Mapping of Fire Perimeter and Reconstruction of Prescribed Grassfire Evolution Using Thermal Imagery From a Low-Altitude UAS

Accurate mapping of fire perimeters and reconstruction of fire evolution are essential for fire behavior analysis, fire spread modeling and prediction, and wildland fire situational awareness. In this paper, we present a new method for generating unitemporal fire perimeter maps and reconstructing grassfire evolution using low-altitude, multitemporal thermal infrared (TIR) imagery from a small Unmanned Aircraft System (sUAS). The proposed method is novel in its ability to correct spatiotemporal variations of fire perimeters caused by longitudinal and lateral overlapping in Structure-from-Motion (SfM)-stitched orthomosaics. It offers a simple and robust approach for mapping a large wildland fire using only a single UAS. Additionally, this approach significantly reduces manual processing efforts while improving the temporal fidelity of fire representation, providing a cost-effective solution to produce spatiotemporally consistent fire representations to support modeling, analysis, and fire operations. The effectiveness of the proposed method is demonstrated and evaluated using a TIR UAS dataset collected during a 30-acre prescribed grassfire at the Anderson County Prairie Preserve, Kansas.