本研究基於機器視覺進行自動化光學檢測系統之研究與開發。並採用影像縫合、影像分離、孔位編號，孔位比對與瑕疵顯示等方法進行系統實現。
首先，以影像縫合方法將每次拍照得到的子影像合併為整體影像。接著以影像二值化方式突顯圓孔，並去除明顯雜點。再對孔位進行編號，以區分各圓孔。最後則是由公式計算出圓心、半徑、真圓度等圓孔品質參數，並以長條圖、靶標圖等進行顯示。
在實驗中，對檢測板上所拍攝到每個子影像的大小為37x47.5mm2，將各子影像合併為一完整影。比對用的標準板則由Excellon格式的原始碼中取得，並計算標準板的圓孔品質參數。藉由比對檢測板與標準板，即可獲得其中之差異。在本研究中，我們以長條圖與靶標圖顯示本系統的比對結果。
經由比對結果可知，本研究所開發之系統可以檢測出各圓孔、缺孔、多孔，並顯示錯誤孔位置。這相當適合用在大量鑽孔時的檢測，並可正確的抓出鑽孔瑕疵。

In this research, an Automatic Optical Inspection (AOI) system is studied and developed based on machine vision. The methods of image registration, threshold, labeling, drill contrast, and defect display are adopted in this research. At first, the image registration is used to connect each sub-image to a whole image. Second, the image threshold is used to divide holes with disturbance. Then, labeling is used to label each object on the whole image. The quality parameters are then calculated and displayed on the bar chart and target plot.
For the experiments, the size of sub-image is 37x47.5mm2, and these images are connected to a whole image, called testing board. The quality parameters of standard board are got from the source code in Excellon format. By comparing the quality parameters of testing board and standard board, the comparing result can be known. Therefore, the target plot and bar chart are used to show the inspection results in this research.
The system developed in this research can inspect each hole, and can show the holes with large position error. It can be used in the large drilling inspection, and can inspect correctly.

[1] 黃建程，"銅箔基板雷射鑽孔的機器視覺檢測"，國立台灣科技大學電機工程學系碩士論文，2003.
[2] 洪維達，"區段式影像對正法及其應用"，國立台灣科技大學電機工程學系碩士論文，2003.
[3] 吳家豪，"印刷電路板圓孔檢測方法之研究"，元智大學電機工程學系碩士論文，2003.
[4] 劉宜德，"軟性電路板自動化視覺導引鑽孔系統之研發"，國立中山大學電機工程學系碩士論文，2001.
[5] S. Oprea, I. Lita, I. B. Cioc, D. Alexandru, Visan, “Determination of Misplaced Drill Holes on a PCB” , 30th International Spring Seminar on Electronics Technology, pp406-409, 2007.
[6] K. C. Lee, “A Vision-Measuring Method for an Imperfect Circle”, Journal of the Chinese Institute of Industrial Engineers, Vol. 22, No. 1, pp.38-45, 2005.
[7] W. Y. Wu, C. C. Chen, “A System for Automated BGA Inspection”, Proceedings of the 2004 IEEE Conference on Cybernetics and Intelligent Systems, Vol.2, pp786-790, 2004.
[8] 邱志豪， "PCB鑽孔設備功能及應用"，機械工業雜誌，pp.162-172，2001年4月.
[9] 林宜弘、陳勝男、唐肇蔚、林逸倫，"機器視覺系統之整合性-光源模組設計與應用"，第七屆全國機構與機器設計學術研討會，pp.66-72，2004.
[10] 楊武志，"二值化影像處理-加標示之處理"，影像處理與辨識，全華科技圖書，1994年12月.
[11] A. Gregory, "Digital Image Processing: Principles and applications", Wiley, New York, 1994.
[12] A. Bugareli, "A Simple and Efficient Connected Components Labeling Algorithm", IEEE proc. of 10th International Conference on Image Analysis and Processing, pp.322-327, 1999.