Static Scene Statistical Non-Uniformity Correction.

摘要

Non-Uniformity Correction (NUC) is required to normalize imaging detector Focal-Plane Array (FPA) outputs due to differences in the end-to-end photoelectric responses between pixels. Currently, multi-point NUC methods require static, uniform target scenes of a known intensity for calibration. Conversely, scene-based NUC methods do not require a priori knowledge of the target but the target scene must be dynamic. The new Static Scene Statistical Non-Uniformity Correction (S3NUC) algorithm was developed to address an application gap left by current NUC methods. S3NUC requires the use of two data sets of a static scene at different mean intensities but does not require a priori knowledge of the target. The S3NUC algorithm exploits the random noise in output data utilizing higher order statistical moments to extract and correct fixed pattern, systematic errors. The algorithm was tested in simulation and with measured data and the results indicate that the S3NUC algorithm is an accurate method of applying NUC. The algorithm was also able to track global array response changes over time in simulated and measured data. The results show that the variation tracking algorithm can be used to predict global changes in systems with known variation issues.