Ground-target infrared signature uncertainties and their effect on simulation and validation

摘要

Abstract: Data and analyses demonstrating the variability and uncertainties in infrared (IR) ground target signatures are presented. The uncertainties are due to a variety of factors ranging from environmental effects to differences in vehicle configurations. Caution must be exercised when using predictive models for simulations because these models are usually pristine and present repeatable signatures for a given set of inputs. Actual vehicle signatures show the effects of wear and tear, aging, poor maintenance, etc., and these effects will vary from vehicle to vehicle. Vehicles encountered in real-life often have a variety of crew- specific signature components that will affect the signature of a vehicle as well, such as stowage of supplies or spare parts. It is important therefore to develop the concept of a `representative' target and take into consideration the expected variations from a baseline signature. It is common in an infrared scene simulation to have a single signature for a given type of vehicle when more than one of the vehicles is present in the scene at the same time. Limiting the target data in this manner can lead to biased results as observers and algorithms can memorize a particular signature. To avoid this, ground target signatures used for training simulations and algorithm development should incorporate variability in the target signatures. Varying ground target signatures in this manner will provide for more realistic sensor performance assessment, training, and algorithm development. The primary signature factor affecting training and algorithm development will be vehicle configuration. Model developers often use temperature deltas when assessing the fidelity of an IR signature model. When validating an infrared signature model, whether it is digital or a target surrogate, the model developer should take into account the uncertainty in the target signature caused by measurement errors and target surface optical variations. Portions of a vehicle where paint has been removed have greatly reduced emissivity and often the reflected radiance will be from the sky. This can lead to temperature errors of tens of degrees Celsius. As paint ages or get dirty its optical characteristics change which can also cause infrared signature variations. All of these surface factors (and more) lead to a general uncertainty in the IR signature of a ground vehicle.!3