Hardware Implementation of a System Classifying the Optoacoustic Signature of Insects Wing-Flap

Abstract

In this paper we present a standalone hardware-implemented system that performs all signal processing stages necessary to classify the species of insects based on their wing-flap imprint as they fly. The recognizer classifies insect’s wing-beat recordings from an array of phototransistors receiving light from an infrared LED or laser. The wing-beat recording is based on the interruption of the emitted light due to the partial occlusion from insect's wings as they fly in typical traps. The classification module and the optoelectronic sensor are inserted in typical insect traps and perform detection, counting, recognition and transmission of results automatically. This work emphasizes the hardware implementation of the classifier performing all steps starting from the analog input to the final transmission of results. We give all necessary implementation details needed to construct all circuit boards. We show recognition results for four insect species. We believe that once optimized the optoacoustic sensor and the standalone recognizer has the potential to revolutionize the way insect monitoring is carried out for a series of insects with large economic impact.