Abstract Entomological photonic sensors enable continuous, automated and non-invasive monitoring of flying insects by recording optical signals of insects transiting in their field-of-view. These instruments can observe extremely large numbers of insects and provide ecological measurements such as aerial density insect/m 3 and biomass density mg/m 3 with temporal resolution down to a minute and minimal downtime. Nevertheless, their taxonomic resolution is limited; since deriving reliable taxonomic signatures from optical signals is difficult, knowing insects represent the most species-rich group of organisms. In this study, we focus on the influence of body orientation during transit, a factor that directly impacts how the body and wings contribute to the recorded signal. Using numerical simulations based on a 3D model of an Apis mellifera (honeybee), and laboratory experiments with Musca domestica (housefly), we analyze extinction signals using Mel-Frequency Cepstral Coefficients. They characterize the time-frequency structure of the waveform, caused by changes in orientation of an insect as it crosses the beam. A Gaussian Process regression model trained on these coefficients predicts symmetry-reduced orientation angles with high accuracy. By correcting for orientation, we can retrieve a more accurate estimate of the insect’s body optical cross-section. This correction may lead to improved identification and mass estimation. Our simulation results show that orientation effects can alter apparent cross-sections by up to 83% in this particular case, however this may be even more pronounced for insects with elongated bodies, underscoring the importance of accounting for them in biomass calculations obtained from photonic sensors. These findings demonstrate that orientation-sensitive signal analysis can refine predictor variables and improve the reliability of photonic sensors in providing ecological metrics beyond abundance, paving the way toward higher taxonomic resolution.
Saha et al. (2026) studied this question.