Research on an Equal Wavelength Spectrum Reconstruction Method of Interference Imaging Spectrometer
XIE Pei-yue1, 2, YANG Jian-feng1, XUE Bin1*, Lü Juan1, HE Ying-hong1, LI Ting1, MA Xiao-long1
1. Xi’an Institute of Optics and Precision Mechanics of Chinese Academy of Sciences, Xi’an 710119, China 2. University of Chinese Academy of Sciences, Beijing 100049, China
Abstract：Interference imaging spectrometer is one of the most important equipments of Chang’E 1 satellite, which is applied to analysis the material composition and its distribution of the surface on the moon. At present, the spectral resolution of level 2B scientific data obtained by existing methods is 325 cm-1. If we use the description way of wavelength resolution, various spectrum is different: the first band is 7.6 nm, the last band is 29 nm, which introduces two questions: (1) the spectral resolution description way mismatch with the way of ground spectral library used for calibration and comparison; (2) The signal-to-noise ratio of the spectra in the shortwave band is low due to the signal entering narrow band is little. This paper discussed the relationship between wavelength resolution and cut-off function based on the reconstruction model of CE-1 interference imaging spectrometer. It proposed an adjustable cut-off function changing with wavelength or wavelength resolution, while selected the appropriate Sinc function as apodization to realize the reconstruction of arbitrary specified wavelength resolution in the band coverage. Then we used this method to CE-1 on orbit 0B data to get a spectral image of 29 nm wavelength resolution. Finally, by usingthe signal-to-noise ratio, principal component analysis and unsupervised classification method on the reconstruction results with 2 grade science data from ground application system for comparison, the results showed that: signal-to-noise ratio of the shortwave band increased about 4 times, and the average increased about 2.4 times, the classification based on the spectrum was consistent, and the quality of the data was greatly improved. So, EWSR method has the advantages that: (1) in the case of keeping spectral information steadiness，it can improve the signal-to-noise ratio of shortwave band spectrum though sacrificed part of spectral resolution; (2) it can achieve the spectral data reconstruction which can set arbitrary band position or specify any wavelength resolution within the band range.
