臺灣博碩士論文加值系統

在許多電機零組件之加工程序中，因使用高速沖壓的加工方式，其工作環境相當吵雜，不利於長期處於此環境中。噪音對工作人員來說是一項非常大的職業傷害，因此，在此種環境下發展自動化生產技術與自動化檢測技術將是未來之趨勢。
本研究提出以機器視覺及類神經網路之方法來進行產品之品質檢驗。我們利用攝影機擷取產品之數位影像，經過適當的影像前處理後擷取產品之重要特徵。這些特徵將作為類神經網路之輸入來進行物理特性之量測及允收、拒收之判斷。
本論文是以電機零組件S鐵心為研究對象，我們考慮之物理特性包含：(1) S鐵心之內徑；(2) S鐵心積片後之厚度；(3) S鐵心積片後之厚度差。本研究所提出之自動化檢測系統是以工廠所收集到之產品，來進行類神經網路之訓練及測試。實驗結果顯示，本研究所提出之檢驗方法可迅速與精確地量測產品之重要品質特性。

In the manufacturing of electrical components, it is not unusual that some working environments are hazardous and not suitable for human beings. In this situation, development of an automatic inspection system is necessary. In this research, a neural network-based inspection system was developed to perform the automatic inspection of product characteristics. Some important features extracted from the results of image processing were used as the inputs to neural network.
In this research, an electrical component, S-rotor, is used as the example to demonstrate the neural network-based inspection system. The quality characteristics considered in this research include: (1) the diameter; (2) the thickness; (3) the difference of thickness at both ends.
Data collected from factory were used for training and testing of neural network. Experimental results indicate that the neural network-based inspection system could accurately measure the quality characteristics.

一、緒論 1
1.1 研究背景與動機 1
1.2 研究目的 2
1.3 研究範圍 2
1.4 評估方式 3
1.5 研究方法與步驟 3
二、文獻探討 7
2.1 影像處理 7
2.1.1 二值化影像 7
2.1.2 灰階值長條圖 9
2.1.3 自動臨界值擷取 9
2.1.4 邊界的偵測 11
2.2 數學運算法則 11
2.2.1 求圓的方法 12
2.2.2 求直線的方法 17
2.3 幾何公差 20
2.3.1 厚度差 21
三、類神經網路 22
3.1 倒傳遞類神經網路之架構 22
3.2 倒傳遞類神經網路之演算過程 23
3.3 倒傳遞類神經網路之演算法 24
3.4 倒傳遞類神經網路應用之要點 29
3.4.1 初始權數 29
3.4.2 網路大小 30
3.4.3 學習速率 30
3.4.4 網路學習終止條件 31
3.4.5 訓練樣本數 31
四、研究方法與實驗結果 32
4.1 影像前處理 32
4.1.1 邊緣追蹤之原理 32
4.2 尋找圓或直線 33
4.2.1 重心法求圓心 34
4.2.2 線性迴歸法求直線以測積長與厚度差 35
4.3 類神經之訓練與測試 35
4.3.1 類神經網路估算內徑 36
4.3.2 類神經網路估算積長與厚度差 38
4.4 實例說明實驗結果 39
4.5 積長與內徑預測之敏感度分析 47
4.5.1 積長之敏感度分析 47
4.5.2 內徑之敏感度分析 53
五、結論與未來展望 56
5.1 結論 56
5.2 未來展望 57
參考文獻 58

1. Bookstein, F. L., ''Fitting conic sections to scattered data," Computer Graphics and Image Processing 9, 56-71 (1979).
2. Chaudhuri, B. B. and P. Kundu, "Optimum circular fit to weighted data in multi-dimensional space", Pattern Recognition Letters 14, 1-6 (1993).
3. Duda, R. O. and P. E. Hart, "Use of the Hough transformation to detect lines and curves inpictures," Comm. Assoc. Comput. Mach. 15, 11-15 (1972).
4. Davies, E. R., "Image space transforms for detecting straight edges in industrial image," Pattern Recognition Letters, Vol. 4, 185-192 (1986).
5. Draper, N. R. and H. Smith, Applied Regression Analysis. 2nd ed. New York: John Wiley & Sons, (1981).
6. Fairhurst, M. C., Computer Vision for Robotic System-An Introduction, Prentice Hall, (1988).
7. Gonzalez, R. C. and R. E. Woods, Digital Image Processing, Addison-Wesley Publishing Company, (1992).
8. Hough, P. V. C., "A method and means for recognizing complex patterns," U.S.Patent 3.069.654, (1962).
9. Joseph, S. H., "Unbiased least squares fitting of circular arcs", GVGIP:Graphical Models and Image Processing 56, 424-432 (1994).
10. Kierkegaard, P., "A method for detection of circular arcs absed on the Hough transform," Machine Vision and Applications 5, 249-263 (1992).
11. Landau, U. M., "Estimation of a circular arc center and its radius," Computer Vision, Graphics, and Image Processing 38, 317-326 (1987).
12. Ostu, N., "A threshold selection method from gray-level histogram," IEEE Trans. Systems Man Cyberentics, Vol. SMC-9, No. 1, 62-66 (1979).
13. Shunmugam, M. S., "Comparison of linear and normal deviations of forms of engineering surfaces," Precision Engineering 9, 96-102 (1987).
14. Sampson, P. D., "Fitting conic sections to very scattered data: an iterative refinement of the Bookstein algorithm," Computer Graphics and Image Processing 18, 97-108 (1982).
15. Thomas, S. M. and Y. T. Chan, "A simple approach for the estimation of circular arc center and its radius," Computer Vision, Graphics, and Image Processing 45, 362-370 (1989).
16. Yi, S., R. M. Haralick, and L. G. Shapiro, "Error propagation in machine vision," Machine Vision and Applications, Vol. 7, 93-114 (1994).
17. 鄭春生, 品質管理, 三民書局, 中華民國84年.
18. 李清安, 機械製圖實務, 全華科技圖書股份有限公司, 中華民國76年.
19. 許忠雄, 應用電腦視覺於電機零組件品質之自動檢測, 新竹, 交通大學, 工業工程研究所碩士論文, 中華民 國82年.
20. 曾佳怡, 數位影像中直線的量測, 中壢, 元智工學院, 工業工程研究所碩士論文, 中華民國83年.
21. 洪志宏, 應用類神經網路估計數位影像中圓弧之圓心及半徑, 中壢, 元智工學院, 工業工程研究所碩士論文 , 中華民國84年.
22. 江行全, 自動視覺檢測技術, 上課用講義, 中華民國86年.
23. 葉怡成, 類神經網路模式應用與實作, 儒林書局, 中華民國87年.