红外成像系统非均匀性快速校正方法

针对传统基于场景的非均匀性校正方法计算量大的缺点，提出一种快速校正方法。通过对图像序列进行分析，得出非均匀性的表现形式可看作是在理想视频上叠加一个固定图案噪声这一结论。首先对前 k 帧图像求取均值，得到包含非均匀性而且背景灰度趋于一致的均值图像；然后提取非均匀性噪声矩阵，将原始非均匀性图像与噪声矩阵做差，便可实现非均匀性的快速校正。利用仿真图像序列与实拍视频对方法进行验证，实验结果表明提出的方法能够有效对非均匀性进行校正，而且计算量较小，具有更快的校正速度。

Fast nonuniformity correction method for IR imaging system

To solve the problem of the large calculation quantity when correcting the nonuniformity using the tradition-al scene-based method, a fast correction method was proposed. Through the analysis of the image sequences, the results showed that the nonuniformity can be seen as that a fixed pattern noise superimposed on the ideal video. Firstly, the mean image of the first k frames was obtained. Then the nonuniformity noise matrix was extracted. The original image could be corrected by subtracting the noise matrix. The method was validated by using simulated video and real video. Experimental results show that the method can correct the nonuniformity with smaller calculation quantity and faster speed.