机载大孔径静态干涉光谱成像数据处理关键技术研究

Large-aperture Static Imaging Spectroscopy (LASIS) with characteristics of high throughput, high stability and high sensitivity performs the imaging spectroscopic technique that can achieve the abilities of high spectral resolution and high spatial resolution because its throughput two orders of magnitude higher than dispersive imaging spectroscopy. However, it is very sensitive in terms of its supporting platform, so as to only be used on the satellites. In order to take advantages of the LASIS principle, this project aims to do researches regarding on the airborne-platform other than the satellite-platform of LASIS, including models of moving errors, retrievals of errors in high-accuracy gestures and recovering of spectra from non-uniform sampled data by the interferometer. Methods of domain transform or data processing are used to bring about appropriate image registration approaches for modulated interference images; and thereafter the image registration and controlled-ground points are used to calculate installation errors of the POS system and the spectrometer. Approaches for recovering spectra of the single sided and non-uniform sampled interferograms will be applied to improve the accuracy of recovered spectra for LASIS data. The research results of this project is capable to achieve motion imaging with high spatial resolution, high spectral resolution and high radiation resolution for High Resolution Foremost Special Projects high resolution earth observations, providing foundations of applications and computing techniques.

大孔径静态干涉光谱成像技术（简称LASIS）具有高通量、高稳定度、高灵敏度等特点，由于光通量比色散型成像光谱仪高两个数量级，能够同时实现高光谱分辨率与高空间分辨率。LASIS的复原光谱受平台运动误差影响较大，目前只应用于卫星平台。为更好地发挥LASIS的原理优势，拓展应用至机载平台，本项目针对目前常用的两种LASIS姿态误差计算及修正方案的不足，开展机载平台运动误差建模、高精度姿态误差反演及非均匀采样干涉数据光谱复原方面的研究。利用域变换或数据处理的方法提出适用于干涉调制图像的图像配准方法，利用图像配准和地面控制点，反演计算POS系统与光谱仪的安装误差，并对POS数据进行修正，寻求适用于单边过零非均匀采样干涉图的光谱复原方法，提高LASIS数据的光谱复原精度。该项目的研究成果将为高分重大专项高分辨率对地观测等实现高空间分辨率、高光谱分辨率、高辐射分辨率运动成像，提供应用基础和技术方法支撑。

项目研制过程中开展了大孔径静态干涉光谱成像技术（以下简称LASIS）成像原理建模；运动成像误差建模，仿真了不同姿态对干涉图及复原光谱的影响；提出了基于条纹消除的干涉图像配准方法，提高干涉图像的配准精度；分析了POS误差对干涉图像的影响，完善了利用POS数据对目标干涉图进行主动追踪的方法；提出多维立体干涉图插值方法，提高干涉图提取的精度；对基于网格化的非均匀采样快速傅立叶变换算法（以下简称NUFFT）和基于规则傅里叶矩阵的NUFFT算法进行了仿真比较，指出基于规则傅里叶变换矩阵的NUFFT算法复原精度更高，更适用于单边过零非均匀采样干涉图的光谱复原。.通过该项目的研究，实现了对LASIS高质量干涉光场信息的提取，可以有效抑制平台运动误差的影响，充分发挥LASIS原理本身的巨大优势，推动LASIS在高分专项等机载遥感领域的应用。项目研究过程中，发表学术论文5篇，申请发明专利3项，授权软件著作权2项，完成了预期目标。