基于自适应滤波的红外焦平面阵列非均匀性校正算法  

Nonuniformity correction algorithm for infrared focal plane arrays based on adaptive-filter

作　　者：周慧鑫[1] 拜丽萍[1,2] 秦翰林[1] 赖睿[3] 王炳健[1] 刘群昌[1] 

机构地区：[1]西安电子科技大学技术物理学院,西安710071 [2]中国电子科学研究院,北京100041 [3]西安电子科技大学微电子学院,西安710071

出　　处：《强激光与粒子束》2009年第11期1611-1614,共4页High Power Laser and Particle Beams

基　　金：教育部科学技术研究重点项目(108114)

摘　　要：红外探测器响应漂移特性会降低红外焦平面阵列(IRFPA)非均匀性校正的精度。针对该问题提出了一种基于场景的IRFPA非均匀性校正算法。该算法利用所获得的序列成像场景信息,采用一种基于快速自适应滤波器的最优化递归估计方法来获得非均匀性校正参数,并利用当前的成像信息来更新校正参数,以此降低探测器响应漂移特性对非均匀性校正的影响。算法仿真实验显示,对非线性参数为26.12%的同一图像,使用该算法、两点校正算法和卡尔曼滤波校正算法校正1 h后,可分别将非线性参数降至1.856%,3.122%和1.893%,说明该算法可获得稳定而较好的非均匀性校正效果。The performance of nonuniformity correction（NUC） would be deteriorated by time shift in the response of detectors in infrared focal plane arrays（IRFPAs）. Therefore, a scene-based NUC algorithm is presented. In this algorithm, an adaptive filter bank is applied to estimate NUC parameters for every detector in IRFPAs. And the parameters are renewed by utilizing current imaging information to reduce the influence imposed by response drift on NUC. For an image with a nonuniformity of 26.12%, nonuniformities after 1 h correction are 3. 122% for two-point NUC algorithm, 1. 893% for Kalman-filter NUC algorithm and around 1. 856% for the presented algorithm, which means the presented algorithm can achieve the best NUC effect steadily.

关 键 词：物理电子学 红外焦平面阵列 非均匀性校正 自适应滤波 

分 类 号：TN81[电子电信—信息与通信工程]