Fig. 1. The schematic diagram of typical (a) thermal and (b) entangled GI. In thermal GI, a laser is commonly used to illuminate a rotating ground glass plate, generating a pseudothermal light field. The intensity distribution of the light field irradiating on the object is recorded by a CCD camera (D_r). For computational GI, the actual CCD camera is unnecessary, with the intensity distribution being calculated. The outgoing light illuminates the object, and the echo reflected from the object is detected by a bucket detector (D_b). Then the reflectance of an object can be obtained by correlating the reference arm with the object arm. For entangled GI, entangled photon pairs are generated by SPDC. One of the photons interacts with the object and is received by D_b, while the other is detected by a photon counting array D_r. Then the image is obtained through correlation.

Fig. 2. Signal transmission process in the detector. A number of photons hit the pixel area during exposure time t₀ and create a number of electrons. The electrons are transferred with the dark noise introduced. Finally, the signal is amplified and digitized for output. η is the probability of converting a photon into a photoelectron, and K is the overall system gain.

Fig. 3. (a) Variation in CNR with signal photoelectrons n¯_s and object size k, where µ = 1. (b) CNR changes with n¯_s under different noise levels µ, where k = 10. (c) CNR changes with n¯_s under different object sizes k, where µ=10.

Fig. 4. Relationship between the number of detected photons and CNR of reconstructed images, under the same noise conditions. The horizontal axis represents the CNR of images, and the vertical axis represents n¯_s. The vertical dark blue chain line represents CNR = 1. (a) Without noise µ=0. (b) With noise µ=10 and µ=20.

Fig. 5. The imaging sensitivity of GI and CI with different detection efficiency. The dotted line represents CI, the solid line represents GI, the blue line represents µ = 10, and the red line represents µ = 20. The detection efficiency of each pixel in the array detector is 60% that of single-pixel detector.

Fig. 6. Flow chart of the simulation process of (a) GI and (b) CI.

Fig. 7. The simulation imaging results of GI and CI in Fig. 6(b), with µ = 20.

Fig. 8. The imaging sensitivity of CI and GI under different noise levels. GI: µ = 2, 5, 10, 15, 20, and CI: µ = 20.

Fig. 9. The simulation imaging results of GI and CI in Fig. 8, with noise levels µ = 2 and μ = 20, respectively.

Fig. 10. (a) The comparison of coincidence counting and fluctuation correlation in Geiger mode. (b) The comparison between Geiger mode and linear mode.

Fig. 11. The variation of CNR with different object sizes k, with q = 10. (a) Theoretical results. (b) Simulation imaging results.

Fig. 12. Sensitivity comparison of GI and CI with q = 10. (a) CI with accumulation of photons, without noise µ = 0. (b) CI with accumulation of photons, with noise µ = 10. (c) CI with pulsed illumination, without noise µ = 0. (d) CI with pulsed illumination, with noise µ = 10.

Fig. 13. The simulation imaging results of GI and CI of Fig. 12(c). The upper row shows the imaging results of GI, and the lower row shows the imaging results of CI.

Fig. 14. The CNR of entangled GI changes with n¯_p. The lines with different colors represent the imaging quality under different quantum efficiencies η and noise conditions μ.

Fig. 15. The CNR of entangled GI with different η, µ_b, and µ_d.

Fig. 16. Sensitivity comparison between entangled GI and CI. The solid line represents GI, and the dotted line represents CI. (a) With dark noise µ_d. (b) With background noise µ_b.

Fig. 17. The simulation imaging results of entangled GI and CI in Fig. 16(b), with background noise µ_b = 2. The upper row shows the imaging results of GI, and the lower row shows the imaging results of CI.

Fig. 18. The DGI and NGI results of Fig. 4, compared to the corresponding results of GI and CI. The yellow dots and purple triangles denote the simulation results of DGI and NGI, respectively.