从红外相机自身特性和使用角度两方面对靶场红外图像序列中不均匀(花纹)背景产生的原因进行了剖析，分析了传统红外图像非均匀校正算法在处理靶场序列红外图像上的优缺点，提出在目标跟踪过程中随着场景变化，基于实时线性标定的非线性红外图像的校正的新方法，克服了传统方法的弊端，实时确定了校正增益系数和校正因子，消除了序列红外图像中不均匀背景。通过对含有弱小目标的靶场实际序列红外图像进行仿真验证表明，新方法去除了图像固定图案噪声，消除了探测器坏元的影响，输出了理想的序列红外图像。

This article takes apart the reason which a serial of the IR images comes into being non-uniformity(speckle) background in the light of the characteristic and using of IR camera. It analyses the merit and defect on traditional non-linear correction methods dealing with a serial of the IR images on the shooting range. It puts forward the new method of non-linear correction of IR images based on the linear calibration in real time serving background changing in course of tracing targets. The new method overcomes traditional abuse and obtains real-time coefficient of gain correction and correction factor. It eliminates non-uniformity background of a serial of the IR images. The experimental results using puny targets of a serial of the IR images on the shooting range shows that the new method not only wipes off the noise of the fixed-pattern and eliminates bad-cell of sensor, but also outputs a serial of the perfect IR images.