New Scene-Based Nonuniformity Correction Algorithm of Infrared Focal-Plane Arrays

Abstract

The technique for nonuniformity correction in infrared focal-plane arrays is a typical blind-estimation problem, and the infrared images without noises are not available. In order to correct the observed images in the presence of the fixed-pattern noise caused by pixel-to-pixel nonuniformity, it is necessary to estimate the corresponding true scene values. A new scene-based nonuniformity correction algorithm is developed, which can use the interframe-prediction technique to achieve the estimates of true scene values. Therefore, we can predict the true value of the next uncorrected frame through the current corrected one. With these scene estimates and the corresponding observed values through a given array, a line-fitting procedure is used to estimate the individual detector response parameters, which can then be used to correct the nonuniformity. Thus, with a frame-by-frame iteration, the response parameters will converge to the optimum values. The strength of the algorithm lies in its high convergent speed despite the level of nonuniformity, and its ability of reducing ghosting artifacts is noticeable. The performance of the algorithm is demonstrated by the use of an ideal infrared image sequence with simulated nonuniformity and an actual infrared image sequence with real nonuniformity.