Multi-model Adaptive Estimation for Nonuniformity Correction of Infrared Image Sequences

Abstract

This paper presents a multiple model parallel processing technique to adaptively estimate the nonuniformity parameters of infrared image sequences. The approach is based on both an optimal recursive estimation based on a fast form of the Kalman filter, and a solution for the uncertainties on the system model by running a bank of those estimators in parallel. The residual errors of these estimators are used as hypothesis to test and assign the conditional probabilities of each model in the bank of the Information form of the Kalman filter. The conditional probabilities are used to calculate weighting factors for each estimation and to compute the final system state estimation as a weighted sum. Then, the weighting factors are updated recursively from one to another sequence of infrared images, providing to the estimator a way to follow the dynamic of the scene recorded by the infrared imaging system. The ability of the scheme to adaptively compensates nonuniformity in infrared imagery is demonstrated by using real infrared image sequences.

Topic: Image and Video Processing and Analysis

This work was partially supported by the ‘Fondo Nacional de Ciencia y Tecnología’ FONDECYT of the Chilean government, project number 1020433 and by Grant Milenio ICM P02-049. The authors wish to thank Ernest E. Armstrong (OptiMetrics Inc., USA) for collecting the data, and the United States Air Force Research Laboratory, Ohio, USA.