Appl. Phys. B 124, 46 (2018) https://doi.org/10.1007/s00340-018-6917-x

According to the World Health Organization, mosquitoes are the greatest animal threat to human health, spreading diseases such as malaria, dengue, yellow fever and the Zika virus, which are collectively responsible for several million deaths every year. Mosquito distributions are closely monitored but methods often rely on human involvement. Now, Adrien Genoud and colleagues at the New Jersey Institute of Technology and Rutgers University in the United States, have developed an infrared optical remote sensing system for characterizing the gender and species of flying mosquitoes. The authors recorded 427 signals from a sample containing males and females of three different mosquito species. The authors attempted to use wing-beat frequency as a predictor for gender alone, as well as species and gender. A continuous-wave laser diode with 1,320-nm wavelength and 3.6 W power illuminated moquitoes 3–4 m away. InGaAs amplified photodetectors detected the light scattered as a mosquito transited the illumination; the modulation of the backscatter due to the insect body is easily separable from the faster (100 Hz to 1,000 Hz) wing beating. Using a Bayesian classification method, gender alone is discriminated with 96.5% accuracy. However, when gender and species are characterized, overall accuracy of classification drops to 62.3%. In real samples, even more species are likely to be present and additional (remotely measurable) variables, such as optical cross-section or depolarization ratio, may be needed to improve accuracy.