隨著現代科技與理論的快速發展，近年來消費性電子產品越趨普及且智慧化，消費 者對於電子產品的要求越來越高，包括功能的多樣性、電池的續航力及攜帶的方便性 等，因此，電子產品的輕量化及薄型化成為主要的發展趨勢，其中為了裝載更多電子元 件，印刷電路板（printed circuit board，PCB)必須朝向多層設計發展且其佈線愈趨細 微，亦即鑽孔孔徑越來越微小，使得微型鑽頭成為PCB製程中的重要耗材之一，因此，微型鑽頭再研磨的需求量與重要性在PCB產業中與日倶增，為求降低成本及穩定產品 的供貨量和品質，自動化再研磨機台的發展受到廣泛的重視。本文的研究目的即運用光 學取像和數位影像處理技術，發展微型PCB鑽頭再研磨機台的自動化光學檢測系統，其中主要的研究内容包含二個部份：第一部份計算刀面影像的缺陷值與檢驗的標準值進 行缺陷檢測，以判定鑽頭再研磨後的良劣，期望能依據缺陷種類反向調整機台的參數設 定；第二部份為嘗試使用類神經網路訓練分類器進行缺陷檢測及分類，期望能降低缺陷 檢測分類的時間，提升整體研磨的效率。根據實驗結果，本文發展的自動化光學檢測系 統能成功檢驗出9種缺陷，分類正確率可達92%以上。

With the rapid development of modern technologies and theories, consumer electronic products are much more popular and smart than before. Hence, customers demand features of high-quality, multi-function, easy-use, and long battery life to fit in with electronic products.
Therefore, slim and lightweight electronic devices become a major trend. To mounting more electronic components on the printed circuit board (PCB), multi-layer designs and subtle wiring technique must be developed under the requirement of holes drilling smaller and improvement to re-sharpen drill bits. That is, in order to cost down and ensure high quality and quantity towards products, PCB drill bits development of automatic grinding machine has widely gained attention. The purpose of this study is to develop automatic optical inspection system for PCB micro drill bits re-sharpening machine by using optical image capturing and digital image processing technologies with two major concerns. The first one is to identify the performance of regrinding drill bits by calculating and comparing the defects in drill bit images with the inspection standards. The second one is to use the artificial neural network training tool to examine and classify PCB drilling defects so that improvement for regrinding efficiency could be made. Experiment results show that in this study automatic optical inspection system can identify 9 types of defects with more than 92% of accuracy.