Unsupervised Acoustic Detection of Queenless Hives in Honeybees (Apis mellifera ligustica)

The absence of a queen bee directly compromises the survival of the colony. Accurately detecting this condition is essential for modern beekeeping, but traditional methods rely on invasive manual inspections. In this paper, we propose an automated and non-invasive approach based on unsupervised machine learning to identify queenless hives through bee sound analysis. Mel spectrograms extracted from real audio recordings of Apis mellifera ligustica were processed using a convolutional autoencoder. The resulting latent representations were clustered using HDBSCAN, and labels were applied only in the final validation step. The model achieved 98.78% accuracy, outperforming supervised approaches reported in the literature. To the best of our knowledge, no previous study has investigated the acoustic detection of queenless honeybee hives using unsupervised learning. Our results show the potential of unsupervised bioacoustic analysis and support advances in precision beekeeping within data-centric e-Science.