随着热像系统时域噪声等效温差(NETD)的不断下降，残余固定图形噪声(RFPN)的不稳定性成为制约系统性能，特别是制约系统识别能力的一个关键因素。通过分析热像系统响应非均匀性产生的机理，介绍一种扫描型热像系统获得场景平均辐照度点的方法，实现了以场景平均辐照度点为校正基准点的动态校正。采用该方法研制的热像系统在(0～50)C场景动态范围内可使系统的RFPN小于15mK。通过样机冬、夏两季外场试验以及各种环境试验，验证了该校正方法在热像系统温度发生变化以及场景辐照度发生变化时具有良好的适应性。

Residual fixed pattern noise (RFPN) is becoming a major limitation for system recognition capability due to the continuous decrease of temporal NETD in thermal imager.Based on the analysis of non-uniformity mechanism in thermal imager,a method to achieve scene averaged irradiance in scanning thermal imager is introduced and the scene averaged irradiance is used as a reference to implement dynamic correction. The achieved system RFPN is less than 15 mK in the scene dynamic range of (0～50) C. The prototype was field tested in various environmental conditions. It was proved that such design enables thermal imager performance immune to the temperature change of thermal imager as well as the change of scene irradiance.