PCB元件焊接對於電子工廠生產線非常重要，波峰焊 (Soldering wave) 品質的好壞會直接影響PCB的良率，其中波峰焊的高度更會直接影響到PCB元件焊接的狀況，高度過高可能會使電路短路，高度過低則可能造成空焊，降低焊接的品質。
　　然而目前在工廠中使用的錫爐大部分只有溫度監控的功能，並沒有錫波高度的監測，目前是請廠務人員利用目測的方式去調整錫波的高度，這個方法過於依賴廠務人員過往對於錫波高度觀測的經驗，缺少科學化的數據，而且也無法即時監控錫波高度。市面上雖有專用的錫波高度量測商用感測器，但它的架設受到嚴格限制，且量測的高度和範圍有限。
　　本論文開發了一套基於樹莓派的影像量測方法，使用Webcam以非接觸式量測錫波高度，主要以動態偵測的方式抓取錫波移動的軌跡，藉由錫波移動的軌跡來計算出當前高度。
　　本論文的實驗分為實驗室環境與工廠環境。在實驗室環境中與商用感測器比較的MAE±SD為0.171±0.115 mm，RMSE為0.217；在工廠環境中與商用感測器比較的MAE±SD為0.213±0.119 mm，RMSE為0.283。

　　PCB components soldering is an important part of the electronics factory production line, and the quality of the wave soldering directly affects the production quality. If the height of soldering wave is too high, the circuit will be shorted. If the height of soldering wave is too low, the circuit will be not connected.
　　However, currently in electronic components factory, the solder machine in the factory only have the temperature monitor, and there is no monitoring of the height of soldering wave. And the height of wave soldering usually estimated by the operator or by using high-temperature resistant glass plate. However, in this situation, it's difficult to control the status of solder machine and keep stability of wave soldering height.
　　Therefore, this paper proposed a method that uses a camera with dynamic detection to monitor the height of soldering wave, which can measure without affecting the operation of the solder machine.
　　The experiments in this paper were divided into laboratory and factory environment. Compared with commercial sensor, the average error was 0.171±0.115 mm(MAE±SD) and RMSE was 0.217 in laboratory environment, and the average error is 0.213±0.119 mm(MAE±SD) and RMSE is 0.283 in factory environment.

[1] The printed circuit report. Accessed: Jan. 22, 2019. [Online]. Available: https://www.prismark.com/printed-circuit-report-pcb.
[2] Taiwan PCB market. Accessed: Jan. 22, 2019. [Online]. Available: http://www.tpca.org.tw/Message?mid=102&itemid=11.
[3] Y. Wang, B. Wang, J. Cai, and T. Wang, “Impact of soldering terminal solderability of component and PCB on solder joint interface,” in Proc. Int. Conf. ICEPT-HDP 2012, Mar. 2012, pp. 877–883.
[4] D. Barbini and J. Bath, “Lead-free wave sodering,” in Springer US, 2011, ch. 3, pp. 45–69.
[5] E. Guene, “Solderability and reliability evolution of no clean solder fluxes for selective soldering,” in Proc. EMPC 2017 - 21st Eur. Microelectron. Packag. Conf. Exhib., Apr. 2018, pp. 1–10.
[6] Q. He, Z. Su, Z. Xie, Z. Zhong, and Q. Yao, “A novel principle for molten steel level measurement in tundish by using temperature gradient,” IEEE Trans. Instrum. Meas., vol. 66, no. 7, pp. 1809–1819, Mar. 2017.
[7] C. Gaber, K. Chetehouna, H. Laurent, C. Rosenberger, and S. Baron, “Optical sensor system using computer vision for the level measurement in oil tankers,” IEEE Int. Symp. Ind. Electron., no. 1, pp. 1120–1124, Nov. 2008.
[8] S. Eppel, “Tracing liquid level and material boundaries in transparent vessels using the graph cut computer vision approach,” arXiv Prepr. arXiv:1501.04691, 2015.
[9] P. Bistak, “Identification and control of hydraulic system using visual feedback,” in Proc. 2016 Int. Conf. Emerg. eLearning Technol. Appl., Jan. 2017, pp. 29–34.
[10] M. Kim, J. Jang, K. Jeong, D. Kim, and J. Paik, “Liquid-level estimation using region-based segmentation for automatic beverage refilling service,” in Proc. Int. Symp. Consum. Electron(ISCE), Aug. 2015, pp. 2–3.
[11] Z. Q. Su, Q. He, Z. Xie, “Molten steel level measurement based on optical flow analysis,” Journal of Northeastern University. Sci., vol. 39, no. 2, pp. 158–161, Feb. 2018.
[12] B. Du, Y. Sun, S. Cai, C. Wu, and Q. Du, “Object tracking in satellite videos by fusing the kernel correlation filter and the three-frame-difference algorithm,” IEEE Geosci. Remote Sens. Lett., vol. 15, no. 2, pp. 168–172, 2018.
[13] S. Y. Chen and S. X. Wang, “Moving object tracking based on five frame difference and improved meanshift algorithm,” Comput. Sci., vol. 43, no. 6A, pp. 203–206, Jun. 2016.
[14] Z. Xu, D. Zhang and L. Du, “Moving object detection based on improved three frame difference and background subtraction,” in Proc. Int. Conf. Ind. Informatics - Comput. Technol. Intell. Technol. Ind. Inf. Integr., 2017, pp. 79-82.
[15] X. J. Zhang and H. Xu, “Moving vehicle detection algorithm based on video processing,” Chinese J. Liq. Cryst. Displays, vol. 27, no.1, pp. 108-113, Feb. 2012.
[16] W. A. Kaminski and G. M. Wojcik, “Liquid state machine built of hodgkin – huxley neurons,” Neurocomputing, vol. 239, pp. 245-251, 2004.
[17] D. S. Suresh and M. P. Lavanya, “Motion detection and tracking using background subtraction and consecutive frames difference method” Int. J. Res. Stud. Sci. Eng. Technol., vol. 1, no. 5, pp. 16-22, Aug. 2014.
[18] H. Wang, S. K. Nguang, and J. Wen, “Robust video tracking algorithm: a multi-feature fusion approach,” IET Comput. Vis., vol. 12, no. 5, pp. 640–650, Jul. 2018.
[19] J. Guo, J. Wang, R. Bai, Y. Zhang, and Y. Li, “A new moving object detection method based on frame-difference and background subtraction,” in Proc. IOP Conf. Ser. Mater. Sci. Eng., pp. 1–4, 2017.
[20] T. Bouwmans, “Traditional and recent approaches in background modeling for foreground detection: An overview,” Comput. Sci. Rev., vol. 11–12, pp. 31–66, May 2014.
[21] J. R. Evans and W. M. Lindsay, “Managing for quality and performance excellence, ” in Cengage Learning, Jan. 2007, pp. 574.
[22] Wave soldering defects. Accessed: Apr. 11, 2019. [Online]. Available: https://www.epectec.com/pcb/wave-soldering-defects/.
[23] Soldering wave wikipedia. Accessed: Apr. 11, 2019. [Online]. Available: https://en.wikipedia.org/wiki/Wave_soldering.
[24] R. H. Todd, D. K. Allen and L. Alting, “Manufacturing processes reference guide,” in Industrial Press, 1994, pp. 393-394.
[25] K. H. J. Buschow et al., “Encyclopedia of materials: science and technology,” in Elsevier, 2001, pp. 2708-2709.
[26] Dual in-line package wikipedia. Accessed: Apr. 18, 2019. [Online]. Available: https://en.wikipedia.org/wiki/Dual_in-line_package.
[27] Eddy current wikipedia. Accessed: Apr. 18, 2019. [Online]. Available: https://en.wikipedia.org/wiki/Eddy_current.
[28] Y. Wang et al, “CDnet 2014: An expanded change detection benchmark dataset,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. Workshops, Jun. 2014, pp. 387–394.
[29] Y. Zhang, X. Wang, and B. Qu, “Three-frame difference algorithm research based on mathematical morphology,” Procedia Engineering, vol. 29, no. 4, pp. 2705–2709, 2012.
[30] N. Otsu, “A threshold selection method from gray-level histogram,” IEEE Trans. Syst. Man Cybern., vol. 9, pp. 62–66, Jan. 1979.
[31] “Raspberry pi 3 Model B plus website” Accessed: Jul. 9, 2019. [Online]. Available: https://www.raspberrypi.com.tw/19429/57/
[32] “Matlab calibration” Accessed: Jun. 13, 2019. [Online]. Available: https://www.mathworks.com/help/vision/ug/camera-calibration.html.
[33] Z. Zhang, “A flexible new technique for camera calibration,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 22, no. 11, pp. 1330–1334, Nov. 2000.
[34] J. Canny, “A computational approach to edge detection,” IEEE Trans. Pattern Anal. Mach. Intell., vol. PAMI-8, no. 6, pp. 679–698, Nov. 1986.
[35] “OpenCV canny edge detector” Accessed: Jul. 10, 2019. [Online]. Available: https://docs.opencv.org/3.4/da/d5c/tutorial_canny_detector.html
[36] “Sobel operator wikipedia” Accessed: Jul. 10, 2019. [Online]. Available: https://en.wikipedia.org/wiki/Sobel_operator
[37] “Mean absolute error wikipedia” Accessed: Aug. 7, 2019. [Online]. Available: https://en.wikipedia.org/wiki/Mean_absolute_error
[38] “Root mean square error wikipedia” Accessed: Aug. 7, 2019. [Online]. https://en.wikipedia.org/wiki/Root-mean-square_deviation