透過您的圖書館登入
IP:44.200.169.91
  • 學位論文

混合式模型在電路板參數調校及缺陷分類問題的發展

The development of hybrid model in PCB parametric regulation and defect classification problems

指導教授 : 張百棧
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


生產及設備製造業不斷地被要求精進,將工廠中的各項設施維持在最佳狀態,並期待未來有更精良的設備。藉由診斷而提供的進階智能,能透過維修策略將資產有效管理,並掌握關鍵元件防止無謂的停機損失。電腦輔助維護取決於 感測器系統訊號的輸入,來提供狀態的監控,在檢測的領域中,影像辨識更是首選的步驟。 本研究透過案例式推導(Case based reasoning, CBR)與基因演算法(Genetic algorithms, GAs)整合的調機系統,應用在雙鐳射量測系統線性度校正問題的探討。當應用在厚度的測量時,實驗結果與其他方法相比較,GA/CBR的混合模型最準確和有效率。本研究的另一個主題,案例式發展的識別法於印製電路板線路圖像的缺陷特徵分割,該模型應用在一個視覺檢查系統的瑕疵分類問題上。實驗結果顯示,案例式發展的演化模型,因為使用案例庫比對以往相類似的問題,所以使用者於缺陷辨識中能節省很多時間,並提出對應問題的處置對策。當一個新的缺陷被發現時,應用案例學習的程序,可將修正過的新缺陷在案例庫中建立新類別。 為了強調預測式的維護工作,工程師和研究人員正把人工智慧應用於電腦輔助的維護。本研究提出的案例演化模型,非常適用於機台狀態監控及製程的維護,詳細的實驗討論及研究結果也在這篇論文裡提出。

並列摘要


Manufacturing and utility industries will continuously be challenged to show insight in optimizing current plant assets and anticipating future asset requirements. The advanced intelligence provided by diagnostics can be the decision factor in effective asset management through maintenance strategies and improved operations aimed at heading off availability losses and forced outages. Computer-aided maintenance depends on sensory systems to provide signal inputs for condition monitoring. Image understanding is the first and foremost step in the inspection. In this research, a hybrid regulation system by evolving case-based reasoning (CBR) with genetic algorithms (GA) is developed for linear regulation problem of a twin laser measuring system. The experimental results show that the GA/CBR model is more accurate and efficient when being applied to the measuring of the thickness than other methods. In another topic, case-based evolutional identification for defect feature segmentation of printed circuit board (PCB) images is developed for a visual inspecting system of defect classification problem. The experimental results show that the CBR evolutionary model compares the similar problem in the case-base, so user can save much time to review the defect and submit the according strategy. When a new defect case is found, the learning procedure will revise the repaired defect into the new defect case. With increasing emphasis on predictive maintenance, engineers and researchers are applying artificial intelligence to computer-aided maintenance. Case-based evolutionay model proposed in this study is very suitable for machine condition monitoring and process maintaining. A detailed discussion of the experiments and the results obtained is also presented in this dissertation.

參考文獻


[1]Spinola, C., Vazquez, M. J. M., Bohorquez, A.F., Bonelo, J.M., and Vizoso, J., “Calibration of thickness measurement instruments based on twin laser sensors. Isoline bilinear look up tables,” Instrumentation and Measurement Technology Conference, IMTC Proceedings of the 18th IEEE, Volume: 2, pp. 1079 -1083, 2001.
[3]Shapiro, L. G., “A Structural Model of Shape,” IEEE Trans. Pattern Anal. Machine Intell., Vol. PAMI-2, No. 2, Mar. 1980, pp. 111-126, 1980.
[4]Moganti, M., Ercal, F., Dagli, C. H. and Tsunekawa, S., “Automatic PCB Inspection Algorithms: A Survey, Computer Vision and Image Understanding,” Vol. 63, No. 2, pp. 287-313, 1996.
[5]Goodall, A. J. E. and Lo, E. K., “A review of inspection techniques applicable to PCB manufacturing and assembly, particularly with respect to SMT,” Advanced Manufacturing Engineering, 3, pp. 18-26, 1991.
[8]Bayro-Corrochano, E., “Review of automated visual inspection 1983-1993 - part 1: Conventional approaches,” Intelligent Robots Computer Vision, pp. 128-158, 1993.

延伸閱讀