Application of Artificial Intelligence in UAV platforms for Early Forest Fire Detection

Abstract

In this paper we propose a platform for early forest fire detection that uses unmanned aerial vehicle for observing given territory. The UAV platforms are equipped with specialized cameras that are either optical or thermal. The focus of the paper falls on a method for using optical camera for fire detection. We pay attention to the powerful machine learning technique that is used for the implementation of object detection in images. We have trained an object detector that can recognize smoke and fire in the images and at the end we have evaluated the performance and accuracy of the model.