Assessment and Validation of a Computer Vision Algorithm for Wildfire Rate of Spread Estimation
Open Access
Abstract
As wildfire activity increases worldwide, developing effective methods for estimating how fast it can spread is critical. This study aimed to develop and validate a computer vision algorithm for fire spread estimation. Using visual flame data from laboratory experiments on excelsior and pine needle fuel beds, we explored fire spread predictions for two types of experiments. In the first, the experiments were conducted in an environment where the flame was maintained visually undisturbed while in the second, real-world scenarios were simulated with visual obstructions. Algorithm performance evaluation was conducted by computing the index of agreement and normalized root mean square deviation (NRMSD) error. Results show that the algorithm estimates fire spread well in pristine visual environments with varying accuracy depending on the fuel type. For instance, the index of agreement between the rate of spread values estimated by the algorithm and the measured values is 0.56 for excelsior fuel beds and 0.51 for pine needle fuel beds. For visual obstructions, varying impacts on the rate of spread predictions were observed. Adding an orange background behind the flame had the least effect on algorithm performance (IAmedian = 0.45), followed by placing a Y-shape element resembling a branch (IAmedian = 0.31) and adding an LED light near the flame (IAmedian = 0.30).Keywords
Funding Information
- National Science Foundation (2318717)
This publication has 27 references indexed in Scilit:
- Development of early fire detection model for buildings using computer vision-based CCTVJournal of Building Engineering, 2022
- Turbulent flame image classification using Convolutional Neural NetworksEnergy and AI, 2022
- Wildland Fire Rate of Spread2018
- Detecting skin in face recognition systems: A colour spaces studyDigital Signal Processing, 2010
- Fire Detection in Color Images Using Markov Random Fields2010
- Comparison of two methods for estimating fire positions and the rate of spread of linear flame frontsMeasurement Science and Technology, 2009
- Fire detection using statistical color model in video sequencesJournal of Visual Communication and Image Representation, 2007
- An image processing technique for fire detection in video imagesFire Safety Journal, 2006
- Average centreline temperatures of a buoyant pool fire obtained by image processing of video recordingsFire Safety Journal, 1995
- Visible structure of buoyant diffusion flamesSymposium (International) on Combustion, 1985