Field evaluation of an automated mosquito surveillance system which classifies Aedes and Culex mosquitoes by genus and sex.

BACKGROUND: Mosquito-borne diseases are a major concern for public and veterinary health authorities, highlighting the importance of effective vector surveillance and control programs. Traditional surveillance methods are labor-intensive and do not provide high temporal resolution, which may hinder a full assessment of the risk of mosquito-borne pathogen transmission. Emerging technologies for automated remote mosquito monitoring have the potential to address these limitations; however, few studies have tested the performance of such systems in the field. METHODS: In the present work, an optical sensor coupled to the entrance of a standard mosquito suction trap was used to record 14,067 mosquito flights of Aedes and Culex genera at four temperature regimes in the laboratory, and the resulting dataset was used to train a machine learning (ML) model. The trap, sensor, and ML model, which form the core of an automated mosquito surveillance system, were tested in the field for two classification purposes: to discriminate Aedes and Culex mosquitoes from other insects that enter the trap and to classify the target mosquitoes by genus and sex. The field performance of the system was assessed using balanced accuracy and regression metrics by comparing the classifications made by the system with those made by the manual inspection of the trap. RESULTS: The field system discriminated the target mosquitoes (Aedes and Culex genera) with a balanced accuracy of 95.5% and classified the genus and sex of those mosquitoes with a balanced accuracy of 88.8%. An analysis of the daily and seasonal temporal dynamics of Aedes and Culex mosquito populations was also performed using the time-stamped classifications from the system. CONCLUSIONS: This study reports results for automated mosquito genus and sex classification using an optical sensor coupled to a mosquito trap in the field with highly balanced accuracy. The compatibility of the sensor with commercial mosquito traps enables the sensor to be integrated into conventional mosquito surveillance methods to provide accurate automatic monitoring with high temporal resolution of Aedes and Culex mosquitoes, two of the most concerning genera in terms of arbovirus transmission.

Automated detection of the yellow-legged hornet (Vespa velutina) using an optical sensor with machine learning.

BACKGROUND: The yellow-legged hornet (Vespa velutina) is native to Southeast Asia and is an invasive alien species of concern in many countries. More effective management of populations of V. velutina could be achieved through more widespread and intensive monitoring in the field, however current methods are labor intensive and costly. To address this issue, we have assessed the performance of an optical sensor combined with a machine learning model to classify V. velutina and native wasps/hornets and bees. Our aim is to use the results of the present work as a step towards the development of a monitoring solution for V. velutina in the field. RESULTS: We recorded a total 935 flights from three bee species: Apis mellifera, Bombus terrestris and Osmia bicornis; and four wasp/hornet species: Polistes dominula, Vespula germanica, Vespa crabro and V. velutina. The machine learning model achieved an average accuracy for species classification of 80.1 ± 13.9% and 74.5 ± 7.0% for V. velutina. V. crabro had the highest level of misclassification, confused mainly with V. velutina and P. dominula. These results were obtained using a 14-value peak and valley feature derived from the wingbeat power spectral density. CONCLUSION: This study demonstrates that the wingbeat recordings from a flying insect sensor can be used with machine learning methods to differentiate V. velutina from six other Hymenoptera species in the laboratory and this knowledge could be used to help develop a tool for use in integrated invasive alien species management programs. © 2022 The Authors. Pest Management Science published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.

A novel optical sensor system for the automatic classification of mosquitoes by genus and sex with high levels of accuracy.

BACKGROUND: Every year, more than 700,000 people die from vector-borne diseases, mainly transmitted by mosquitoes. Vector surveillance plays a major role in the control of these diseases and requires accurate and rapid taxonomical identification. New approaches to mosquito surveillance include the use of acoustic and optical sensors in combination with machine learning techniques to provide an automatic classification of mosquitoes based on their flight characteristics, including wingbeat frequency. The development and application of these methods could enable the remote monitoring of mosquito populations in the field, which could lead to significant improvements in vector surveillance. METHODS: A novel optical sensor prototype coupled to a commercial mosquito trap was tested in laboratory conditions for the automatic classification of mosquitoes by genus and sex. Recordings of > 4300 laboratory-reared mosquitoes of Aedes and Culex genera were made using the sensor. The chosen genera include mosquito species that have a major impact on public health in many parts of the world. Five features were extracted from each recording to form balanced datasets and used for the training and evaluation of five different machine learning algorithms to achieve the best model for mosquito classification. RESULTS: The best accuracy results achieved using machine learning were: 94.2% for genus classification, 99.4% for sex classification of Aedes, and 100% for sex classification of Culex. The best algorithms and features were deep neural network with spectrogram for genus classification and gradient boosting with Mel Frequency Cepstrum Coefficients among others for sex classification of either genus. CONCLUSIONS: To our knowledge, this is the first time that a sensor coupled to a standard mosquito suction trap has provided automatic classification of mosquito genus and sex with high accuracy using a large number of unique samples with class balance. This system represents an improvement of the state of the art in mosquito surveillance and encourages future use of the sensor for remote, real-time characterization of mosquito populations.