臺灣博碩士論文加值系統

目前存在於台灣製造業的主要產品檢測技術，絕大部分仍停留在人工檢測的方式，雖有國內廠商引進影像處理之檢測系統，但仍得依賴其技術支援，無法真正掌握系統之關鍵技術。有鑑於此，本研究是以縮短產品製程時效、節省成本為主要目的，提出一套結合影像處理技術與類神經網路的辦識系統。
於本研究方法中，前置處理先將影像正規化並分離出RGB三原色之分色圖，下一步則對分離後之分色圖各別進行臨界值處理（Thresholding），進而求得特徵值所在點，最後結合類神經網路的訓練方式，形成一自動辨識系統，以提升整體辨識速度及效能。本研究爰提信用卡為實證範本，透過信用卡成品中之良品和不良品的辨別，驗證出本研究方法之可行性，以達成信用卡成品製造之品質管制的目標。
本計畫所研發之影像處理系統，於多樣範本不斷測試下，其影像辨識率已達業界評比標準之上。未來其以更完善的模組化、物件工具設計來因應不同類型檢測工作之彈性需求，希冀在產品製造流程中達到成品品質之嚴格控管。

At present, most of the manufacturing industries in Taiwan use the man-made inspection for their certain finished products inspection. Although a few firms import the digital image processing techniques from the foreign, it still needs techniques support. In consideration of this disadvantage, we propose an identification system that combines the digital image processing techniques and the artificial neural network to shorten the procedure and reduce the cost of the personnel matters.
At this study, the image normalization and the transformation of RGB components are performed. Next, the thresholding is used to find the features of each component. This study takes the credit card as the example. Using the identification of the credit card to verify the probability of applying this study to achieve the goal of the quality management of the finished products of the credit card manufacturing.
The developed inspection system has been tested on a case study from a credit card manufacture located in Tainan to verify the functionality and superiority of the developed system.

中文摘要 Ⅰ
英文摘要 Ⅱ
目錄 Ⅲ
表目次 Ⅴ
圖目次 Ⅵ
第一章 緒論 1
第一節 研究背景與問題 1
第二節 研究目的 1
第三節 研究範圍 2
第四節 研究內容 2
第五節 研究流程 3
第二章 文獻探討 5
第一節 影像處理之研究與論述 5
第二節 類神經網路研究與輿論 6
第三節 文獻討論 7
第三章 類神經網路理論與架構 8
第一節 類神經網路定義 8
一、 類神經網路基本定義 8
二、 類神經網路之分類 10
三、 類神經網路的運作原 11
四、 類神經網路之特性 11
第二節 類神經網路的基本架構 12
一、 網路架構 12
二、 處理單元 13
三、 層架構 15
四、 網路 16
第三節 倒傳遞演算法 17
一、 學習過程 17
二、 回想過程 19
三、 參數的設定 20
四、 範例處理 21
頁碼
第四章 研究方法 23
第一節 影像資料處理與分析 23
一、 影像擷取 23
二、 影像調整與裁切 24
三、 影像分離 25
四、 特徵值的擷取 28
第二節 影像檢測模組 36
一、 影像檢測模組之選擇 36
二、 倒傳遞網路之基本理論 37
三、 影像檢測模式之構建 38
第五章 實驗測試與結果 42
第一節 硬體環境 42
第二節 樣本的類型與數量 42
一、 樣本的選取 42
二、 樣本的類型與數量 43
第三節 影像分析與處理 43
一、 影像擷取與裁切 43
二、 RGB三原色頻譜分離 44
三、 特徵值之擷取 45
第四節 類神經網路設定 46
一、 實驗設計 46
二、 模式評估指標 47
三、 預測模式之訓練與測試 47
四、 實驗結果與比較 52
第六章 研究結論與未來研究方向 54
第一節 研究結論 54
第二節 研究缺失與侷限 54
第三節 未來展望 55
第四節 總結 55
參考文獻 56
附錄一、 影像特徵擷取程式 63
附錄二、 類神經網路主程式 64
附錄三、 貝氏定理影像辨識程式 67
表目次
表3.1 轉換函數形式 14
表4.1 類神經網路模式比較表 36
表5.1 統計直方圖轉換 43
表5.2 訓練樣本型態 44
表5.3 影像分離與二值化 44
表5.4 輸入特徵值之樣本 45
表5.5 標準化後的Z值 46
表5.6 N= 2預測模組績效值 48
表5.7 N= 4預測模組績效值 49
表5.8 N= 8預測模組績效值 49
表5.9 N= 16預測模組績效值 50
表5.10學習率變動之績效值 51
表5.11 實驗樣本之辨識率 52
表5.12 實驗辨識率之比較一 53
表5.13 實驗辨識率之比較二 53
圖目次
圖1.1 研究流程圖 4
圖3.1 人工神經元模型 9
圖3.2 類神經網路架構圖 12
圖3.3 處理單元的作用 13
圖4.1 本系統之研究流程 23
圖4.2 硬體架構 24
圖4.3 影像調整與裁切 24
圖4.4 單位元影像 26
圖4.5 灰階影像 27
圖4.6 RGB彩色模型 27
圖4.7 RGB三原色的灰階圖 29
圖4.8 圖形二值化轉換 29
圖4.9 固定臨界值二值化基本示意 31
圖4.10灰階值等分區間示意圖 34
圖4.11 特徵值擷取示意圖 35
圖4.12 影像檢測網路架構 39
圖5.1 樣本影像 42

一、 中文文獻
陳飛龍、唐為璽（1999），「結合影像處理與類神經網路之沖壓件檢測技術」，Journal of the Chinese Institute of Industrial Engineers , Vol.16, No.2, pp.253-263.
江行全、林泰宗、王建智（1999），應用機器視覺系統於表面瑕疵分類之研究，Journal of the Chinese Institute of Industrial Engineers , Vol.16, No.4, pp.443-453.
侯東旭、彭銘德（1999），使用影像投影法與類神經網路於外型辨識之研究，Journal of the Chinese Institute of Industrial Engineers , Vol.16, No.3, pp.333-344.
連國珍著（民81），數位影像處理，儒林圖書。
葉怡成編著（民82），類神經網路模式應用與實作，儒林圖書。
葉怡成編著（民86），應用類神經網路，儒林圖書。
廖紹綱（民88），數位影像處理 活用─MATLAB。
二、 英文文獻
Adrian Biran and Moshe Breiner, (1995), “MATLAB for Engineers,” Addison Wesley.
Ahmed M. Darwish and Anil K. Jain, (1998), “A Rule Based Approach for Visual Pattern Inspection,” IEEE Transactions on Pattern Analysis and Machine Intelligence, VOL. 10, NO. 1, pp. 618-630.
Chris Roman and Charles Reinholtz, (1998), “Robust Course-Boundary Extraction Algorithms for Autonomous Vehicles,” IEEE Intelligent Systems, pp. 32-39, November/December.
G. Pasquariello; G. Satalino; V. la Forgia; and F. Spilotros, (1998)“Automatic Target Recognition for Naval Traffic Control Using Neural Networks,” Image and Vision Computing, VOL. 16, pp. 67-73.
G.J. Edwards; A. Lanitis; C.J. Taylor; and T.F. Cootes, (1998), “Statistical Models of Face Images – Improving Specificity,” Image and Vision Computing, VOL. 16, pp. 203-211.
H.H. Ehricke; G. Daiber; and W. Straber, (1993), “The Vision Camera: An Interactive Tool for Volume Data Exploration and Navigation,” IEEE Visualization’93, pp. 25-30.
H.H. Ehricke; K. Donner; W. Koller; and W. Straber, (1994), “Visualization of Vasculature from Volume Data,” Computers and Graphics, VOL. 18, NO. 3, pp. 395-406.
H.H. Ehricke; T. Grunert; T. Buck; R. Kolb; and M. Skalej, (1994), “Medical Workstations for Applied Imaging and Graphics Research,” Computerized Medical Imaging and Graphics, VOL. 18, NO. 6, pp. 403-411.
Haruo Yoda; Yozo Ohuchi; Yuzo Taniguchi; and Masakazu Ejiri, (1998), “An Automatic Wafer Inspection System Using Pipelined Image Processing Techniques,” IEEE Transactions on Pattern Analysis and Machine Intelligence, VOL. 10, NO. 1, pp. 648-660, January.
Hung, C., C. Lim., and M. C. Liu, (1992), Comparison of Image Algorithms and Neural Networks in Machine Vision Inspection, Computers and Industrial Engineering, No23, PP.105-108.
J. R. Parker,(1994),Practical Computer Vision Using C, John Wiley&Sons, Inc. New York.
Jiang, Bernard C. and Jiang, S. J., (1998) “Machine Vision Based Inspection of Oil Seals,” Journal of Manufacturing System, Vo17, N0.3, pp.159-166.
John M. Gauch, (1999), “Image Segmentation and Analysis via Multiscale Gradient Watershed Hierarchies,” IEEE Transactions on Image Processing, VOL. 8, NO. 1, pp. 69-79, January.
Kjoelen and S.E. Umbaugh; and M. Zuke, (1998), “Compression of Skin Tumor Images,” IEEE Engineering in Medicine and Biology, VOL. 17, NO. 3, pp. 73-80, May/June.
Kjoelen; M. Thompson; S.E. Umbaugh; R. Moss; and W. Stoecker, (1995), “Performance of Artificial Intelligence Methods In Automated Detection of Melanoma,” IEEE Engineering in Medicine and Biology, VOL. 14, NO. 4, pp. 411-416, July/August.
Kok F. Lai and Roland T. Chin, (1998), “On Modeling, Extraction, Detection and Classification of Deformable Contours from Noisy Images,” Image and Vision Computing VOL. 16, pp. 55-62.
Krzyzak; W. Dai; and C.Y. Suen, (1991), “On the Recognition of Handwritten Characters Using Neural Networks,” Pattern Recognition: Architectures, Algorithms and Applications, pp. 115-135.
M.R. Abairid; G. Martin; and C.Y. Suen, “Evaluation of Center-Lines of Digital Patterns Produced by Maximal-Square and Maximal-Cricle Algorithms,” Pattern Recognition: Architectures, Algorithms and Applications, pp. 229-248, 1991.
Milan Sonka; Vaclav Hlavac; and Roger Boyle, Image Processing, Analysis and Machine Vision, Chapman and Hall, 1993.
Muthu Kumaran and Scott E. Umbaugh, “A Dynamic Window-Based Runlength Coding Algorithm Applied to Gray-Level Images,” Graphical Models and Image Processing, VOL. 57, NO. 4, pp. 267-282, July 1995.
Nelson, M. M. and W.T. Illingworth,(1991) A Practical Guide to Neural Nets, Addison-Wesley.
Niemann, H., H. Briinig, R. Salzbrunn, and S. Schrader, (1990) “A Knowledge-Based Vision System for Industrial Applications ” Machine Vision and Applications, 3, 201-229
Osama Moselhi and Tariq Shehab-Eldeen, “Automated Detection of Surface Defects in Water and Sewer Pipes,” Automation in Construction, VOL. 8, pp. 581-588, 1999.
Ostu, N., (1979), “ A Threshold Selection Method from Gray-Level Histogram,” IEEE Transactions on System, Man and Cybernetics, Vol.9, No1, pp.62-66.
P.N. Suganthan and H. Yan, “Recognition of Handprinted Chinese Characters by Constrained Graph Matching,” Image and Vision Computing, VOL. 16, pp. 191-201, 1998.
Piironen, T., O., Silven, M., Pietikainen and T., Laitinen, (1990), Automated Visual Inspection of Rolled Metal Surface, Machine Vision and Applications, Vol.3, pp.247-254.
Piironen, T.; O. Silven; M. Pietikainen; and T. Laitinen, “Automated Visual Inspection of Rolled Metal Surface,” Machine Vision and Applications, Vol.3, pp.247-254, 1990.
Qinghan Xiao and Hazem Raafat, “Combining Statistical and Structural Information for Fingerprint Image Processing, Classification and Identification,” Pattern Recognition: Architectures, Algorithms and Applications, pp. 335-354, 1991.
Raafat, H. and Taboun, S., “An Integrated Robotics and Machine Vision System for Surface Flaw Detection and Classification,” Computer in Industrial Engineers, Vol.30, No.1, pp.27-40, 1996.
Raafat, H. and Taboun, S.,(1996) An Integrated Robotics and Machine Vision System for Surface Flaw Detection and Classification, Computer in Industrial Engineers, Vol.30, No.1, pp.27-40.
Ragael C. Gonzalez and Richard E. Woods,(1993) Digital Image Processing, Adison-Wesley.
Rasanen, J., R. Silvennoinen, K. E., Peiponen and T., Asakura, (1994), “On Surface Damage Detection of Sightly Rough Metal Surfaces”, Optics and Lasers in Nginerring, VOL.20, PP.65-69.
Richard R. Schultz and Robert L. Stevenson, (1994),“A Bayesian Approach to Image Expansion for Improved Definition,” IEEE Transactions on Image Processing, VOL. 3, NO. 3, pp. 233-242.
Robert M. Lougheed and Robert E. Sampson, (1998), “3-D Imaging Systems and High-Speed Processing for Robot Control,” Machine Vision and Applications VOL. 1, pp. 41-57.
Rosenfeld, A..(1988) “Computer Vision: Basic Principles, ” Transactions of IEEE, 76,863-868.
S.E. Umbaugh; R.H. Moss; and W.V. Stoecker, (1993), “Automatic Color Segmentation Algorithms: With Application to Skin Tumor Feature Identification,” IEEE Engineering in Medicine and Biology, VOL. 12, NO. 3, pp. 75-82.
Scott E. Umbaugh, (1998), Computer Vision and Image Processing, Prentice-Hall.
Scott E. Umbaugh; Yansheng Wei; and Mark Zuke, (1997),“Feature Extraction in Image Analysis – A Program for Facilitating Data Reduction in Medical Image Classification,” IEEE Engineering in Medicine and Biology, VOL. 16, NO. 4, pp. 62-73.
Strouthopoulos and N. Papamarkos, “Text Identification for Document Image Analysis Using a Neural Network,” Image and Vision Computing, VOL. 16, pp. 879-896, 1998.
Sun, Y. N. and C. T., Tsai, (1992), A New Model-based Approach for Industrial Visual Inspection, Pattern Recognition, Vol.25, pp1327-1336.
Susan S. Young; Peter D. Scott; and Cesar Bandera, (1998), “Forveal Automatic Target Recognition Using a Multiresolution Neural Network,” IEEE Transactions on Image Processing, VOL. 7, NO. 8, pp. 1122-1135, August.
W. Brent Seales; Cheng J. Yuan; Wei Hu; and Matthew D. Cutts, (1998), “Object Recognition in Compressed Imagery,” Image and Vision Computing, VOL. 16, pp. 337-352.
Woods, W.; Taylor, C. J.; Copper, D.H.; and Dixon, R. N., , (1987), “The Use of Geometric and Gray-Level Models for Industrial Inspection,” Pattern Recognition, Vol.5, pp.11-17.
Wu,W., Wang, M. J., Liu, C., (1996) ,Automated Inspection of Printed Boards through Machine Vision, Computer in Industry, Vol.28, No.2, pp. 103-111, May.
X. Maldague; J.C. Krapez; and D. Poussart, “Thermographic Nondestructive Evaluation (NDE), (1990), : An Algorithm for Automatic Defect Extraction in Infrared Images,” IEEE Transactions on Systems, Man, and Cyberbernetics, VOL. 20, NO. 3, pp. 722-725, May/June.
Ying Liu and Sargur N. Srihari, (1997), Document Image Binarization Based on Texture Features, IEEE Transaction on Pattern Analysis and Machine Intelligence, Vol.19, NO.5, pp.540-544.
Zurada , j. M.,(1992), “ Introduction to Artificial Neural Network ”, West Publishing Company, St. Paul, MN.