基于灰度归一化的HJ-1A星HSI图像条带噪声去除方法

A Method to Destripe HJ-1A HSI Data Based on Nomalized Grey Level

针对HJ-1A星HSI图像上的条带噪声,在分析传统条带噪声去除方法的基础上,提出了一种新的基于灰度归一化去除条带噪声的方法。首先,设计孤立点检测窗口,将随机斑点噪声与条带噪声分离;然后,利用灰度归一化方法建立图像各列像元灰度级与基准列像元灰度级对应的查找表,根据查找表对HSI图像条带噪声进行消除。实验结果表明,该方法在很好地保留图像光谱信息的情况下,能够有效地消除图像条带噪声和随机斑点噪声。

The first national self-developed Hyper-Spectral Imaging（HSI） sensor was aboard on HJ-1A satellite which was launched on March 30,2009 successfully.The HSI data quality was influenced by the stripe noise in the first 20 spectral bands severely.It is an urgent need to study the method to destripe the HSI data.In this paper,a new destriping method was developed based on an analysis of the main causes and characteristics of the stripe noise in the hyperspectral images.The method firstly devised a filter window to separate the random noise and stripe noise.Secondly,a look-up table between the gray level of each column and that of the standard column was calculated to destripe the stripe noise of the HSI data.The results indicate that this method can remove the stripe noise and random noise effectively and at the same time keep the spectral radiation information properly.