Many animals and birds can not be seen with the human eye as they either only come out in the dark, live underground or live deep in the ocean waters. Fortunately for researchers, these animals often emit audio calls which can be picked up with specialised microphones and analysed on computers. Historically, this analysis has involved having to spend many hours trawling through many tens or hundreds of gigabytes of data which has been both costly and highly prone to human error.
With the general advances in computer technology , software has been created to recognise particular events in a stream of audio such as pulses of amplitude or changes in frequency that occur in a ‘natural’ way, indicating that they were created by an animal. This software has been specifically tailored to animal calls, whether it be audible to humans or in the ultrasonic range and further processing has allowed specific animal species to be identified with an associated probability. The end result is that audio files are split up into chunks that have a probability of anything above zero and those that have a probability of zero, which are deleted, so even the ‘unlikely’ audio chunks are retained. Furthermore, each audio chunk is assigned a probability of good classification and renamed accordingly eg: 98%_Plecotus_Auritus_2020_22_07_14_55_43.wav, which also has a time stamp written into the name.
A further advance in technology has recently allowed the use of ‘Small Board Computers’ (SBCs) to record, analyse and classify audio ‘in the wild’, rather than in the researchers’ office. This enables results to be presented in near real time so that detection of an interesting animal is immediately presented to the researchers, enabling them to make better decisions such as where to position the microphone or whether to continue recording or not.
Our SBC device can be run off 12 volt batteries and be deployed in the wild for days / weeks at a time with data being transmitted every now and again via a LoRa radio or 4G cell phone link, depending on what connection is available. Essentially, it records about 30 second chunks of audio, does a quick analysis of that audio using machine learning, and then renames the audio file with species, confidence and date if something interesting was detected. All other recordings are automatically deleted to save disc space and time for the biologist / researcher.
Sampling rate ranging from 48kHz up to 384 kHz 16 bit.
Bandwidth from 1Hz up to 192 kHz.
Mono up to 4 channels simultaneous sampling.
Results can be displayed in real-time with 30 second delay in either text or spectogram or bar chart format.
Wide power supply from 12 V battery or an external power supply from 6 to 16 V.
Software is optimized for power saving and speed by using asynchronous classification.
Average battery life is about 5 hours using 10 x 1.2 V LiMH AA re-chargeable batteries.
Automatically classifies the subject in a choice of resolutions eg animal / genus / species.
Retains data even if it is only 1% confident up to set limit eg 5 GB and then starts deleting the worst of it to prevent data clogging.
Batch data processing mode can be used for re-evaluating any previous data or new data from other sources.
Open source software – anybody can adapt / improve the code
New models for new geographical zones can be trained using the core software. Wide sound database with sound coming from professional
The core ingredients of this project are:
➤ Raspberry Pi 4 + 4 Gb RAM or Nvidia Jetson Nano ➤ Case and fan ➤ ADC Pi shield for sensing battery and supply voltages. ➤ EVM3683-7-QN-01A Evaluation Board for supplying a steady 5v to the Nano. ➤ Dragino LoRa GPS hat. ➤ Adafruit Si7021 temperature and humidity sensor. ➤ 5 Inch EDID Capacitive Touch Screen 800×480 HDMI Monitor TFT LCD Display. ➤ 12 V rechargeable battery pack. ➤ WavX bioacoustics R software package for wildlife acoustic feature extraction. ➤ Random Forest R classification software. ➤ In house developed deployment software. ➤ Full spectrum (384 kb per second) audio data. ➤ UltraMic 384 or Ultramic384-Quattro usb microphones. ➤ Waterproof case Max 004.
➤ 3.3 K wired resister. ➤ 48.1 K wired resistor. ➤ 270 0805 SMD resistor, ±0.1% 0.125W ➤ RS PRO Panel Mount, Version 2.0 USB Connector, Receptacle