HJ-1A星HSI数据2级产品条带噪声特征分析及去除方法

Improving method of eliminating level 2 Hyper-spectral data stripes of HJ-1A satellite

HJ-1A星在成像过程中由于线阵传感器相邻探测器在响应和输出时的奇偶效应,故其获取的影像存在明显的条带噪声.首先,对影像亮度统计直方图进行分析;然后,使用全局去条带法对影像进行条带去除.根据处理效果以及直方图特征分析,将影像分为3类.结果证明,全局去条带法只适用于亮度直方图符合正态分布的影像.针对全局去条带法的局限性,提出了一种改进的条带去除方法,使用该方法基本上可以获得相对较好的处理结果.

Since the linear array sensor adjacent the detector in the response and output effect of parity,there is some stripe noise on HJ-1A HIS（Hyperspectral Imaging Radiometer） data product.At first,the image brightness statistical histograms are analyzed,and the globally eliminate strip method is used to deal with images.According to the analysis and the result,images are divided into three categories.The results show that the globally eliminate strip method is suitable for the image within a specific range of DN（Digital Number）.In order to solve this problem,a improving eliminate strip method is proposed.The method is more suitable for the level 2 hyperspectral data of HJ-1 Asatellite.