Fig. 1. Schematic diagram of the CUP-VISAR system

Fig. 2. Data compression process of CUP

Fig. 3. Structural framework of TSVT

Fig. 4. Structural framework of SPM

Fig. 5. Overall framework of the algorithm

Fig. 6. Simulated stripe image. (a) Original 3D stripe image; (b) images at different time points t; (c) a certain image along y direction

Fig. 7. Actual noise images and simulated images. (a) Space light background noise; (b) background noise of CCD; (c) encoded image; (d) observed image of measurement values

Fig. 8. Observed data from the experiment. (a) Observed data; (b) effective signal data

Fig. 9. The 25th frame of the original data and reconstruction results of different algorithms along the t direction. (a) Original data; (b) E-3DTV; (c) ADMM-TV; (d) TVAL3; (e) proposed algorithm

Fig. 10. Original data and line-VISAR images of different algorithms reconstructed results at y=263. (a) Original data; (b) E-3DTV; (c) ADMM-TV; (d) TVAL3; (e) proposed algorithm

Fig. 11. Comparison of reconstruction image quality using four algorithms. (a) PSNR; (b) SSIM

Fig. 12. Comparison of maximum velocity errors between all line-VISAR images of four algorithms and the original data

Fig. 13. Line-VISAR velocity field at y = 263 reconstructed by proposed algorithm and its error with original data. (a) Original data; (b) line-VISAR velocity field reconstructed by proposed algorithm; (c) velocity error

Fig. 14. Images collected from the experiment. (a) Encoded image; (b) observed image; (c) line-VISAR image; (d) static interference fringe image

Fig. 15. Reconstruction results of different algorithms at the 15th frame along the t direction. (a) E-3DTV; (b) ADMM-TV;

Fig. 16. Line-VISAR images of different algorithm reconstructed results at y=263. (a) E-3DTV; (b) ADMM-TV; (c) TVAL3;

Fig. 17. Standard velocity field and line-VISAR velocity fields of different algorithms reconstructed results at y=263. (a) Standard velocity field; (b) E-3DTV; (c) ADMM-TV; (d) TVAL3; (e) proposed algorithm

Fig. 18. Relative velocity error of line-VISAR of different algorithms reconstructed results at y = 263. (a) ADMM-TV; (b) TVAL3; (c) proposed algorithm