Study On Scene-based Non-uniformity Correction Algorithm For IRFPA

The IRFPA has a simple and compact structure.It also has high SNR and sensitivity.Therefor it is one of the most promising infrared imaging devices.However,because of the short of materials and the immature manufacturing technology,there is nonuniformity problem in the infrared imaging system,which affects the quality of infrared imaging system seriously.Hence,it is necessary to correct the nonuniformity.The scene-based algorithms do not need the blackbody calibration and do not affect the work of the system.Because there are adaptive,the scene-based algorithms now have become the focuses of the research field.Therefor two scene-based nonuniformity correcting algorithms are proposed in this thesis.Firstly,the traditional temporal high-pass filter algorithm is studied.Experiment shows that there are ghosts appearing in the corrected images and some real scene vanishs away.Based on this,a nonuniformity correction algorithm combining the spatial filter with temporal filter is proposed.First the original nonuniformity image is divided into two parts one of which is the spatial high frequency information part and the other is the low part by the global filter.Then the correction parameters are estimated from the high frequency information.At last,the feasibility of the algorithm is proved with simulation result.Then,the standard moment match algorithm is studied.Experiment shows that the corrected images have serious grayscale distortion.On this basis,an adaptive correction algorithm based on moment match is proposed.Firstly,the nonuniformity image is filtered by the bilateral filter.Then the correction parameters are estimated from the filtered image.At last,the pretreatment image is iterated in the time domain.Experiment proves that the proposed algorithm is fast and can effectively correct the nonuniformity of the uncooled IRFPA without bringing in the distortion effect and the “ghoat” artifact.