Female mosquito detection by means of AI techniques inside release containers in the context of a Sterile Insect Technique program

TL;DR

This study addresses the risk of releasing female Aedes mosquitoes in SIT programs by developing a quality-control workflow that detects female wingbeat sounds inside release containers. It combines audio processing with TRILL-based embeddings and two one-class anomaly detectors (iForest and OC-SVM) trained on male-only data to flag female presence as anomalies, using a four-microphone setup and 30-second clips parsed into 4-second chunks. The results show that detection performance varies with the day after sex-sorting, peaking on day seven (iForest achieving up to 87.5% accuracy for mixed containers and 81% for all-male containers), with an overall average around 73%. The work demonstrates a practical, noninvasive QC step that can enhance SIT reliability by reducing false negatives before releasing sterile males.

Abstract

The Sterile Insect Technique (SIT) is a biological pest control technique based on the release into the environment of sterile males of the insect species whose population is to be controlled. The entire SIT process involves mass-rearing within a biofactory, sorting of the specimens by sex, sterilization, and subsequent release of the sterile males into the environment. The reason for avoiding the release of female specimens is because, unlike males, females bite, with the subsequent risk of disease transmission. In the case of Aedes mosquito biofactories for SIT, the key point of the whole process is sex separation. This process is nowadays performed by a combination of mechanical devices and AI-based vision systems. However, there is still a possibility of false negatives, so a last stage of verification is necessary before releasing them into the environment. It is known that the sound produced by the flapping of adult male mosquitoes is different from that produced by females, so this feature can be used to detect the presence of females in containers prior to environmental release. This paper presents a study for the detection of females in Aedes mosquito release vessels for SIT programs. The containers used consist of PVC a tubular design of 8.8cm diameter and 12.5cm height. The containers were placed in an experimental setup that allowed the recording of the sound of mosquito flight inside of them. Each container was filled with 250 specimens considering the cases of (i) only male mosquitoes, (ii) only female mosquitoes, and (iii) 75% males and 25% females. Case (i) was used for training and testing, whereas cases (ii) and (iii) were used only for testing. Two algorithms were implemented for the detection of female mosquitoes: an unsupervised outlier detection algorithm (iForest) and a one-class SVM trained with male-only recordings.

Figures (5)

• Figure 1: Spectrogram of (top) 250 male mosquitoes, (middle) 250 female mosquitoes and (bottom) 250 mosquitoes with 75% male and 25% female ratios. Spectrogram of original audio signals subsampled at 8 kHz sampling rate.
• Figure 2: (a) release container (b) inside container (c) open recording setup (d) recording setup
• Figure 3: Recording procedure carried out. This framework was carried out twice per day. In every interaction with a container, 8 audios of 30 seconds were recorded. The first recording container corresponds to the male container that have been used as a training set.
• Figure 4: Comparison of two spectrograms of the same container of males. The upper spectrogram corresponds to the first recorded clip (as soon as the container is introduced into the setup) and the second one once the container has been in the setup for 3.5 minutes.
• Figure 5: (a) Spectrogram of a 25%/75% female/male container audio of the 6th day since sex-sorting (b) iForest system output (c) OCSVM system output. The prediction point at $t=n$ seconds is the prediction for $t=n$ seconds to $t=n+4$ seconds audio segment.