改进卡尔曼滤波的IRFPA非均匀性校正算法  

Improved Kalman filtering algorithm for nonuniformity correction in IRFPA

作　　者：代少升[1] 马东鸽[1] 

机构地区：[1]重庆邮电大学信号与信息处理重庆市重点实验室,重庆400065

出　　处：《华中科技大学学报（自然科学版）》2012年第10期13-16,共4页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61071196);信号与信息处理重庆市市级重点实验室建设项目(2009CA2003);重庆市自然科学基金资助项目(2009BB2287;2010BB2398;2010BB2411);重庆市科技攻关计划资助项目(2011AB2008)

摘　　要：提出一种改进的卡尔曼滤波的IRFPA非均匀性校正算法,该算法使用求逆引理对传统迭代公式进行求逆变换,使用误差协方差逆阵形式实现迭代过程,使算法计算复杂度由nk×nk阶矩阵求逆降低至nk维对角矩阵求逆,简化了传统卡尔曼滤波算法的递推过程,提高了实时性能.通过对实际采集的红外图像序列进行仿真实验,结果表明:该算法在继承传统校正算法校正精度的同时,兼顾提高了算法校正速度,运行效率可以提高至2倍.Kalman filtering approach to non-uniformity correction in infrared focal plane array （IRFPA） was improved. In the algorithm, matrix inverse lemma was used to inverse transformation of the traditional iterative formulas, with inverse error covariance form. The traditional algorithm in- volves the inversion of the matrix, and the improved algorithm involves the inversion of the dimension diagonal matrix, so the improved algorithm simplifies the recursive process and improves the real-time performance. Meanwhile, the algorithm was tested by simulating the real infrared image sequences. Results show that the correction in the proposed algorithm reaches twice than the one in traditional al- gorithms at the same correction accuracy.

关 键 词：红外焦平面阵列 卡尔曼滤波 非均匀校正 矩阵求逆引理 误差协方差 

分 类 号：TP215[自动化与计算机技术—检测技术与自动化装置]