Fig. 1. Experimental setup for computational ghost imaging
Fig. 2. Original image of five-pointed star
Fig. 3. CGI reconstructed image of five-pointed star using the correlation algorithm when the reconstruction times is 200 and K=19.5%
Fig. 4. CSHGI reconstructed images of five-pointed star using SP reconstruction algorithm when the light source is Hadamard speckle. (a) The number of iterations is 100 and K=9.8%; (b) the number of iterations is 150 and K=14.6%; (c) the number of iterations is 200 and K=19.5%
Fig. 5. CSHGI reconstructed images of five-pointed star using OMP reconstruction algorithm when the light source is Hadamard speckle. (a) The number of iterations is 100 and K=9.8%; (b) the number of iterations is 150 and K=14.6%; (c) the number of iterations is 200 and K=19.5%
Fig. 6. Original image of CUST
Fig. 7. CGI reconstructed image of CUST using the correlation algorithm when the reconstruction times is 1000 and K=24.4%
Fig. 8. CSHGI reconstructed images of CUST using SP reconstruction algorithm when the light source is Hadamard speckle. (a) The number of iterations is 600 and K=14.6%; (b) the number of iterations is 800 and K=19.5%; (c) the number of iterations is 1000 and K=24.4%
Fig. 9. CSHGI reconstructed images of CUST using OMP reconstruction algorithm when the light source is Hadamard speckle. (a) The number of iterations is 600 and K=14.6%; (b) the number of iterations is 800 and K=19.5%; (c) the number of iterations is 1000 and K=24.4%
Fig. 10. Schematic of CSHGI experiment
Fig. 11. Object to be measured in the experiment
Fig. 12. CSHGI reconstructed images using the SP reconstruction algorithm with Hadamard speckle in the experiment. (a) The number of iterations is 100 and K=9.8%; (b) the number of iterations is 150 and K=14.6%; (c) the number of iterations is 200 and K=19.5%
Fig. 13. CSHGI reconstructed images using the OMP reconstruction algorithm with Hadamard speckle in the experiment. (a) The number of iterations is 100 and K=9.8%; (b) the number of iterations is 150 and K=14.6%; (c) the number of iterations is 200 and K=19.5%
Number of iterations | 50 | 100 | 150 | 200 | 250 |
---|
K/% | 4.9 | 9.8 | 14.6 | 19.5 | 24.4 | /s | 0.343 | 0.543 | 1.559 | 3.094 | 7.594 | TOMP /s | 0.172 | 0.266 | 0.734 | 1.563 | 2.266 | SSSIM-SP | 0.2552 | 0.3534 | 0.5047 | 0.6888 | 0.7597 | SSSIM-OMP | 0.2277 | 0.3316 | 0.4368 | 0.6820 | 0.7476 |
|
Table 1. Comparison of reconstruction results using two algorithms when the object to be measured is a binary image
Number of iterations | 400 | 600 | 800 | 1000 | 1200 |
---|
K/% | 9.8 | 14.6 | 19.5 | 24.4 | 29.3 | /s | 79.058 | 267.336 | 391.707 | 666.890 | 1052.539 | TOMP /s | 16.062 | 41.221 | 82.532 | 157.416 | 254.331 | SSSIM-SP | 0.5449 | 0.6908 | 0.8552 | 0.9255 | 0.9369 | SSSIM-OMP | 0.5165 | 0.6722 | 0.7954 | 0.8504 | 0.8840 |
|
Table 2. Comparison of reconstruction results using two algorithms when the object to be measured is a gray-scale image
Number of iterations | 50 | 100 | 150 | 200 | 250 |
---|
K/% | 4.9 | 9.8 | 14.6 | 19.5 | 24.4 | /s | 0.173 | 0.642 | 1.572 | 3.671 | 6.678 | TOMP /s | 0.163 | 0.321 | 0.703 | 1.412 | 2.310 | SSSIM-SP | 0.2051 | 0.3900 | 0.4949 | 0.6396 | 0.7379 | SSSIM-OMP | 0.1268 | 0.2903 | 0.4240 | 0.6184 | 0.7215 |
|
Table 3. Comparison of reconstruction results of two algorithms in the experiment