這篇論文的目的是為了改善太赫茲成像的解析度和測量速度。我們推出一種基於太赫茲時域光譜的3D成像演算法，它與傳統的計算機斷層掃描演算法相比，可提供更好的圖像品質，並且有更快的測量時間。這個演算法使用了時間和振福特徵的優點來加強3D太赫茲影像的品質。我們的演算法達到了0.64%的均方誤差值，這比傳統斷層掃描演算法好五倍。此外，我們演算法算出來的影像品質以三分之一的投影數就超越了傳統的計算機斷層掃描。
本文也介紹了不同類型的2D太赫茲成像，這個對在進行3D太赫茲成像前，會有幫助。其中的穿透式成像和反射式成像，他們皆有不同的應用。穿透式成像用穿透待測物的訊號來重建太赫茲影像。其設置簡單、訊雜比高，並可直接量測物體。行李安檢和藥品檢測等應用皆可被證明。反射式成像用被待測物反射的訊號來重建太赫茲影像。反射訊號可以被用來區分不同層之結構。因此，汽車塗層測量和古書分析等應用可以被實現。

The aim of this thesis is to improve the resolution and measurement speed of terahertz imaging. We presented an algorithm for terahertz time-domain spectroscopy (THz-TDS) 3D imaging, which offer superior image quality compared with conventional computed tomography (CT) algorithm with fast measurement time. This algorithm takes advantages of both temporal and amplitude information to increase the quality of 3D terahertz image. A 0.64% mean square error value was achieved by our algorithm which was five times better than conventional CT algorithm. Moreover, the image quality of our algorithm outperformed CT with only one-third number of projections.
Different type of 2D terahertz imaging methods were demonstrated in this thesis. This is helpful before performing the 3D terahertz image. The transmission imaging and reflection imaging of the 2D imaging have different applications. Transmission imaging uses signal of terahertz wave which penetrates tested sample to reconstruct terahertz image. Transmission imaging is a simple setup that can measure object directly with higher signal-to-noise ratio. Applications like luggage screening and pharmaceutical inspection can be demonstrated. Reflection imaging uses signal of terahertz wave which reflected by tested sample to reconstruct terahertz image. The reflected signal can be used to distinguish different layer structure. Therefore, applications like car coating detection and ancient book analysis can be realized.

摘要 i
Abstract ii
致謝 iv
Contents v
List of Figures vii
Chapter 1 Introduction 1
1.1 Background of terahertz wave 1
1.2 Application of terahertz wave 3
Chapter 2 Terahertz imaging setup and measurement 7
2.1 Terahertz source 7
2.1.1 Terahertz continuous wave 7
2.1.2 Terahertz pulse 8
2.2 Setup of terahertz imaging 9
2.2.1 Terahertz time-domain spectroscopy 10
2.2.2 Asynchronous optical sampling system 12
2.2.3 Terahertz transmission and reflection imaging 14
2.3 Automation control 18
2.3.1 Devices of automation control 18
2.3.2 Control method 23
2.4 Imaged sample in algorithm improvement 29
2.4.1 Imaged sample manufacturing 29
2.4.2 Imaged sample alignment 32
2.4.3 Ground Truth of terahertz images 35
Chapter 3 Data and principle of imaging 38
3.1 Measured data form 38
3.2 Measured data alignment 40
3.3 Terahertz image in transmission imaging 43
3.3.1 Signal in time domain 43
3.3.2 Signal in frequency domain 46
3.3.3 Results of terahertz transmission imaging 47
3.4 Terahertz image in reflection imaging 54
3.4.1 Signal in time domain 54
3.4.2 Results of terahertz reflection imaging 56
3.4.3 Sample alignment 62
Chapter 4 Computed tomography and differential pulse delay computed tomography 67
4.1 Computed tomography 67
4.2 Challenge and motivation 72
4.3 Algorithm of differential pulse delay computed tomography 74
4.4 Results of computed tomography and differential pulse delay computed tomography 77
Chapter 5 Conclusion 83
Chapter 6 Future work 85
Reference 86

1. McIntosh, A.I., et al., Terahertz spectroscopy: a powerful new tool for the chemical sciences? Chemical Society Reviews, 2012. 41(6): p. 2072-2082.
2. Taylor, Z., et al., Reflective terahertz imaging of porcine skin burns. Optics letters, 2008. 33(11): p. 1258-1260.
3. Breitenstein, B., et al., Introducing terahertz technology into plant biology: A novel method to monitor changes in leaf water status. Journal of Applied Botany and Food Quality, 2012. 84(2): p. 158.
4. Liu, H.-B., et al., Terahertz spectroscopy and imaging for defense and security applications. Proceedings of the IEEE, 2007. 95(8): p. 1514-1527.
5. Kawase, K., et al., Non-destructive terahertz imaging of illicit drugs using spectral fingerprints. Optics express, 2003. 11(20): p. 2549-2554.
6. Busch, S., et al., Optical properties of 3D printable plastics in the THz regime and their application for 3D printed THz optics. Journal of Infrared, Millimeter, and Terahertz Waves, 2014. 35(12): p. 993-997.
7. Squires, A. and R. Lewis, Feasibility and characterization of common and exotic filaments for use in 3D printed terahertz devices. Journal of Infrared, Millimeter, and Terahertz Waves, 2018. 39(7): p. 614-635.
8. Ok, G., et al., Foreign object detection by sub-terahertz quasi-Bessel beam imaging. Sensors, 2013. 13(1): p. 71-85.
9. Wei, X., et al., Generation of arbitrary order Bessel beams via 3D printed axicons at the terahertz frequency range. Applied optics, 2015. 54(36): p. 10641-10649.
10. Slocum, D.M., et al., Atmospheric absorption of terahertz radiation and water vapor continuum effects. Journal of Quantitative Spectroscopy and Radiative Transfer, 2013. 127: p. 49-63.
11. Zhong, H., et al., Nondestructive defect identification with terahertz time-of-flight tomography. IEEE Sensors Journal, 2005. 5(2): p. 203-208.
12. Takayanagi, J., et al., High-resolution time-of-flight terahertz tomography using a femtosecond fiber laser. Optics express, 2009. 17(9): p. 7533-7539.
13. Ferguson, B., et al., T-ray computed tomography. Optics Letters, 2002. 27(15): p. 1312-1314.
14. Ferguson, B.S., S. Wang, and X.-C. Zhang, Transmission mode terahertz computed tomography. 2006, Google Patents.
15. Jewariya, M., et al., Fast three-dimensional terahertz computed tomography using real-time line projection of intense terahertz pulse. Optics express, 2013. 21(2): p. 2423-2433.
16. Recur, B., et al., Investigation on reconstruction methods applied to 3D terahertz computed tomography. Optics express, 2011. 19(6): p. 5105-5117.
17. Beylkin, G., Discrete radon transform. IEEE transactions on acoustics, speech, and signal processing, 1987. 35(2): p. 162-172.
18. Helgason, S. and S. Helgason, The radon transform. Vol. 2. 1980: Springer.
19. Huang, Z., Extensions to the k-means algorithm for clustering large data sets with categorical values. Data mining and knowledge discovery, 1998. 2(3): p. 283-304.
20. Likas, A., N. Vlassis, and J.J. Verbeek, The global k-means clustering algorithm. Pattern recognition, 2003. 36(2): p. 451-461.
21. Wagstaff, K., et al. Constrained k-means clustering with background knowledge. in Icml. 2001.
22. Hung, Y.-C. and S.-H. Yang. Terahertz Deep Learning Computed Tomography. in 2019 44th International Conference on Infrared, Millimeter, and Terahertz Waves (IRMMW-THz). 2019. IEEE.
23. Kim, Y. and D.-S. Yee, High-speed terahertz time-domain spectroscopy based on electronically controlled optical sampling. Optics letters, 2010. 35(22): p. 3715-3717.