Error Analysis And Correction Of Spatial Heterodyne Remote Sensing Data

Spatial Heterodyne Spectroscopy(SHS)is a new hyperspectral analysis technology.Because of its advantages of obtaining ultra-high spectral resolution spectral data,it is widely used in atmospheric detection,satellite remote sensing and other fields.However,errors will be introduced in the making process of space heterodyne spectrometer or the change of detection working environment,which will make the interferogram data inaccurate and need error correction.On the satellite platform,due to the change of the working environment,the correction parameters measured on the ground are no longer applicable to the spatial heterodyne interferogram data,especially the change of the modulation error(phase error and non-uniform error),which greatly affects the accuracy of the correction spectrum.Therefore,this paper mainly studies the acquisition and correction of non-uniform errors and phase errors of spatial heterodyne remote sensing data,and uses the modulation error correction method and spectral interferogram bidirectional correction method to correct these two kinds of errors.Firstly,the modulation error correction method was theoretically derived,and the measured spectrum obtained by Greenhouse gases Monitoring Instrument(GMI)on GF-5was used as the data processing object,and the phase error and non-uniform error were corrected by the modulation error correction method.For non-uniform error,12 measured spectra were divided into two groups.One group was used to calibrate the non-uniform error curve,and the other group was used to correct the error.Then the standard deviation(STD),mean square error(MSE)and signal-to-noise ratio(SNR)of the measured spectrum after non-uniform error correction are calculated.The results show that both the STD and MSE are reduced to a certain extent,while the SNR increases.Among them,the measured spectrum 7 has the best correction effect,with the STD reduced by 0.0849,SNR increased by 2.0640 d B,and MSE decreased by 0.0603.After the non-uniform error correction,the phase error was corrected,and the MATLAB simulation error surface was added to the error-free spectrum simulated by SCIATRAN,and then compared with the measured spectrum to determine the phase error surface.Finally,the modulation error correction method was used to correct the phase error.The results show that for the phase error surface presented in this paper,the STD and MSE of spectra 7,8,9 and 10 are all reduced to a certain extent,SNR improved,and the correction effect of spectra 9 is the best,with the STD reduced by 0.0274,MES reduced by 0.015,SNR increased by 1.2089 d B.The above shows that the modulation error correction method has a certain effect on error correction,but there is still room for improvement.To further enhance the correction effect,from two aspects of spectrum and the interference figure,analysis for error separation and error factors,the modulation error parsing,think that the error of the spatial heterodyne remote sensing data mainly comes from efficient spectral peak frequency and intensity of interferogram amplitude changes,spectral interferogram bidirectional correction method is proposed.In calibration process for selecting measured spectrum’s spectrum peaks position and SCIATRAN simulation to simulate the linear spectral spectrum peak position,get the correct spectral peak position,and then after fitting coordinates to make use of the simulation spectrum interpolation,calculated with the measured spectral spectrum peak position corresponding to the strength of the actual change,finally the results for other correction of the measured spectra.The STD,MSE and SNR of the corrected spectra are calculated,and the results show that both the STD and the MSE are significantly reduced,while the SNR is significantly increased,and the STD is basically below 0.07,and the SNR can reach more than 20 d B.The results show that the spectral interferogram bidirectional correction method is effective in the correction of spatial heterodyne remote sensing data.