Acoustic detection and localisation system for Hylotrupes bajulus L. larvae using a MEMS microphone array

Abstract

A novel system for acoustic detection of the presence of xylophagous insect larvae inside structural timber beams is presented. It is based on an extensive array of MEMS microphones that allows the acoustic detection and localisation of the larvae when they are active. In a first phase, the activity of the larvae is continuously detected by means of frequency filtering and a sliding energy estimator, and after that, a set of short-duration segmented signals is generated, which obtains the spatial localisation of the larvae, by means of a shaping algorithm based on delay-sum beamforming techniques. The tests carried out demonstrate that it is possible to detect and locate multiple larvae of Hylotrupes bajulus L. inside structural-sized pieces of wood of Pinus syilvestris L., as well as their internal trajectory. In the future, the system could address the identification of the specific type of xylophage responsible for the deterioration by using machine learning or equivalent techniques, based on the temporal and frequency information of the detected sound events. The aim of this work is to control unintentional infestions in the international timber trade, in the assembly and the use of infested timber and, in all cases, to be able to carry out selective, targeted and localised treatments and to verify their success.