时空联合调制干涉成像光谱仪的坏像元校正  

Bad Pixel Correction in Temporally and Spatially Modulated Fourier Transform Imaging Spectrometer

作　　者：李谦[1] 袁艳[1] 操俊 

机构地区：[1]北京航空航天大学,北京100191

出　　处：《航天返回与遥感》2016年第6期76-84,共9页Spacecraft Recovery & Remote Sensing

基　　金：国家自然科学基金(61471018)

摘　　要：时空联合调制干涉成像光谱仪通过获取景物目标的干涉图像序列,根据傅里叶变换光谱学理论,复原目标光谱信息。仪器探测器上坏像元的存在会破坏目标干涉图的连续性,造成干涉信息不能正确提取,导致目标反演光谱失真。文章分析了坏像元校正后的精度与不同光程差位置对目标复原光谱的影响,提出了一种基于径向基神经网络的校正方法,方法复杂度低,实现简单。试验结果表明,文章提出的方法与空间维线性插值、双线性插值、邻域均值三种遥感影像常用坏像元校正方法相比,精度有明显改善。Temporally and spatially modulated fourier transform imaging spectrometer gains interference image sequences of the target through the process of pushing, by using Fourier Transform for spectrum reconstruction. The existence of bad pixels in its detector destroys the continuity of interference images, resulting information of interference images being picked up incorrectly, which leads to the distortion of target＇s spectrum. We show the effects of recovering spectra by bad pixels in different position of optical path difference, and present a new correction method based on radical basis function neural networks, which has higher accuracy than that of the rrtost popular correction methods. The advantages of our method are discussed and demonstrated with simulation data.

关 键 词：时空联合调制干涉成像光谱仪 干涉图像 光程差 坏像元 校正 高光谱遥感 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]