高灵敏度热红外成像系统的灵敏度是最重要的指标。经过简单非均匀性校正的红外图像已不能满足高灵敏度场合的需要，需要后续对红外图像条纹噪声进行处理，以提高红外图像的灵敏度水平。为了进一步提高热红外成像系统的应用效果，全面分析了系统的图像噪声模型，提出并实施了一种基于硬件系统的条纹噪声去除方法，通过一套自行研制的制冷型热红外成像系统的实测数据实际验证了提出方法的实际效果，由非均匀性噪声引起的 NETD从 26 mK降到 19 mK。

Sensitivity is a key index for high resolution & high sensitivity thermal infrared imaging system. Such noise cannot be removed by any routine NUC algorithm, the IR image after NUC cannot satisfy the demand in any high sensitivity case, because there is still some non-uniform spatial noise in the IR image after NUC. To improve the application effect of the thermal infrared imaging system, the system image noise model was comprehensively analyzed, various methods for inhibiting the image spatial noise after image nonuniformity correction were taken, the actual effect of these methods was verified by the data measured from a series of independently designed refrigerating thermal infrared imaging system, and the NETD caused by the nonuniform noise was reduced from 26 mK to 19 mK.