微型PCB鑽頭再研磨機台自動化光學檢測系統之發展

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產業中與日俱增，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，其中主要的研究內容包含二個部份：第一部份計算刃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。

關鍵字

PCB鑽頭再研磨；自動化光學檢測；類神經網路；分類器

並列摘要

With the rapid development of modern technologies and theories, consumer electronics products are much more popular and intelligent in recent years. Hence, there are increasing demands from customers for electronic products, including high-quality, multi-function, longer battery life, and more convenient for carrying out. Therefore, thin and lightweight electronic products become a major development trend. For mounting more electrical components on the printed circuit board (PCB), its development must be toward multi-layer designs and its wiring is growing increasingly subtle. That is, drilling holes become smaller in size and PCB drill bits play an important role in the PCB manufacturing process. Requirement of drill bits re-sharpening has become increasingly important for the PCB industry. In order to reduce costs and ensure the high quality and sufficient quantity of its products, the development of automatic grinding machine for PCB drill bits has gained wide attention. The purpose of this study to develop the automated optical inspection system for the PCB micro drill bits re-sharpening machine by using optical image capturing and digital image processing technologies. There are two major parts of this study. The first part is to determine the regrinding drill bits good or bad by compared with calculating defects in the image of drill bit and product inspection standards. The second part is to use the training classifiers of artificial neural network for examining defects, making classifications and evaluating how they perform in promoting productive efficiency. Experimental results show that the presented automated optical inspection system in this study can identify 9 types of defects, namely the percentage of correctly classified is more than 92%.

並列關鍵字

PCB drills regrinding ； Automated optical inspection ； Neural network classifier

參考文獻

[13] 呂雅鈞, PCB鑽頭再研磨自動化機台之光學檢測, 中原大學機械工程學系碩士論文, 2012.

[4] R. C. Gonzalez and R. E. Woods, Digital Image Processing. Reading, Mass.: Addison-Wesley, 1992.

[14] W. S. McCulloch and W. H. Pitts, “A logical calculus of the ideas imminent in nervous activity,” Bulletin of Mathematical Biophysics, Vol. 5, pp. 115-133, 1943.

[15] D. O. Hebb, The Organization of Behavior: A Neuropsychological Theory, Wiley, New York, 1949.

[16] F. Rosenblatt, “The perceptron: a probabilistic model for information storage and organization in the brain,” Psychological Review, Vol. 65, No. 6, pp. 386-408, 1958.

國際替代計量

微型PCB鑽頭再研磨機台自動化光學檢測系統之發展

未授權

主題瀏覽