A Novel Method for Fuel Maps in Urban-Rural Interface Zones

Climate change and the expansion of urban areas into rural ones have significantly increased the risk of wildfires, with serious and lasting consequences for the environment and society. Automatic generation of fuel maps from satellite imagery is a key element for environmental monitoring and a valuable tool for fire prevention and management. In this paper, we propose an approach based on image clustering and segmentation techniques to identify the vegetation present and generate a fuel matrix. Specifically, the K-means algorithm was employed for cluster segmentation, while the Otsu method was used to classify vegetation according to green channel intensity. The analysis of the different nature of fuel provides essential information for estimating the rate of spread of a fire and can be used as input for realistic simulations of its temporal evolution. The generated maps are a key decision support for relevant authorities, facilitating environmental monitoring and land protection.