Fig. 1. The generation process of illumination light fields by the fiber laser array.

Fig. 2. Simulation results of illumination light fields of different arrays. (a₁)–(a₃) Array configurations: top to bottom, rectangle, hexagon, and Fermat spiral; (b₁)–(b₃) 3D g(2) distribution for each configuration; (c₁)–(c₃) g(2) distribution along the u-axis; (d₁)–(d₃) g(2) distribution along the v-axis; (e₁)–(e₃) vertical view of g(2) distribution.

Fig. 3. Imaging simulations of rectangle, hexagon, and Fermat spiral arrays. Left: binary “3-slits”; Right: grayscale “drone.” Reconstructed via DGI and CS-TV.

Fig. 4. Simulation results of g(2) for different N. (a₁)–(a₄) Array configs with different N; (b₁)–(b₄) 3D g(2) distributions; (c₁)–(c₄) vertical views for g(2) with different N.

Fig. 5. Simulation imaging results of different N.

Fig. 6. RMSEs of the images of different N.

Fig. 7. System schematic of SPI illuminated by the Fermat spiral fiber laser array.

Fig. 8. High-speed data synchronous acquisition system diagram. (a) 20 kHz synch signals; (b) the time-varying phase modulation attached to the beam by the phase modulator; (c) the acquired illumination light field; (d) the time-varying normalized light intensity values detected by the SPD.

Fig. 9. Results of 2000 acquired illumination light fields. (a) An acquired illumination light field; (b) the 3D distribution of g(2); (c) viewing the 3D distribution of g(2) from the positive direction of the v-axis; (d) viewing the 3D distribution of g(2) from the positive direction of the u-axis.

Fig. 10. Experimental imaging results. Images reconstructed by DGI, CS-TV, and UNN at different samples. The object is a transmissive “2.” The red text indicates the lowest samples for distinguishable imaging.

Fig. 11. The distribution of g(2) for different samples. (a₁)–(a₄) Viewing the 3D distribution of g(2) from the positive direction of the v-axis, (b₁)–(b₄) Viewing the 3D distribution of g(2) from the positive direction of the u-axis.

Fig. 12. SSIM values at different samples, with the g(2)(u,v;u0,v0) at 2000 samples serving as the reference.