Research On Nonuniformity Correction Algorithm For Infrared Detector

Non-uniformity,which is the fixed-pattern noise of infrared image,comes with infrared focal plane arrays(IFPA).Non-uniformity has a bad effect on infrared image.There are mainly two kinds of algorithm,blackbody-based and scene-based,for non-uniformity correction of infrared image in the world at present.Firstly,the composition of infrared image system and the development of infrared detector are analyzed in the dissertation.And the category and the development of non-uniformity correction algorithm for infrared image are also present in the dissertation.Secondly,the definition and source of non-uniformity and the response model of infrared detector are introduced in the dissertation.Meanwhile,the basic principle and processing steps of kinds of typical algorithm are analyzed.And also merits and demerits are analyzed in detail for every algorithm.Then,the basic principle,processing steps and merits and demerits of kinds of image registration and non-uniformity correction algorithm based on image registration are introduced in the dissertation.The impact of moving target on the algorithm is analyzed.The problem of non-uniformity correction of infrared image,which contains moving target,can be solved through replacing the radiance of the object in the last frame with that of the same location in the current frame.Tests with synthetic and true infrared scene sequences indicate that the proposed algorithm can solve the problem of non-uniformity correction for infrared image with moving target efficiently.At last,the way,direction and range of deformation target are analyzed.And the principle of judging a deformation target is given.And then,the way how deformation target can be replaced is obtained.Tests with synthetic infrared sequences indicate that the proposed algorithm can solve the problem of non-uniformity correction for infrared image with deformation target efficiently.