Fig. 1. Flow of NLRSTV simultaneous spatial and spectral information recovery algorithm

Fig. 2. Washington D.C. hyperspectral image (HSI) and its simulated interferogram. (a) Original HSI; (b) simulated spatial-interferogram image; (c) spectral curve and (d) simulated interferogram curve at the center pixel

Fig. 3. PSNR and SSIM values of each band using different recovery methods in Case 6. (a) PSNR; (b) SSIM

Fig. 4. SAD images of recovery results using different methods in Case 6. (a) DCT method; (b) DCT+LRTV method; (c) NLRSTV method

Fig. 5. Visual comparison of recovered HSI by different methods. (a) Original HSI; (b)-(d) recovered HSI by DCT, DCT+LRTV, and NLRSTV in Case 1; (e)-(g) recovered HSI by DCT, DCT+LRTV, and NLRSTV in Case 6

Fig. 6. Influence of interferogram noise on recovered spectrum by different methods. (a) Original HSI; (b) recovered HSI by DCT; (c) recovered HSI by DCT+LRTV; (d) recovered HSI by NLRSTV

Fig. 7. Comparison of recovery results of IIM Orbit 0234 data using different recovery methods. False-color-composite images of recovery results with (a) DCT, (b) DCT+LRTV, and (c) NLRSTV methods; identification results of bad points of (d) DCT, (e) DCT+LRTV, and (f) NLRSTV methods

Fig. 8. Estimated signal-to-noise ratio of recovered HSI by different methods

Fig. 9. Sensitivity analysis results of parameters of NLRSTV model. (a) MPSNR; (b) MSSIM; (c) MSAD

Fig. 10. Convergence analysis results of NLRSTV model. (a) MPSNR; (b) MSSIM; (c) MSAD