Research On Data Processing Method Of Micro-spectrometer Based On ZYNQ

The traditional Fourier transform infrared spectrum analyzer is large,not portable,high cost in manufacturing,poor and real-time acquisition;which cannot meet the requirements of on-site real-time detection and data processing.In order to fulfil requirements the data processing and spectrum restoration after the interferogram is collected,This article will research the data processing method in the process of transforming interference data into spectral data.The principle of Fourier transforms spectrometer and the interference signal acquisition methods analyzed in this article.For example,digital method resampling,equal time interval sampling and equal optical path difference sampling.The shortcomings of the traditional digital method of resampling have been improved,and the experiments results show that the improved digital method resampling has higher accuracy compare the traditional digital method resampling.According to the characteristics of the interference signal,the techniques related to noise reduction,apodization and phase correction in the interference signal are studied,and several common phase correction methods are compared and analyzed.Finally,the Mertz phase correction method is used to correct the phase slight difference generated during the collecting.When using digital methods to resample the interference signal,the sample light and the reference light need to be interpolated,the common interpolation methods are studied and experimentally compared.Finally,cubic spline interpolation is used to interpolate the interference signal.At the same time,this paper uses an improved wavelet threshold function to denoise the spectrum.Finally,a real-time data processing experimental platform for the Fourier transform micro-spectrometer is built based on the MEMS micromirror and ZYNQ hardware environment.Detoe,interpolation,phase correction and filtering are performed on the collected 1310 nm interference signal on this platform.The experimental results show that it is feasible to preprocess the data on this system.