Forest Monitoring System Through Sound Recognition

Abstract

The surface of land covered by forests is in continuous decline, causing a significant problem for the environment and the population. Loss of forest space through illegal deforestation followed by natural disasters, such as fires, landslides or floods are all affecting our planet. In the past years, the statistics which have been made indicated substantial fall in forest areas. This problem has been raised not just in Romania, but also internationally and researchers are attempting in different ways to combat deforestation and prevent disasters induced by big losses of wood. This article proposes a forest monitoring process to prevent illegal logging. The goal of the system is to identify the possible dangers by continuously recording the noise in the forest, by processing segments of the recorded signals and decide upon the nature of each of these segments. The system will implement an algorithm able to isolate the audio signal corresponding to cutting wood material (for example, by using a saw, etc.). Eventually, the system will produce an alarm to register the specific time when the event was detected. Alarms will be transmitted in real time and will notify the authorities to take actions to prevent illegal deforestation.