Machine Learning Approaches to Early Wildfire Detection

Abstract

A comparative study of convolutional and transformer-based models for detecting early-stage wildfire indicators from satellite and drone imagery, with a deployed prototype evaluated in field conditions.