Spectrum recovery method analysis on nonuniform sampling interference data
Fengzhen Huang, Jingzhen Li, Jun Cao · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2015
Temporally and Spatially Modulated Fourier Transform Imaging Spectrometer (TSMFTIS) is a new imaging spectrometer without moving mirrors and slits. The interferogram of the target point can be consisted by sequentially arranging the interference information extracted from the same target point of the sequential images, and the spectrum can be recovered by using fast Fourier transform. In the practical application, there is nonuniform sampling in the interference data, and many researchers have carried out researches on nonuniform sampling with the fast Fourier transform algorithm. As to the issue of interference data in the nonuniform sampling, the nonuniform sampling degree’s impact on the recovered spectrum precision is currently and mainly analyzed. This paper has adapted several typical nonuniform fast Fourier transform (NUFFT) methods, carried out spectrum recovery precision comparison on the interferogram of the nonuniform sampling point with the above methods, and further analyzed the impact of kernel function type, oversampling ratio and kernel function width’s on spectrum recovery precision in the above mentioned methods. The experiment result indicates that, when the oversampling ratio is 4 and the kernel function width is 4, the spectrum recovery precision with NUFFT based on Blackman type kernel function is optimal, however, the Gaussian kernel function is stable.