Detection and Identification of Beehive Piping Audio Signals (D. Fourer and A. Orlowska)

ABSTRACT: Piping signals are particular sounds emitted by honey bees during the swarming season or sometimes when bees are exposed to specific factors during the life of the colony. Such sounds are of interest for beekeepers to predict the state of a beehive and to detect an imminent swarming. The present study introduces a novel publicly available dataset made of several honey bee piping recordings allowing for the evaluation of future audio-based detection and recognition methods. First, we propose an analysis of the most relevant timbre features for discriminating between tooting and quacking sounds which are distinct piping signals. Second, we propose and assess several machine-learning-based methods for the detection and the identification of piping signals through a beehive-independent 3-fold cross-validation methodology.

Materials

We introduce a novel dataset of bee piping audio signals which was built by collecting 44 different recordings which were published by various beekeepers on the YouTube platform. These audio recordings were obtained in a real-world scenario using a microphone located close to a beehive when a piping signal is emitted. Each recording has a duration varying from 2 to 13 seconds and is annotated according to the beekeeper comment respectively as Tooting or Quacking. We extracted and segmented the audio from 14 distinct videos from which the signal is stored without a loss of quality into a WAVE file with a sampling frequency of Fs=22.05kHz and a sample precision of 16 bits. % After removing the silent frames, the resulting dataset contains 36 tooting signals and 8 quacking signals which correspond to a duration of 145 seconds for tooting and 60 seconds for quacking (total 205 seconds). % For possible copyright reasons, we only made publicly available the \ac{stft} matrices and timbre descriptors computed using a matlab implementation of the timbre toolbox \cite{peeters2011timbre}. A more detailed description of the dataset containing the links of the original Youtube videos and matlab codes is provided on IEEE DataPort.

Signal analysis




Results




Code & Dataset


https://github.com/dfourer/beepiping IEEE DataPort.