固定图案噪声是红外焦平面阵列的一种固有噪声，该噪声在低温场景尤其突出，影响了探测器的成像质量。基于探测器读出电路结构，理论分析了相关双取样电路（CDS）对焦平面阵列非均匀性的影响，并通过数据仿真给出了CDS噪声的空间和时间分布特征。针对CDS噪声特征，提出将迭代原理和黑体定标校正相结合的噪声滤波算法。实验结果表明，算法有效抑制了CDS电路噪声，提高了非均匀性校正的精度，具有较强的实用价值。

Fixed pattern noise, an inherent noise of infrared focal plane array (IRFPA), is serious in the low-temperature scene and affects detector imaging quality. The influence of Correlated Double Sampling (CDS) to the nonuniformity in IRFPA is theoretically analyzed based on the structure of internal readout circuit, moreover, the space and time distribution features of CDS noise are described according to simulation data. Aiming at these features, a noise filtering algorithm using the combination of iterative theory and blackbody calibration is proposed. Experimental results show the proposed algorithm can effectively eliminate CDS circuit noise, increase nonuniformity correction precision and have strong utility value.