Fig. 1. PSF of optical synthetic aperture imaging system when only piston error is considered. (a) Schematic diagram of piston error; (b1), (b2) PSF corresponding to different piston errors at single wavelength; (c1), (c2) PSF corresponding to different piston errors under broadband light illumination

Fig. 2. Structure diagram of MobileNet

Fig. 3. Experimental platform and control flow chart for closed-loop co-phasing experiment

Fig. 4. Experimental optical synthetic aperture imaging systems and corresponding ideal PSFs. (a),(b) Configuration of two-aperture system and corresponding ideal PSF; (c),(d) configuration of three-aperture system and corresponding ideal PSF

Fig. 5. Histogram of residual RMSE distribution on training set and test set using two-aperture experimental images

Fig. 6. Comparison between loaded and predicted piston errors in two-aperture system

Fig. 7. Results of closed-loop correction experiment on two-aperture system. (a) Convergence curve of residual; (b) image without closed-loop correction; (c) image with closed-loop correction

Fig. 8. Histogram of residual RMSE distribution on training set and test set using three-aperture experimental images

Fig. 9. Comparison of loaded and predicted piston errors in three-aperture system. (a) Measured sub-aperture 1; (b) measured sub-aperture 2

Fig. 10. Results of closed-loop correction experiment on three-aperture system. (a) Convergence curve of residual; (b) image without closed-loop correction; (c) image with closed-loop correction