表面電漿共振 (Surface Plasmon Resonance, SPR) 生物感測器常用於檢測生物及化學分子，具有高靈敏度 (high sensitivity)、即時性 (real-time)、不須事先標定生物分子 (label-free) 等優點，但傳統上幾乎都是使用一般電腦來處理後端的影像分析，其中牽涉到需手動來選取影像中有表面電漿共振反應的區域，既耗時費力又容易造成人為誤差，而無法達到自動化的檢測分析。
　　因此本研究透過嵌入式系統 (embedded system) 開發一套自動化影像分析系統，搭配感光耦合元件 (Charge-coupled Device, CCD) 來擷取表面電漿共振之影像，並利用多執行緒 (multi-threading) 架構使影格率 (Frame Per Second, FPS) 增快約 81.4 %，提高了檢測的時間解析度 (temporal resolution)，亦能使影像取樣 (sampling) 時能更均勻 (uniform)，進而增加實驗與檢測結果的準確度。當系統收到影像後，會立即執行已預先寫好的影像切割演算法，自動化找出有表面電漿共振反應的區域，再將反應區域的平均強度數值，透過藍芽傳輸的方式實時 (real-time) 傳給智慧型裝置 (smart device)，以提供使用者即時確認檢測結果的平台。
　　而藉由自動化影像切割 (automatic image segmentation) 演算法，可降低處理時間及人為誤差的可能性，實質達到即時性與自動化檢測分析的效果，並能提高實驗結果之準確度，而經過實驗得到的結果可知目前感測器之靈敏度 (sensitivity) 可達到 9.6*10^-6 RIU。而使用嵌入式系統取代傳統電腦亦能大幅縮減後端運算裝置的體積與成本，並且可與前端表面電漿共振生物感測器整合成為一可攜式 (portable) 系統裝置。

Surface plasmon resonance (SPR) biosensors are used to detect biological and chemical molecules. They have serval advantages such as high sensitivity, real-time detection, and label-free. However, backend image analysis traditionally relies on a personal computer to process, and involves manual selection of SPR reactive region. It is time-consuming, liable to cause artificial error, and unable to detect automatically.
　　Therefore, an automatic image analysis system has been proposed by using an embedded system, combing charge-coupled device (CCD) to capture SPR images. The multi-threading architecture of the main program increases FPS by about 81.4 %, which also increases temporal resolution of detection. Moreover, the sampling time interval is more uniform, causing experiment and testing results become more accurate. After received SPR images, the system will immediately execute a pre-defined image segmentation algorithm to find SPR reactive region. And the mean value of intensity in SPR reactive region will be real-time transmitted to the smart device by Bluetooth. It provides users with the platform of checking testing results.
　　The system can reduce processing time and artificial error by using an automatic image segmentation algorithm, providing real-time and automatic detection. It also increases the accuracy of experiments and sensitivity, which reaches 9.6*10^-6 RIU. Replacing the personal computer with an embedded system greatly reduces the volume and cost of the backend computing device. Moreover, it can be integrated with SPR biosensors into a portable device.

口試委員會審定書 i
誌謝 ii
中文摘要 iii
ABSTRACT iv
目錄 v
圖目錄 vii
表目錄 x
第 1 章 序論 1
1.1 前言 1
1.2 研究動機與貢獻 1
1.3 論文架構 3
第 2 章 基本原理與文獻回顧 4
2.1 表面電漿共振原理簡介 4
2.2 稜鏡式表面電漿共振感測器架構 5
2.2.1 Otto 組態 5
2.2.2 Kretschmann 組態 6
2.3 量測方式 7
2.3.1 角度調變 7
2.3.2 強度調變 7
2.3.3 相位調變 8
2.3.4 波長調變 8
2.4 自動化表面電漿共振影像切割文獻回顧 8
第 3 章 研究方法及步驟 12
3.1 多執行緒之嵌入式系統架構 12
3.1.1 嵌入式系統Raspberry Pi簡介 12
3.1.2 多執行緒 (Multi-threading) 架構 13
3.2 表面電漿共振 (SPR) 影像之擷取 18
3.3 自動化影像切割演算法 19
3.3.1 演算法流程 19
3.3.2 直方圖切割 (Histogram Slicing) 22
3.3.3 大津演算法 (Otsu Algorithm) 26
3.3.4 尋找輪廓及質心 (Find Contour and Centroid) 29
3.3.5 限制對比度適應性直方圖均衡化 (CLAHE) 34
3.3.6 中值濾波器 (Median Filter) 38
3.3.7 基於標記的分水嶺演算法 (Marker-based Watershed Algorithm) 40
3.4 藍芽資料傳輸 44
3.5 Android app開發 45
3.6 實驗設計及量測步驟 48
3.6.1 光學實驗裝置之架設 48
3.6.2 葡萄糖水溶液之濃度檢測實驗 49
3.6.3 多執行緒架構之效能檢驗 49
第 4 章 實驗結果與討論 52
4.1 自動化影像分析系統之操作及使用 52
4.2 自動化影像切割結果 55
4.3 葡萄糖水溶液之濃度檢測實驗結果 59
4.4 多執行緒架構之效能檢驗結果 61
第 5 章 結論與未來發展 63
參考文獻 64

[1]Wood, R.W., On a remarkable case of uneven distribution of light in a diffraction grating spectrum. Proceedings of the Physical Society of London, 1902. 18(1): p. 269.
[2]Gupta, B.D. and R.K. Verma, Surface plasmon resonance-based fiber optic sensors: principle, probe designs, and some applications. Journal of sensors, 2009.
[3]Otto, A., Excitation of nonradiative surface plasma waves in silver by the method of frustrated total reflection. Zeitschrift für Physik, 1968. 216(4): p. 398-410.
[4]Kretschmann, E., Decay of non radiative surface plasmons into light on rough silver films. Comparison of experimental and theoretical results. Optics Communications, 1972. 6(2): p. 185-187.
[5]Ong, B.H., et al., Optimised film thickness for maximum evanescent field enhancement of a bimetallic film surface plasmon resonance biosensor. Sensors and Actuators B: Chemical, 2006. 114(2): p. 1028-1034.
[6]Homola, J., Electromagnetic theory of surface plasmons. Surface plasmon resonance based sensors, Springer, 2006. p. 3-44.
[7]Yan, Y., et al. Molecule binding parameter estimation from Surface Plasmon Resonance measurements. SoutheastCon 2017, IEEE, 2017.
[8]Chen, L., ROI detection in SPR measurements and molecule binding parameter estimation. 2015. Western Carolina University.
[9]Halfacree, G. and E. Upton, Raspberry Pi User Guide. Wiley Publishing, 2012. p. 262.
[10]Silberschatz, A., P.B. Galvin, and G. Gagne, Operating System Concepts. Wiley Publishing, 2008. p. 992.
[11]Yuan, L., et al. New watershed segmentation algorithm based on hybrid gradient and self-adaptive marker extraction. Computer and Communications (ICCC), 2016 2nd IEEE International Conference, 2016.
[12]Otsu, N., A threshold selection method from gray-level histograms. IEEE Transactions on systems, man, and cybernetics, 1979. 9(1): p. 62-66.
[13]Hasan, M.M. and P.K. Mishra, Improving morphology operation for 2D hole filling algorithm. International Journal of Image Processing (IJIP), 2012. 6(1): p. 635-646.
[14]Canny, J., A computational approach to edge detection. IEEE Transactions on pattern analysis and machine intelligence, 1986(6): p. 679-698.
[15]Krutsch, R. and D. Tenorio, Histogram equalization. 2011.
[16]Pizer, S.M., et al., Adaptive histogram equalization and its variations. Computer vision, graphics, and image processing, 1987. 39(3): p. 355-368.
[17]Zuiderveld, K. Contrast limited adaptive histogram equalization. Graphics gems IV, Academic Press Professional Inc, 1994.
[18]Meyer, F. Color image segmentation. in Image Processing and its Applications. International Conference, 1992.
[19]Yeh, Y.-L., Real-time measurement of glucose concentration and average refractive index using a laser interferometer. Optics and Lasers in Engineering, 2008. 46(9): p. 666-670.