Performance Synthesis (Electro-Optical Sensors)

Analytical models are developed for evaluating and predicting the performance of observers augmented by electro-optical sensors when viewing images of comparatively simple geometry. These models provide a method of associating a signal-to-noise ratio with an image as a function of the image radiance and spatial dimension. Through psychophysical experimentation, the probability of the observers detecting certain images was established using images of known signal-to-noise ratio. By matching the sensor capability with the observer requirements, overall sensor system performance is predicted. The models are perfectly general and apply to all electro-optical sensors. Their use is illustrated by applying them to the prediction of a number of representative sensors including image dissectors, forward-looking infrared scanners, direct view light amplifiers, and low-light-level television. Methods of predicting the range capability of sensor-augmented observers are developed and applied to both active and passive systems. These models take into account properties of the scene, atmospherics, and level of target discrimination.