Wavelet applications to forest-fire monitoring and measurement

Wavelets have become a powerful image-processing tool capable to cope with uncontrolled conditions and high levels of noise common in natural environments. The combination of wavelets with fuzzy logic permits an easy particularization of the methods to specific problems using heuristic knowledge. This paper presents two wavelet-based techniques for forest fire monitoring. The first one is a fuzzy-wavelet threshold selection method for fire segmentation in infrared images. The second one is a fuzzy-wavelet filter designed to cancel unnatural fire-spread behavior originated by disturbances such as noise, inaccuracies of the cameras and errors in the calibration. These techniques were used in the "Forest-Fire Monitoring System", developed in the INFLAME project funded by the European Commission in the Environment and Climate Programme.