Fire Affected Area Estimation For Long Duration Fire Events In Odisha Using Multi-Sensor Data Fusion And Machine Learning

Résumé fourni par la source

The proposed methodology leverages multi-sensor data fusion to identify long-duration fire events (LDFEs) by integrating active fire locations from various sensors (MODIS and VIIRS) enabling comprehensive monitoring of persistent fires in both spatial and temporal dimensions. This dynamic approach enables the real-time tracking and understanding of LDFE behavior. We employed a supervised Random Forest classification technique for rapid burnt area assessment using spectral indices and satellite imagery. The results showcase the successful detection and monitoring of LDFEs, aiding forest managers in resource optimization and effective fire management. The integration of satellite imagery and spectral features facilitates precise burnt area evaluation, offering insight into fire spread patterns. Overall, this approach contributes to a holistic comprehension of fire behavior over time and space, thus enhancing fire management strategies.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.