本論文研究目的為使用電腦視覺對射出成型(Injection Moulding)的驗孕片外殼進行自動化瑕疵偵測，並結合SCARA機械手臂吸取輸送帶上移動的驗孕片外殼進行瑕疵篩選。本論文實驗中，驗孕片外殼由小型輸送帶運送模擬現實工廠的流水線環境，其中輸送帶被設計成前後高低兩段，兩段中間放置自行印製的3D列印斜面擋板模型，藉此達到驗孕片外殼掉落時自動翻面的機制。兩台工業攝影機分別拍攝驗孕片外殼的正反面影像進行瑕疵檢測，本實驗中需檢測的四種瑕疵分別為：油漬、黑線黑點、裁剪殘料、缺料。為去除驗孕片外殼以外的環境背景，使用影像二值化與連通物件法對影像進行前處理，而對每種瑕疵分別採用不同的演算法進行分析與偵測，其中利用了HSV色彩空間分離油漬與驗孕片外殼背景以偵測油漬瑕疵；使用了Neighboring Difference Filter (NDF)圖像濾波器，有效的抓取模糊表面上的瑕疵，以偵測黑線黑點瑕疵；使用道格拉斯-普克近似輪廓演算法判斷邊緣是否有殘料殘留，以偵測裁剪殘料瑕疵；先透過驗孕片外殼面積比例分析是否為本體缺料，再於HSV色彩空間偵測接腳部分分析是否為斷腳缺料, 以偵測缺料瑕疵。檢測後，若驗孕面外殼任一面被檢測出瑕疵存在，則命令SCARA機械手臂吸取驗孕片外殼進行瑕疵篩選。最後將討論論文實驗中四種瑕疵偵測方法於不同參數下的偵測結果與準確率。

The main purpose of this thesis is to use computer vision to automatically detect the shell defects on the pregnancy test kit, and use SCARA robot to pick up and exclude the defective products of them on the moving conveyor belt for realizing the filtering function. In this study, the shell of the pregnancy test kit is transported by a small conveyor belt, to simulate a real factory assembly line. The conveyor belt is designed as two sections: the front section is higher and the rear section is lower. Then when the shell of the pregnancy test kit is transported from high section and drops to the low section, at the same time, it will turn over from one side to the reverse side because of a baffle made by 3D printer at the junction of two sections. Two industrial cameras take the photo of the shell at the high and low sections respectively. Therefore, four kinds of defects need to be detected, which are black spots and lines, oil stains, cutting residual materials, and lack of materials. In order to remove the background outside the shell, the image is pre-processed with binarization and connected-component labeling. It is noted that different defects are analyzed and detected by different algorithms, respectively. For instance, the oil stain and the background can be separated in HSV color space. Black spots and lines are detected by the Neighboring Difference Filter, which can effectively detect defects on blurred surfaces. The cutting residual on the edge is detected by the Douglas-Puck contour algorithm. In the case of the lack of materials, the proportion of the area of the pregnancy test kit is firstly analyzed to determine whether the body is short of material, and then analyzed whether the color of the pin in the HSV color space is broken. If defects are detected on any surfaces, the image taken by the industrial camera on the low section of the conveyor belt will be applied to calculate its moving speed and position coordinate X and Y at the current time. After then, the SCARA robot arm is commanded to pick the shell up if any defect is found. Finally, there are some experiments to show the proposed algorithms are effective for the defects detection of shells and the accurate operations of the SCARA robot.

[1] Z. Yu, X. Li, H. Yu, D. Xie, A. Liu and H. Lv, "Research on surface defect inspection for small magnetic rings," in Proc. International Conference on Mechatronics and Automation, Changchun, 2009, pp. 994-998.
[2] H. Tao, Z. Jinshan, D. Na, C. Zhihui and D. Li, "The defects identify algorithm and the automated detecting system for the screw thread," in Proc. International Conference on Electrical and Control Engineering, Wuhan, 2010, pp. 2292-2294.
[3] C. Wang, F. Li, Y. Li, H. Chen and X. Cao, "A defect status detecting method for external gear in railway," in Proc. IEEE 3rd International Conference on Image, Vision and Computing, Chongqing, 2018, pp. 123-127.
[4] Y. Park and I. S. Kweon, "Ambiguous surface defect image classification of AMOLED displays in smartphones," IEEE Transactions on Industrial Informatics, vol. 12, no. 2, pp. 597-607, 2016.
[5] F. Li, Z. Hang, G. Yu, G. Wei and C. Xinyu, "The method for glass bottle defects detecting based on machine vision," in Proc. The 29th Chinese Control And Decision Conference, Chongqing, 2017, pp. 7618-7621.
[6] J. Lu and X. Hu, "Detecting defects on citrus fruit surface by circularity gradient threshold segmentation," in Proc. The 13th International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery, Guilin, 2017, pp. 2312-2316.
[7] M. Song, X. Zhou, Y. Cai and X. Mou, "Algorithm for cheese yarn defects detecting based on multi-directional matching filter," in Proc. IEEE 3rd International Conference on Image, Vision and Computing, Chongqing, 2018, pp. 696-700.
[8] A. S. Siva Sarma, R. Janani and A. S. V. Sarma, "Detecting the surface defects on hot rolled steel sheets using texture analysis," in Proc. International Conference on Advanced Electronic Systems, Pilani, 2013, pp. 157-159.
[9] Y. F. Meng, S. R. Gong and C. P. Liu, "A fast computer vision system for defect detection of rubber keypad," in Proc. International Conference on Computer Application and System Modeling, Taiyuan, 2010, pp. 155- 160.
[10] X. Yang and S. Wu, "A rapid defect detecting algorithm for printed matter on the assembly line," in Proc. International Conference on Systems and Informatics, Yantai, 2012, pp. 1842-1845.
[11] G. S. Huang and Z. H. Tian, "Application of a SCARA manipulator with a visual system in industries," International Journal of Computer, Consumer and Control, vol. 4, no. 4, pp. 9-17, 2015.
[12] C. Hu, C. Chen, C. Tseng, A. P. Yudha and C. Kuo, "Visual servoing spanner picking and placement with a SCARA manipulator," in Proc. IEEE International Conference on Industrial Technology, Taipei, 2016, pp. 1632-1637.
[13] A. Nagchaudhuri, S. Kuruganty and A. Shakur, "Introduction of mechatronics concepts in a robotics course using an industrial SCARA robot equipped with a vision sensor," Mechatronics, vol. 12, pp. 183-193, 2002.
[14] H. Zhu, J. Xu, D. He, K. Xing and Z. Chen, "Design and implementation of the moving workpiece sorting system based on LabVIEW," in Proc. The 26th Chinese Control and Decision Conference, Changsha, 2014, pp. 5034-5038.
[15] C. Y. Tsai, C. C. Wong, C. J. Yu, C. C. Liu and T. Y. Liu, "A hybrid switched reactive-based visual servo control of 5-DOF robot manipulators for pick-and-place tasks," IEEE Systems Journal, vol. 9, no. 1, pp. 119-130, 2015.
[16] J. Cunha, R. Ferreira and N. Lau, "Computer vision and robotic manipulation for automated feeding of cork drillers," Materials and Design, vol. 82, pp. 290-296, 2015.
[17] J. M. Lee, B. S. Park, Y. S. Lee, J. S. Ahn, S. H. Lee, S. J. Lim, and C. S. Han, "The development of the robot manipulator for an intelligent service robot," in Proc. Proceedings of SICE-ICASE International Joint Conference, Busan, 2006, pp. 282-287.
[18] H. Nomura and T. Naito, "Integrated visual servoing system to grasp industrial parts moving on conveyer by controlling 6DOF arm," in Proc. IEEE International, 2000, pp. 1768-1775.
[19] Y. Sahin, M. Tinkir, and A. Ankarali, "Neuro-fuzzy trajectory control of a SCARA robot," in Proc. Proceeding of the IEEE International Conference on Computer and Automation Engineering, Singapore, 2010, pp. 298-302.
[20] H. S. Ali, L. Boutat-Baddas, Y. Becis-Aubry, and M. Darouach, "H-infinity control of a SCARA robot using polytopic LPV approach," in Proc. The 14th Mediterranean Conference on Control and Automation, Italy, 2006, pp. 1-5.
[21] 華碩官方網站，2019年6月。
Available at：https://www.asus.com/tw/
[22] Blackfly 3.2 MP Color GigE PoE (Sony IMX265), June 2019.
Available at：https://www.ptgrey.com/blackfly-32-mp-color-gige-poe-sony-imx265-2
[23] 台達電(SCARA工業機器人)，2019年6月。
Available at：http://www.deltaww.com/Products/CategoryListT1.aspx?CID=060601&PID=ALL&hl=zh-TW
[24] 台灣氣立CHELIC – 氣源處理元件，2019年6月。
Available at：http://www.chelic.com/website/tw/products-MOin-3.html
[25] 台灣氣立CHELIC – 控制閥類，2019年6月。
Available at：http://www.chelic.com/website/tw/SO-series.html
[26] Vacuum Generator VU - Product - PISCO PNEUMATIC EQUIPMENT, June 2019.
Available at：https://www.pisco.co.jp/product/detail/e/e02/
[27] MP7528 5W USB-LED磁鐵飛輪燈 - 明沛實業有限公司，2019年6月。
Available at：http://www.mpbattery.com.tw/product_info.php?products_id=979
[28] Connected-component labeling – Wikipedia, June 2019.
Available at：https://en.wikipedia.org/wiki/Connected-component_labeling
[29] Image Thresholding - OpenCV Documentation, June 2019.
Available at：https://docs.opencv.org/master/d7/d4d/tutorial_py_thresholding.html
[30] HSL and HSV – Wikipedia, June 2019.
Available at：https://en.wikipedia.org/wiki/HSL_and_HSV
[31] Emgu CV: OpenCV in .NET (C#, VB, C++ and more), June 2019.
Available at：http://www.emgu.com/wiki/index.php/Main_Page
[32] Bilateral filter – Wikipedia, June 2019.
Available at：https://en.wikipedia.org/wiki/Bilateral_filter
[33] Top-hat transform – Wikipedia, June 2019.
Available at：https://en.wikipedia.org/wiki/Top-hat_transform
[34] Ramer–Douglas–Peucker algorithm – Wikipedia, June 2019.
Available at：https://en.wikipedia.org/wiki/Ramer%E2%80%93Douglas%E2%80%93Peucker_algorithm
[35] Projection matrix – Wikipedia, June 2019.
Available at：https://en.wikipedia.org/wiki/Projection_matrix
[36] GitHub - tzutalin/labelImg: LabelImg is a graphical image annotation tool and label object bounding boxes in images, June 2019.
Available at：https://github.com/tzutalin/labelImg