以往對電子連接器的接腳進行檢測大多使用人工目視判斷產品好壞，但隨著產品要求愈精密且微小的情況下，人工檢測已愈難判斷且人力成本高。因此近年來逐漸被機器視覺檢測給取代，並朝著自動化光學檢測的趨勢發展。視覺檢測若是從電子連接器的接腳頂端來擷取影像，往往會因為接腳屬於金屬材質，具有光源高反射特性，影像中會產生亮點及光暈現象，再加上接腳端面面積極小，變成亮點會影響接腳端面特徵，需要額外影像處理技術解決此部分，否則會產生誤判影響檢測結果。為了開發出能消除接腳亮點問題，又能夠得到與接腳端面觀察結果一致的位置度結果，本研究之目的為開發低成本視覺檢測系統，由電子連接器的接腳側面來對連接器接腳進行位置度檢測。為了降低量測成本，本研究使用單一攝影機進行量測。然而使用單攝影機從接腳側面觀察只能看到接腳在單一方向上的位置誤差，無法量測其位置度。因此本研究提出演算法可於接腳側面以二個不同量測角度量測以克服此問題。
最後對兩種連接器各三顆之不同歪斜量的接腳進行位置度檢測，同時使用二維影像量測儀從接腳端面量得的位置度作為參考，與本研究所開發的影像檢測系統計算的位置度做比較，來得知二個系統的檢測差異。本研究所開發之影像系統對第一種連接器的位置度檢測差異為0.06 mm，而第二種連接器的位置度檢測差異為0.08 mm。差異主要來源應是由於從接腳端面和側面所量測到的邊緣不同所致。雖然兩種方法所量測到的結果有些微差異，但結果是一致的。故本研究所開發的方法，可做為當由端面無法量測時的替代方案。

The common practice of the inspection of connector’s pin is through Manual Vision Inspection (MVI), a method of using human eyes. But, the connector’s pins are becoming finer and more delicate. This tendency not only makes this conventional strategy more difficult to implement, but also increases the cost. As a result, machine vision is utilized as a new inspection method and further developing toward Automatic Optical Inspection (AOI).
During the inspection by machine vision, the image taken from top of the connector’s pin will contain light spots due to the high reflectivity of the metallic materials. Moreover, since the top surface area of the pin is relatively small, the light spots would blur edges of the pin. Thus, additional image processing techniques are required to avoid the deterioration of the accuracy of the inspection caused by light spots.
The objective of the research is to develop a low cost machine vision inspection system, which inspects connector’s pins from the side-view of pins in order to solve the problem owing to light spots, meanwhile, obtain the coincident result from the top of the pin. To reduce the cost of the inspection, a single camera is applied in the research. However, a single camera is able to only inspect position deviation in a specific direction of the pin but not in the planar direction. Consequently, an algorithm by inspecting the pin at two various angles from the side of the pin is developed.
Finally, each three of two different types of connectors are inspected in various inclinations. In the meanwhile, two dimensional coordinate measuring machine (2D CMM) is utilized to inspect pins from top-view. Data obtained from 2D CMM is used as a reference to compare with that from the system developed in the research. For the first and second kind of connectors, the difference of the position deviation between two systems is 0.06mm and 0.08mm, respectively. These differences are attributed to the reason that edges are detected differently from top-view or side-view of the pin. Although there is subtle difference between two inspection techniques, the result is consistent. In conclusion, the technique developed in the research is implementable as an auxiliary strategy when the inspection method from top of the pin falls short.

[1]電子時報，企劃，滿足各種應用需求連接器市場與產業化走向全球化競爭，2009年12月.
Available: in
http://digitimes.com.tw/tech/dt/n/shwnws.asp?cnlid=13&packageid=3104&id=0000162419_nda36vbz2t46u788ad9ev
[2]D. Zhao, W. Feng, G. Sun, and Y. Peng, "High precision measurement system of micro-electronic connector based on machine vision," Journal of Applied Sciences, vol. 13, pp. 5363-5369, 2013.
[3]S. Shirmohammadi and A. Ferrero, "Camera as the instrument: the rising trend of vision based measurement," IEEE Instrumentation and Measurement Magazine, vol. 17, pp. 41-47, 2014.
[4]L. Stroppa and C. Cristalli, "Stereo Vision System for Accurate 3D Measurements of Connector Pins’ Positions in Production Lines," Experimental Techniques, vol. 41, pp. 69-78, 2017.
[5]M. Li, D. Weng, Y. Li, L. Zhang, and H. Zhou, "High-accuracy 3D measurement system based on multi-view and structured light," in Proceedings of SPIE - The International Society for Optical Engineering, 2013.
[6]Y. Furukawa and H. Sakuma, "Automatic Detection of Defects among Small Pins Group - Inspection of Connector-Plug Pins by the Use of Image Processing Associated with Neural Network," CIRP Annals - Manufacturing Technology, vol. 41, pp. 589-592, 1992.
[7]G. Luo, H. Fang, Z. Fang, and G. Mu, "Method for the deviation inspection of integrated circuit chip pin based on Harr wavelet transform," Optical Engineering, vol. 39, pp. 1712-1716, 2000.
[8]H. Qiao, Y. Yang, Z.-L. Fang, and F.-L. Liu, "A High-accuracy Measurement of IC Chip Pin Positions Deviation Based on Subpixel Techniques," IEEE International Symposium on Electromagnetic Compatibility, vol. 13, pp. 949-952, 2002.
[9]Q. Zhong, X. Zhang, and Z. Chen, "Online coplanarity measurement of IC leads based on single camera vision system," Transactions of the Institute of Measurement and Control, vol. 38, pp. 1365-1380, 2016.
[10]Q. Zhong, Z. Chen, X. Zhang, and G. Hu, "Feature-based object location of IC pins by using fast run length encoding BLOB analysis," IEEE Transactions on Components, Packaging and Manufacturing Technology, vol. 4, pp. 1887-1898, 2014.
[11]C. J. Tay, X. Kang, C. Quan, X. Y. He, and H. M. Shang, "Height measurement of microchip connecting pins by use of stereovision," Applied Optics, vol. 42, pp. 3827-3831, 2003.
[12]GD&T. Position Tolerance.
Available: in
http://www.gdandtbasics.com/true-position/
[13]MINWT. 鏡頭焦距、等效焦距、感光元件的相關性，2013年.
Available :in
https://www.minwt.com/dc/7693.html
[14]照明節能產品應用手冊，財團法人台灣綠色生產力基金會，2011年.
Available: in
http://www.ecct.org.tw/Knowledge/knowledge_more?id=1597cdb08575446cbcf9f8b8ed7d2b5e
[15]張振堶、金美敬，自動光學檢測市場及技術發展趨勢調查，工研院IEK系統能源組，2002年.
Available: in
http://www.iek.org.tw/pubinfo-detail.screen?type=itispub&domain=0&pubid=2678
[16]莊倩雅，"顯微遠心鏡頭設計" 國立屏東科技大學 機械工程系 碩士論文，2008年.
[17]National Instruments, IMAQ Vision Concepts Manual, 2000.
[18]鍾國亮，影像處理與電腦視覺: 台灣東華書局股份有限公司，2015年8月.