基于场景的红外成像系统盲元检测及校正  

Blind pixel detection and correction of IR imaging system based on scene

作　　者：李召龙[1] 韩玉龙[2] 

机构地区：[1]海军航空工程学院控制工程系,山东烟台264001 [2]海军航空工程学院指挥系,山东烟台264001

出　　处：《激光与红外》2017年第4期465-469,共5页Laser & Infrared

基　　金：国家自然科学基金项目(No.61303192)资助

摘　　要：根据盲元的响应特性,给出了无需依赖黑体辐射定标的、基于时域平均野值提取的盲元检测算法,然后又提出了基于相关像素加权插值的盲元校正算法。详细探讨了盲元检测过程中阈值的设置方法,研究了参与加权估计盲元灰度值的像素数目对估计误差的影响。最后通过仿真实验对检测算法的准确性及校正算法的有效性进行了验证,实验结果表明:本文提出的检测算法结果准确,误检测率低。校正算法校正效果好,结果图像的RMSE低于两种常用校正算法。According to the response characteristics of the blind pixel,a blind pixel detection algorithm was given based on time domain-average and outliers-extraction,and it didn＇t depend on black-body radiationcalibration.Then,a blind pixel correction algorithm was put forwardbased on weighted interpolation.The setting method of threshold in blind pixel detection was discussed in detail,and in the estimation of blind pixel gray level,the influence of the pixel number on the estimation error was studied.Finally,the accuracy of the detection algorithm and the effectiveness of the calibration algorithm were verified by simulation experiments.The results show that the detection algorithm has high precision and low false alarm rate,and the correction algorithm has good effect.The RMSE of the results obtained with the proposed algorithm is less than that of two common algorithms.

关 键 词：盲元 响应特点 盲元检测 相关像素 

分 类 号：TN216[电子电信—物理电子学]