ESTIMATING THE LIMIT OF BIO-AEROSOL DETECTION WITH PASSIVE INFRARED SPECTROSCOPY

摘要

To investigate the detection limits of biological aerosols using passive infrared measurements, we have developed a computational model that relies on physics-based simulations to generate a statistical sample. The simulation consists of three principal models: an atmospheric turbulence model, a radiative transfer model and a target detection model. The turbulence model is used to generate microscale atmospheric variability. Resulting temperature and density profiles, along with custom aerosol profiles, are used to generate inputs for MODTRAN5, which produces simulated atmospheric spectral radiance. The simulated data is then analyzed by using an optimal detection algorithm and a hypothesis test, resulting in receiver operating characteristic (ROC) curves.