Automatic detection of wildfire active fronts from aerial thermal infrared images

Infrared (IR) imagery presents a great applicability in the field of wildland fires. However, most of the research conducted thus far has either focused on pre- and post- fire applications or relied on sophisticated technology. Despite the benefits it would potentially bring to operational fire management teams, real time active fire monitoring has not been achieved yet. The present article proposes a preliminary algorithm to detect and locate the fuel burning interface from thermal IR images. Since it is based on gradient instead of intensity thresholds, those images could ideally be captured with a simple low-cost camera without thermogrammetry capacities, i.e. no absolute temperature values are needed. First results are encouraging and suggest a further developed system could be deployed under operational conditions.