Characterization of infrared imaging performance within a general statistical framework for environmental impacts on battlefield signals and sensing

摘要

The Environmental Awareness for Sensor and Emitter Employment (EASEE) software models the impacts of terrain and weather on a diverse range of battlefield sensing systems. The goal is to provide mission planning tools that realistically capture complex environmental factors impacting sensor performance, yet are simple enough for users with little specialized training. This paper describes incorporation of infrared (IR) modeling into EASEE, and the subsequent challenges of supporting imaging sensors within a framework that had previously evolved primarily for non-imaging sensors, such as acoustic and seismic. The design requires independently interchangeable modules for signature generation, propagation, and signal processing. Sensor performance metrics, such as probability of detection, are characterized statistically rather than through simulation of actual images. Some key enhancements needed to support imaging sensors were: (1) geometric models for targets, (2) packaging of multiple attributes representing target image properties (radiance, projected area, and spatial spectrum), (3) explicitly distinguishing between signals for the background, target of interest, and nuisance targets, and (4) calculation of apparent temperature differences (as opposed to incoherent energy summation). Target signatures are generated using MuSES (Multi-Service Electro-optic Signature), whereas the IR background properties are generated using FASST (Fast All-Season Soil STrength) and numerical weather prediction models. Propagation is handled primarily with MODTRAN (MODerate resolution atmospheric TRANsmission), although simpler models such as a line-of-sight calculation can also be employed. The Johnson criteria were added to the available library of detection algorithms.