基于中值滤波的红外焦平面阵列非均匀性神经网络校正

统的神经网络校正算法存在收敛速度慢和校正精度低的缺点。当背景噪声较大时，它更难以获得令人满意的校正效果。 针对其不足之处， 提出一种基于中值滤波的红外焦平面阵列(IRFPA)非均匀性神经网络校正算法。该算法首先利用中值滤波对强噪声进行预处理，在此基础上 采用改进的神经网络校正算法对IRFPA非均匀性进行自适应校正。实验结果表明，该算法与传统的神经网络方法相比具有收敛速度快和校正精 度高等特点，并且使图像的峰值信噪比至少提高了10dB。

Neural Network Non-uniformity Correction Method for IRFPA Based on Median Filtering

Since the traditional neural network non-uniformity correction algorithm has lower convergence speed and lower correction precision, it is more difficult to obtain satisfactory correction result when high background noise is present. For this reason, a neural network method for correcting the non-uniformity of an infrared focal plane array (IRFPA) based on median filtering is proposed. In the method, the median filtering is first used to preprocess the high noise and then an improved neural network algorithm is used to correct the non-uniformity of the IRFPA adaptively. The experimental result shows that the algorithm has higher convergence speed and higher correction precision than the traditional method and can make the peak signal to noise ratio be improved by 10dB at least.