A beehive monitoring system based on the Helium Developer Kit and a RasPi with Seed’s 2-Mics ReSpeaker PiHat for audio acquisition and basic audio analysis with classification of queenless vs. with queen. Telemetry is done via the Helium network (it’s based on LoRaWAN).
[image source: LongHive - Hackster.io]
Project description
Source Code
Audio analysis
Audio analysis was realized a bit cumbersome: First convert audio to a graphical representation, in this case a spectrogram, then analyse this graphics with deep learning, in this project a pre-trained Tensorflow Lite Interpreter.
[image source: LongHive - Hackster.io]
Sensors
Beside weight and temperature they measured CO2 and air quality . Unfortunately they used the worst weight sensor you can get, the unreliable bathroom sensors:
