Extracted features from opto-acoustic audio recordings of aphids and beetles

Hassall, Kirsty L; Dye, A.; Bell, James

Abstract

Opto-acoustic recorders have been utilised to record the flight of weak-flying insects including aphids (Hemiptera: Aphididae) and beetles (Coleoptera). These data include 52 extracted features from 4928 audio recordings collected in 2019 and 2020 of 8 distinct species.

Citation

Hassall, K. L., Dye, A., & Bell, J. (2020). Extracted features from opto-acoustic audio recordings of aphids and beetles. [Dataset]. https://doi.org/10.23637/rothamsted.981w8

• Publication Date: 2020
• Deposit Date: Feb 9, 2024
• Publicly Available Date: Feb 9, 2024
• DOI: https://doi.org/10.23637/rothamsted.981w8
• Publisher URL: https://repository.rothamsted.ac.uk/item/981w8/extracted-features-from-opto-acoustic-audio-recordings-of-aphids-and-beetles
• Collection Method: Opto-acoustic recorders capture the variation of light when an insect passes through a light beam. Both the main body and the wings cast a shadow in the emitter’s light beam, known as the extinction of light principle, and this shadow is subsequently detected by a receiver photodiode array (14). The Wingbeat Recorder® (Insectronics, Chania, Crete, Greece) was placed underneath a 15,000 ml heavy-walled glass beaker (Duran™).; ; Insects were collected and placed inside the beaker. Over a period of two days, insects were free to disperse in and around the sensor. Flights were automatically triggered as an insect enters the field of view of the LED array, generating a recording lasting 0.6 seconds. All flights were saved as an audio file on an SD card within the sensor along with average temperature and humidity covariates. These audio recordings were subject to internal signal processing (Potamitis I, Rigakis I. Large aperture optoelectronic devices to record and time-stamp insects’ wingbeats. IEEE Sensors Journal. 2016;16(15):6053-61.) after which an extensive data processing pipeline as described in Hassall et al (In Preparation) was applied to extract the 52 features contained in this dataset.