Integrated Forest Monitoring System for Early Fire Detection and Assessment

Abstract

This paper presents a novel system for automatic early detection of wild forest-fire using optical and thermal cameras at ground station and mounted on Unmanned Aerial Systems (UAS). The proposed system can detect and identify forest fires threats in real-time, and at the same time it is capable to notify the interested parties and authorities by providing alerts and important information (e.g. specific location, environmental conditions, etc.). Early recognition and detection of forest fires is a very challenging problem. Numerous potential sources of error leads to an increased rate of false positives (false alarms). The proposed system addresses them by exploiting the complementarities between thermal and optical cameras located at a panoramic static ground location together with the same type of cameras mounted onboard a (or potentially more) UAS. Also, the system is equipped with sensors to monitor and take into account the environmental conditions for the on the fly threat assessment. All data are fused together for an automated risk assessment. The proposed system has been integrated and it is ready for experimental testing and validation in close-to-operational conditions in field fire experiments with controlled safety conditions carry out in Pano Platres forest in Cyprus.