Abstract
Animal activity acoustic monitoring is becoming one of the necessary tools in agriculture, including beekeeping. It can assist in the control of beehives in remote locations. It is possible to classify bee swarm activity from audio signals using such approaches. A deep neural networks IoT-based acoustic swarm classification is proposed in this paper. Audio recordings were obtained from the Open Source Beehive project. Mel-frequency cepstral coefficients features were extracted from the audio signal. The lossless WAV and lossy MP3 audio formats were compared for IoT-based solutions. An analysis was made of the impact of the deep neural network parameters on the classification results. The best overall classification accuracy with uncompressed audio was 94.09%, but MP3 compression degraded the DNN accuracy by over 10%. The evaluation of the proposed deep neural networks IoT-based bee activity acoustic classification showed improved results if compared to the previous hidden Markov models system.
Highlights
Technology development in the last century has transformed all segments of human society
Parameter values were optimized in an empirical way, using the Kaldi automatic speech recognition (ASR) guidelines and the evaluation result based on a random test subset
The bee activity acoustic monitoring evaluation was carried out on a 32-min test set containing 643 individual recordings excluded from the training set
Summary
Technology development in the last century has transformed all segments of human society. The presented solution could carry out automatic classification of bee swarm activity based on captured audio. The main contribution of this article is to apply the deep neural networks for bee swarm activity acoustic classification. The usage of DNN acoustic modeling approaches from the field of automatic speech recognition is proposed This represents an efficient solution, which adds to rapid development. These criteria are essential when considering practical applications Another important contribution of this article is the analysis of lossy MP3 codec audio compression for bee sounds. The main article objective is to develop an efficient deep learning acoustic activity modeling technique for uncompressed or lossy compressed bee sounds.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
