Health Classification of Beehive Using Sounds

Abstract

The health of beehives plays a crucial role in
agriculture and biodiversity. Monitoring beehive health
helps in preventing diseases and colony collapse
disorder, thus ensuring a stable population of
pollinators. Traditional methods of inspection are timeconsuming
and can stress the bees. Therefore, this
project presents an intelligent, non-invasive beehive
health classification system using audio analysis and
machine learning. This system leverages the power of
Python, Librosa for audio feature extraction, and a
machine learning classifier, specifically a Random
Forest model, to predict the health status of a beehive
based on sound recordings. A user-friendly graphical
interface built with Tkinter allows users to easily
upload an audio file and receive the classification
result, along with the option to visualize various audio
features like FFT, MFCCs, and spectral features. The
aim is to provide an efficient, accurate, and accessible
solution to monitor and maintain the wellbeing of bee
colonies.