Moving scene based nonuniformity correction algorithm

Abstract

An improved nonuniformity correction (NUC) algorithm based on moving scenes has been brought forward in this paper. It is developed for nonuniformity correction of infrared focal plane arrays (FPA). Matching technology of images and bidirectional updating strategy on correction coefficients are used in this algorithm. Compared with general scenebased nonuniformity correction algorithm, this algorithm reduces requirements on statistic characteristics of scenes and then has faster convergence speed. First, two images from neighbor frames are matched. Shift parameters of images between two neighbor frames are estimated. After that, bidirectional updating strategy on correction coefficients is used to realize adaptive updating of correction coefficients. In the end, Updated correction coefficients are then used to correct nonuniformity of images. Utilizing matching technology won’t induce blur of scene when correction coefficients are updated. In the meantime, utilizing bidirectional updating strategy on correction coefficients can ensure that coefficient of every pixel is updated at least one time in each frame. Using collected image sequences, simulation is carried out to evaluate this correction algorithm. Results demonstrate that the algorithm presented in this paper exhibits excellent correction effect over general scene-based nonuniformity correction algorithms.