UPC).

This work was supported by the project 'Integration of robust perception, learning, and navigation systems in mobile robotics' (J-0929).

Peer Reviewed

et al.