Wildfires stand as one of the most relevant natural disasters worldwide, particularly more so due to the effect of climate change and its impact on various societal and environmental levels. In this regard, a significant amount of research has been done in order to address this issue, deploying a wide variety of technologies and following a multi-disciplinary approach. Notably, computer vision has played a fundamental role in this regard. It can be used to extract and combine information from several imaging modalities in regard to fire detection, characterization and wildfire spread forecasting. In recent years, there has been work pertaining to Deep Learning (DL)-based fire segmentation, showing very promising results. However, it is currently unclear whether the architecture of a model, its loss function, or the image type employed (visible, infrared, or fused) has the most impact on the fire segmentation results. In the present work, we evaluate different combinations of state-of-the-art (SOTA) DL architectures, loss functions, and types of images to identify the parameters most relevant to improve the segmentation results. We benchmark them to identify the top-performing ones and compare them to traditional fire segmentation techniques. Finally, we evaluate if the addition of attention modules on the best performing architecture can further improve the segmentation results. To the best of our knowledge, this is the first work that evaluates the impact of the architecture, loss function, and image type in the performance of DL-based wildfire segmentation models.

In wildfire research, systems that are able to estimate the geometric characteristics of fire, in order to understand and model the behavior of this spreading and dangerous phenomenon, are required. Over the past decade, there has been a growing interest in the use of computer vision and image processing technologies. The majority of these works have considered multiple mono-camera systems, merging the information obtained from each camera. Recent studies have introduced the use of stereovision in this field; for example, a framework with multiple ground stereo pairs of cameras has been developed to measure fires spreading for about 10 meters. This work proposes an unmanned aerial vehicle multimodal stereovision framework which allows for estimation of the geometric characteristics of fires propagating over long distances. The vision system is composed of two cameras operating simultaneously in the visible and infrared spectral bands. The main result of this work is the development of a portable drone system which is able to obtain georeferenced stereoscopic multimodal images associated with a method for the estimation of fire geometric characteristics. The performance of the proposed system is tested through various experiments, which reveal its efficiency and potential for use in monitoring wildfires.

This paper reports two experimental fires conducted at field-scale in Corsica, across a particular mountain shrubland. The orientation of the experimental plots was chosen in such a way that the wind was aligned along the main slope direction in order to obtain a high intensity fire. The first objective was to study the high intensity fire behavior by evaluating the propagation conditions related to its speed and intensity, as well as the geometry of the fire front and its impact on different targets. Therefore, an experimental protocol was designed to determine the properties of the fire spread using UAV cameras and its impact using heat flux gauges. Another objective was to study these experiments numerically using a fully physical fire model, namely FireStar3D. Numerical results concerning the fire dynamics, particularly the ROS, were also compared to other predictions of the FireStar2D model. The comparison with experimental measurements showed the robustness of the 3D approach with a maximum difference of 5.2% for the head fire ROS. The fire intensities obtained revealed that these experiments are representative of high intensity fires, which are very difficult to control in the case of real wildfires. Other parameters investigated numerically (flame geometry and heat fluxes) were also in fairly good agreement with the experimental measurements and confirm the capacity of FireStar3D to predict surface fires of high intensity.

A safe separation distance (SSD) needs to be considered during firefighting activities (fire suppression or people evacuation) against wildfires. The SSD is of critical interest for both humans and assets located in the wildland–urban interfaces (WUI). In most cases, the safety zone models and guidelines assume a flat terrain and only radiant heating. Nevertheless, injuries or damage do not result exclusively from radiant heating. Indeed, convection must be also considered as a significant contribution of heat transfer, particularly in the presence of the combined effects of sloping terrain and a high wind velocity. In this work, a critical case study is considered for the village of Sari-Solenzara in Corsica (France). This site location was selected by the operational staff since high-intensity fire spread is likely to occur in the WUI during wind-blown conditions. This study was carried out for 4 m high shrubland, a sloping terrain of 12° and a wind speed of 16.6 m/s. The numerical simulations were performed using a fully physical fire model, namely, FireStar2D, to investigate a case of fire spreading, which is thought to be representative of most high wildfire risk situations in Corsica. This study is based on the evaluation of the total (radiative and convective) heat flux received by two types of targets (human bodies and buildings) located ahead of the fire front. The results obtained revealed that the radiation was the dominant heat transfer mode in the evaluation of the SSD. In addition, the predictions were consistent with the criterion established by the operational experts, which assumes that in Corsica, a minimum SSD of 50 m is required to keep an equipped firefighter without injury in a fuelbreak named ZAL. This numerical work also provides correlations relating the total heat flux to the SSD.