Bee Hive Acoustic Monitoring and Processing Using Convolutional Neural Network and Machine Learning

Abstract

This article focuses on utilizing artificial intelligence in monitoring bee colonies by analyzing their sound manifestations through smart sensor networks. It examines available datasets of recorded audio data from the hives. The collection of sound samples is realized through embedded systems consisting of micro-controllers, microphones and various sensors (e.g. temperature, humidity, pressure, weight, etc.) in the nodes of the smart sensor networks located in the hives of our interest. Subsequently, we test the ability of deep learning algorithms to predict the presence of the queen bee in the hive. The work implements models of two deep-learning algorithms and evaluates the created models’ performance in processing sound for one-dimensional signals.