SBIR-STTR Award

Automatic target recognizer (atr) computers are causing a revolution in the design of forward looking infrared (flir) sensors. Historically, flirs have under-sampled because of the small number of detectors available on the focal plane arrays imagery. However, fpa detector densities have improved and atrs are sensitive to aliasing. As a result, sampling rate is now considered to be a critical design parameter. New flir design and performance figures of merit cannot be verified since all existing sensors are limited to one sample rate. In phase i of this effort, an innovative design of a variable sample rate laboratoryflir will be developed that, while not practical for a operational system, will be ideal for the proposed laboratory unit. Combining a starting fpa with a step-stare two axis scanner will be studied. The flir will generate tv line rates so that the data can be displayed to an operator without image tearing artifacts and detector and aperture sizes will be selected to provide anti-aliasing filters. Phase 2 will consist of the final design, flir fabrication and the collection of datato verify the figures of merit.

Forward Looking Infrared (FLIR) sensors, traditionally designed for use by human operators, are now being developed for automatic target recognition (ATR) applications in which computers are responsible for target acquisition tasks. Advanced FLIR design and performance figures of merit are attained through models that have not been verified because of the inherent lack of flexibility in existing FLIR systems. The lack of verified performance models is delaying the progress of numerous ATR systems now under development. A laboratory FLIR system that facilitates variable performance conditions in support of this research is the subject of this proposal. The proposed Variable Parameter FLIR (VP FLIR) system is an investigative tool designed to explore the effects of temporal and spatial sampling on focal plane array (FPA) imagery, to verify existing software models, and to characterize the sensitivity of ATR computers to aliasing effects. In Phase II of this effort, the VP FLIR system, including a 128 x 128 INSB focal plane array, optics, and imaging electronics, will be assembled. Options to the baseline program comprise other FPA configurations and a single instruction multiple data (SIMD) class parallel processor to facilitate direct execution of ATR algorithms on FPA data in real time.