Audio-Based Wildfire Detection on Embedded Systems

The occurrence of wildfires often results in significant fatalities. As are notorious for their high speed spread, the ability to identify wildfire at its early stage is essential quickly obtaining control fire and reducing property loss preventing life. This work presents a machine learning detecting data pipeline that can be deployed on embedded systems remote locations. proposed consists three main steps: audio preprocessing, feature engineering, classification. Experiments show capable effectively with precision sound over forest’s background soundscape. When being Raspberry Pi 4, takes 66 milliseconds process 1 s clip. To knowledge author, this first edge-computing implementation an audio-based detection system.