Deep learning-based wildfire detection from satellite imagery

Abstract

Hundreds of forest fires burned over 9 million acres of land in 2015, causing millions of dollars in property damage and immeasurable loss and pain to those families affected. Understanding, tracking, and effectively fighting forest fires is crucial in terms of minimising this damage and loss. Using satellite imagery of potentially detected forest fires can greatly aid in this process considering the increased prevalence and severity of wildfires in recent times, the minimisation of response delay has become even more crucial for improved wildfire mitigation. In this paper, we tend to build a ‘Satellite Imagery Based Wildfire Detection System’ in order to detect the wildfire through a convolutional neural network-based model. The objective is to make a computationally efficient, stable and train the convolutional neural network with the help of the transfer learning method and use a window-based analysis approach to increase the fire detection rate. Using satellite imagery of probably detected forest-fires can greatly aid during this process considering the increased prevalence and severity of wildfires in recent times, the minimisation of response delay has become even more crucial for improved wildfire mitigation.