Title Page
ABSTRACT
Contents
Abbreviation 8
1. Introduction 9
2. 3D Image visualization under photon-starved conditions 11
2.1. Introduction 11
2.2. Theory 11
2.2.1. Integral imaging 11
2.2.2. Synthetic aperture integral imaging 14
2.2.3. Computational photon counting 14
2.2.4. Photon counting integral imaging 16
2.2.5. Contrast limited adaptive histogram equalization 18
2.2.6. Histogram matching 20
2.2.7. proposed method 21
2.3. Experimental 23
2.3.1. Simulation setup 23
2.3.2. Simulation results 23
2.3.3. Experimental setup 27
2.3.4. Experimental result 27
2.4. Conclusion 31
3. Donut shape mask Fourier Peplography 32
3.1. introduction 32
3.2. Theory 33
3.2.1. Conventional Peplography 33
3.2.2. Fourier peplograhpy 35
3.2.3. Our method 36
3.3. Results 43
3.4. Conclusion 44
4. Conclusion 45
Bibliography 46
Summary in Korean 49
Table 1. PSR results of simulation. 26
Table 2. PSR results of optical experiments. 30
Figure 1. Images under the harsh conditions, (a) is foggy condition, and (b) is photon-starved condition. 10
Figure 2. Same 3D images with different light conditions. (3D images are focus on right car) 11
Figure 3. Overview the sequence of the integral imaging 12
Figure 4. Overview of the synthetic aperture integral imagine (SAII) 14
Figure 5. Images recorded under the photon-starved conditions. (a) is the conventional imaging system, and (b) is the photon counting imaging system. 15
Figure 6. 3D visualization in photon-starved conditions. (a) is conventional integral imaging, and (b) is photon counting integral imaging. Both 3D images are focused on the left car. 16
Figure 7. Photon counting integral imaging by MLE(a) and MAP(b). 17
Figure 8. (a)Orignal image under photon-starved conditions, (b)reconstructed imae by conventional histogram equalization, (c)reconstructed image by adaptive histogram equalization, and... 19
Figure 9. illustrates overview of the CLAHE sequence. 20
Figure 10. shows the histogram matching. (a) is original image, (b) is target image, and (c) is histogram matched image. 21
Figure 11. Histogram of (a) the histogram match image, and (b) the original image. 21
Figure 12. Overview of our proposed technique. 22
Figure 13. illustrate the simulation setup and shows the 3D scene. 23
Figure 14. Result of the simulation. 24
Figure 15. Peak sidelobe ratio (PSR) results for (a) house, (b) wolf, and (c) police car where 800 extracted photons are used. 25
Figure 16. shows the optical setup and scene. (a) is the optical setup and (b) is the scene. 27
Figure 17. Result of the optical experiment. (a) is conventional photon counting (b) is our proposed method. 28
Figure 18. PSR result for (a) left car, (b) right car, and (c) bunny where 10,000 extracted photons are used. 29
Figure 19. 3D reconstructed images by (a) 4*4 segmented image, (b) 16*16 segmented image, (c) 0.01 clip limit, and (d) 0.15 clip limit, respectively. (a) and (b) have the same clip limit 0.06. (c) and... 31
Figure 20. Illustration of photons that get through the scattering medium. 32
Figure 21. Results of the conventional peplography technique. (a) is original image, (b) is coventional peplography with digital processing, and (c) is conventional peplogarphy with... 34
Figure 22. shows the fourier tansform of the original image and scattering image. (c) is fourier transform (a) original image, and (d) is fourier transform of (b) the scattering image. 35
Figure 23. illustrates the cross mask filtering. (a) is the scattering image, and (b) is cross mask. 36
Figure 24. Overview of the donut shape mask. 37
Figure 25. illustrates the experiment setup. 39
Figure 26. illustrates the sequence of calculation of the scattering medium. 40
Figure 27. illustrate the sequence of optimization of gain per each image. 41
Figure 28. illustrates the combination of two graphs 42
Figure 29. Result of donut shape mask fourier peplography with fuzzy technique. (a) is original image, and (b) is Fourier peplography by using donut mask image. 43
Figure 30. utilize our method on partial scattering medium. (a) is original image, (b) is peplography image. 44