A Neural Network for Nonuniformity and Ghosting Correction of Infrared Image Sequences

Abstract

In this paper, an adaptive scene-based nonuniformity and ghosting artifacts correction algorithm for infrared image sequences is presented. The method simultaneously estimates detector parameters and carry out the non-uniformity and ghosting artifacts correction based on the retina-like neural network approach. The method incorporates the use of a new adaptive learning rate rule into the estimation of the gain and the offset of each detector. This learning rule, together with the consideration of the dependence of the detector’s parameters on the retinomorphic assumption used for parameter estimation, may sustain an efficient method that could not only increase the original method’s ability for estimating the non-uniformity noise, but also increase the capability of mitigating ghosting artifacts. The ability of the method to compensate for nonuniformity and reducing ghosting artifacts is demonstrated by employing several infrared video sequences obtained using two infrared cameras.

This work was partially supported by Grant Milenio ICM P02-049. The authors wish to thank Ernest E. Armstrong (OptiMetrics Inc., USA) and Pierre Potet (CEDIP Infrared Systems, France) for collecting the data, and the United States Air Force Research Laboratory, Ohio, USA.