成像光谱仪星上光谱定标的数据处理  

Data processing of on-orbit spectral calibration of space-borne high resolution imaging spectrometer

作　　者：张军强[1,2] 邵建兵[1] 颜昌翔[1] 吴清文[1] 陈伟[1] 

机构地区：[1]中国科学院长春光学精密机械与物理研究所,吉林长春130033 [2]中国科学院研究生院,北京100039

出　　处：《中国光学》2011年第2期175-181,共7页Chinese Optics

基　　金：国防预研基金资助项目(No.O5001SA050)

摘　　要：为了提高仪器光谱定标精度,降低谱线偏移对地物反射光谱数据反演精度的影响(特别是大气吸收峰附近),根据星上光谱定标的特点,介绍了星上光谱定标数据处理的常用方法。通过对实验室光谱定标时两种典型工况的比较,实现了星上定标数据处理算法的地面仿真,验证了算法的可行性,并比较了各自的优缺点。结果表明,基于谱线匹配的标准差法、相关系数法和最小差值法计算精度较高,但计算效率较低;基于多项式拟合的极值法计算精度较低,但是计算效率较其他算法提高1个数量级。用极值法对数据进行预处理,快速确定谱线偏移量,再用谱线匹配算法在小范围内精确计算谱线偏移量,可以在不影响计算精度的前提下提高运算速度,计算精度优于1 nm,满足成像光谱仪星上光谱定标的精度要求。In order to improve the spectral calibration precision of a High Resolution Imaging Spectrometer（HRIS） and to reduce the influence of the spectral shift on the surface reflectance retrieving of the ground spectra,especially in those wavelengths mostly affected by gaseous absorptions,the common data processing algorithms used in the on-orbit spectral calibration are introduced.The feasibility of these different algorithms is verified and their advantages and disadvantages are emulated with two cases of the laboratory spectral calibration under a thermal vacuum environment.It is shown that the algorithm based on spectra-matching,such as standard deviation method,correlation coefficient method and the minimized difference method can offer an advantage in the processing precision,and the extreme value method based on polynomial-fitting works much better in the processing speed.If polynomial-fitting method confirms the spectral shift fleetly in the pretreatment phase firstly,the spectral bound for spectral matching processing will be narrowed.A new algorithm with high processing precision and fast processing speed which combines spectra-matching and polynomial-fitting ideas was presented,it can offer the band-center wavelength accuracy better than 1 nm and can meet the need of the on-orbit spectral calibration precision of the HRIS.

关 键 词：成像光谱仪 星上定标 光谱定标 数据处理 

分 类 号：TP73[自动化与计算机技术—检测技术与自动化装置] O433[自动化与计算机技术—控制科学与工程]