T. Saha, B.P. Thomas
New Jersey Institute of Technology,
United States
Keywords: entomology, insects, laser sensing, optical remote sensing, environment, polarization, MFCC, machine learning, classification, biomass monitoring
Summary:
Insects play essential roles in ecosystems as pollinators, decomposers, and food sources for countless species, yet their global decline in abundance and diversity has become a major environmental concern. The ability to monitor their population is key to addressing this crisis, but conventional approaches, principally traps, are labor-intensive, destructive, and limited in both temporal and spatial coverage. Optical and laser-based methods offer a compelling alternative, capable of continuously monitoring insect activity without capture, and providing quantitative information on abundance, biomass, and even species traits through light–insect interactions. Our work focuses on the development and application of laser-based entomological photonic sensors that exploit backscattering and extinction phenomena to characterize flying insects in both laboratory and field environments. Using near-infrared continuous-wave lasers, these instruments record the optical signatures of individual insects crossing the beam, from which multiple predictor variables are extracted, including wingbeat frequency, spectral and polarimetric cross-sections, and the harmonic structure of the signal. In controlled experiments, we demonstrated that these optical features allow for the discrimination of species, sex and sometimes payloads using machine learning algorithms. Building on these laboratory results, we developed the Entomological Bistatic Optical Sensor System (eBoss), a compact and eye-safe bistatic instrument designed for autonomous outdoor deployment. Deployed continuously for multiple seasons, the eBoss has observed millions of flying insects, measuring both aerial insect density (number of insects per cubic meter) and biomass density (mass per cubic meter) with a one-minute temporal resolution. This unprecedented dataset reveals pronounced diurnal and seasonal variations, capturing daily activity peaks at dawn and dusk, as well as the onset of flight activity as a function of ambient temperature, humidity, and wind speed. Subsequent analysis established empirical models linking wingbeat frequency to air temperature, enabling real-time temperature correction and more accurate taxonomic clustering. Comparisons with traditional light-trap collections confirmed strong correlations in abundance trends, while highlighting the advantages of laser-based sensing in temporal resolution, sampling volume, and non-destructive operation. Furthermore, advanced signal-processing approaches, such as Mel-Frequency Cepstral Coefficients (MFCCs), have been applied to optical data to account for insect orientation during flight, further refining estimates of body cross-section and mass. Altogether, these developments position photonic sensing as a powerful approach for ecological monitoring. By coupling high-precision laser instrumentation with data-driven analysis, it becomes possible to continuously quantify insect populations over extended periods, capturing both fast behavioral dynamics and long-term environmental trends. This work demonstrates how laser-based remote sensing can expand the reach of entomology, transforming insects from elusive targets into quantifiable optical scatterers, and how photonics can contribute directly to addressing one of the most pressing biodiversity challenges of our time.