本研究旨在應用影像處理技術及類神經網路理論於Array電路工程瑕疵檢測，首先利用二維的傅立業轉換及反傅立業轉換再使用二值化影像分割技術將瑕疵凸顯出來。並經由點標示(Pixel Labeling)、瑕疵缺口偵測(Defective Breach Detection)及膨脹(Dilation)等處理還原瑕疵點的原貌。再由瑕疵圖形的特性選取面積、周長、圓滑度、包含物體的最小長方形、瑕疵點所佔四方向區域面積及細線化後瑕疵點所佔四方向區域面積等六種圖形特徵，做為該瑕疵點的特徵值。
最後以特徵值當做類神經網路輸入訓練及測試參考，經由自組織映射(Self-Organizing Map Neural Network, SOM)及蟻群為基礎之自組織映射圖網路變形之系統(Ant-Based Growing SOM)、倒傳遞(BPN)及監督式自組織映射(supervised SOM)。針對Array電路工程最常見的三種金屬殘留瑕疵：圓形殘、中空圓形殘、不規則殘以及刮痕共四種瑕疵類型，並加入無瑕疵樣本進行辨識。從實驗結果可明顯地看出此四類系統可以正確的辨識出五種Array電路工程瑕疵種類，辨識率皆為100% 。我們可以將此研究實際應用於Array電路工程瑕疵檢測的線上辨識系統，以降低人為檢測的誤差。

The purpose of this research is to apply image processing technology and neural network to the defective detection of array panel. Firstly, two-dimensional Fourier transform and inverse Fourier transform are utilized and then let defects be even conspicuous by the use of binary image segmentation technology. By the help of pixel labeling, defective breach detection and dilation and some other processing techniques to make the defective spots returned to the original condition. According to the image characteristics, we identify six graphic features: area, perimeter, roundness, length and width of the least rectangle covering whole spot, four-directional areas of defective spot, and four-directional areas of defective spot after thinning, that serve as feature values of that defective spot. All the six image characteristics are served as the characteristic values for that defected spot detection.
In conclusion, the feature values can be employed as input training and testing data for neural network paradigms. The recognition process can then conducted by the application of Self-Organizing Map (SOM) neural network, ant-based SOM, BPN and supervised SOM for three common metal residue defects, such as round residue, concentric residue, irregular residue, and scrape. Samples without defects are also added to do recognition. Experimental findings clearly show that these four neural network paradigms are capable of recognizing five kinds of defects in the array panel. The recognition correctness rate can achieve 100%. We may apply this research result to the online recognition system for defective detections of the array panel to reduce human detection errors.

誌 謝....................I
摘 要....................II
Abstract..................III
目 錄.....................IV
表目錄.....................VII
圖目錄.....................VIII
第一章 諸論.................1
1.1研究背景及動機............1
1.2研究目的.................2
1.3研究範圍與限制............4
1.4研究流程..................4
第二章 文獻探討...............7
2.1 TFT-LCD製程簡介..........7
2.2圖形識別系統..............14
2.3圖形辨識―特徵選取..........18
2.4瑕疵檢測..................22
2.4.1瑕疵類型................22
2.4.2轉換法..................25
2.4.3 TFT-LCD的瑕疵檢測法.....28
2.5類神經網路理論.............31
2.5.1類神經網路...............31
2.5.2類神經網路的運作過程......32
2.5.3類神經網路的架構..........34
2.5.4隱藏層處理單元數目之選定...35
第三章 研究方法................38
3.1二維傅立葉轉換..............39
3.2影像分割：二值化.............46
3.3瑕疵點還原..................48
3.3.1點標示....................48
3.3.2瑕疵缺口偵測...............49
3.3.3膨脹......................51
3.4特徵值選取...................52
3.4.1面積......................52
3.4.2周長......................52
3.4.3包含物體的最小長方形........53
3.4.4圓滑度或稠密度.............54
3.4.5 瑕疵點所佔四方向面積.......54
3.4.6 細線化後瑕疵點所佔四方向面積...54
3.5分類器(Classifier)：類神經網路...57
3.5.1非監督式類神經網路.............57
3.5.2監督式類神經網路...............59
3.5.2.1倒傳遞類神經網路訓練及測試.....61
3.5.2.2 Supervised SOM類神經網路訓練及測試...62
第四章 研究結果與分析.................63
4.1瑕疵影像處理各參數變動對於瑕疵偵測之影響.....63
4.1.1選擇濾除寬度範圍w................64
4.1.2保留半徑r 範圍...................65
4.1.3二值化門檻值範圍n.................67
4.2瑕疵點還原及特徵選取影像處理各階段結果..68
4.2.1瑕疵點還原影像處理結果.............68
4.2.2特徵值選取影像處理結果.............70
4.3類神經網路辨識結果...................73
4.3.1 Kohonen SOM及ABSOM類神經網路聚類效能比較.....73
4.3.2倒傳遞類神經網路辨識結果............75
4.3.3 sSOM類神經網路辨識結果............76
4.3.4 BPN及sSOM最佳模式辨識結果比較......79
4.3.5瑕疵點還原前BPN及sSOM辨識結果.......80
第五章 結論及建議........................82
參考文獻................................84
一、英文文獻.............................84
二、中文文獻.............................88
附錄一...................................92
附錄二..................................100
附錄三..................................108
附錄四..................................111
附錄五..................................114

一、英文文獻
[1]Bryant, A., and Bryant, J., “Recognizing shapes in planar binary images ,” Pattern Recognition, Vol. 22, No. 2, pp.155-164, 1989.
[2]Chan, Y.C., K.C., Huang and X. Dai, “Nondestructive defect detection in multiplayer ceramic capacitors using an improved digital speckle correlation method with wavelet packet noise reduction processing,” IEEE Transactions on Packaging and Manufacturing Technology, Vol. 46, pp.80-87, 2000.
[3]Chen, P.W., T.C. Liang and H.F. Yau, “Classifying textile faults with a back-propagation neural network using power spectra,” Textile Research Journal, Vol. 68, No. 2, pp. 121-126, 1998.
[4]Chan, C.H., and G.K.H. Pang, “Fabric defect detection by fourier analysis,” IEEE Transactions on Industry Application, Vol. 36, No. 5, pp. 1267-1276, 2000.
[5]Chang, C.C., Hwang, S.M., and Buehrer, D.J., “A shape recognition scheme based on relative distances of feature points from the centroid,” Pattern Recognition, Vol. 24, No. 11, pp. 1053-1063, 1991.
[6]Dubois, S.R., and Glanz, F.H., “An atoregressive model approach to tow-dimensional shape classification”, IEEE Transactions on Pattern Annalysis and Machine Intelligence, Vol. 8, No. 1, pp. 67-75, 1986.
[7]Erturk, S. and T.J. Dennis, “Image sequence stabilization based on DFT filtering”, IEE Proc.-Vis. Image Signal Process, Vol. 147, pp. 95-102, 2000.
[8]Fu,K.S., Sequential methods in pattern recognition and machine learning, Academic Press, New York,1968.
[9]Fu, K.S., Syntactic pattern recognition and applications, Prentice-Hall, Englewood Cliffs, New Jersey, 1982.
[10]Fukunaga,K., Introduction to statistical pattern recognition, Academic Press, San Diego, CA, 1972.
[11]Fujiwara, H., Z. Zhong and K. Hashimoto, “Toward automated inspection of textile surfaces removing the textural information by using wavelet shrinkage,” IEEE International Conference on Seoul Korea, pp. 3259-3534, 2001.
[12]Gupta, L., M.D. Srinath, “Invariant planar shape recognition using dynamic alignment,” Internation Pattern Recognition, Vol. 21, No.3, pp. 235-239, l988.
[13]Gonzalez R.C. and R.E.Woods, “Digital image processing,2nd edition, Addison-Wesley, Boston,MA, 2002.
[14]Hu M.C. and I.S. Tasi, “Fabric inspection based on best wavelet packet bases,” Textile Research Journal, Vol. 70, No. 8, pp. 662-670, 2000.
[15]Hu, M. K., “Visual pattern recognition by moment invariants,” IRE Transactions on Information Theory, Vol. 8, pp. 179-187, 1962.
[16]Horaud, R., S. Olympieff, and J. P. Charras, “Shape and position recognition of mechanical parts from their outlines,” Proceedings of the 1st International Conference on Robot Vision and Sensory Controls, pp. 125-134, 1997.
[17]Huynh L.V., “A vision system for in-process surface quality assessment,” Society of Manufacturing Engineers (SME), Vol. 87, pp. 12-51, 1987.
[18]Oh, I.S., and C.Y.Suen, “Distance features for neural network-based recognition of handwritten characters,” International Journal on Document Analysis and Recognition ( IJDAR), Vol. 1, No. 1, pp. 73-88, 1998.
[19]Jagannathon, S., “Automatic inspection of wave solder joints using neural network,” Journal of Manufacturing Systems, Vol. 16, No. 6, pp. 389-398, 1997.
[20]Jin, L.W. and J.Z. Qin, “Car plate number characters recognition using Gabor orientantion features and neural networks,” Proceedings of the 2003 International Conference on Neural Networks and Signal Processing, Vol. 2, pp. 1628-1631, Dec. 2003.
[21]Jain, A.K., “Fundamentals of digital image processing,” Prentice Hall Information and System Sciences Series, pp. 409-411, 1989.
[22]Kumar, A. and G. Pang, “Identification of surface defects in textured materials using wavelet packets,” Proceedings of the 2001 IEEE Industrial Applications Conference, Vol. 1, pp. 247-251, 2001.
[23]Kertesz, A., V. Kertesz, and T. Muller, “An on-line image processing system for registration number identification, ”IEEE International Conference on Neural Networks, Vol. 6, pp. 4145-4148, 1994.
[24]Kido, T., “In-process inspection technique for active-matrix LCD panels, ” International Test Conference, pp. 795-799, 1992.
[25]Kido, T., “In-process functional inspection technique for TFT-LCD array, ” Journal of SID, Vol. 1, No. 4, pp. 429-435, 1993.
[26]Kido, T., N. Kishi, and H. Takahashi, “Optical charge-sensing method for testing and characterizing thin-film transistor arrays,” IEEE Journal of Selected Topics in Quantum Electronics, Vol. 1, No. 4, pp. 993-1001, 1995.
[27]Kittler, J. and J. Illingworth, “Minimum error thresholding, ”Pattern Recognition, Vol. 19, No. 1, pp. 41-47, 1986.
[28]Kohonen, T., “The self-organizing map,” Proceedings of the IEEE, Vol. 78, pp. 1464-1480, 1990.
[29]Kim, S., M.H. Lee and K.B. Woo, “Wavelet analysis to fabric defects detection in weaving processes,” Proceedings of the IEEE International Symposium on Industrial Electronics, Vol. 3, pp. 1406-1409, 1999.
[30]Latif-Amet, A.L., A. Ertuzun and A. Ercil, “Texture defect detection using subband domain co-occurrence matrices,” Image Analysis and Interpretation, pp. 205-210, 1998.
[31]Looney, G.G., Pattern recognition using neural network, Oxford University Press, Inc., 1997.
[32]Liu, S.S. and M.E. Jernigan, “Texture analysis and discrimination in additive noise,” Computer Vision, Graphics, and Image Processing, Vol. 49, pp. 52-67, 1990.
[33]Lin, C.S., W.Z. Wu, Y.L. Lay and M.W. Chang, “A digital image-base measurement system for a LCD backlight module,” Optics & Laser Technology, Vol. 33, No. 7, pp. 499-505, 2001.
[34]Lambert, G. and F. Bock, “Wavelet method for texture defect detection,” Proceedings of the IEEE International Conference on Image Processing, Vol. 3, pp. 201-204, 1997.
[35]NEC (Nippon Electric Company) , “Improvements in or relating to character recognition apparatus,”U.K. Patent, Vol. 1, pp. 124-130, Aug. 1968.
[36]Nakashima, K., “Hybrid inspection system for LCD color filter panels, ” Proceedings of the Tenth International Conference on Instrumentation and Measurement Technology, Hamamatsu, Vol. 2, pp.689-692, 1994.
[37]Neubauer, C. , “Intelligent X-Ray inspection for quality control of solder joints, ”IEEE Transformation on Component, Packaging, and Manufacturing Technology, Vol.20, No.2, pp.111-120, 1997.
[38]Pu, H., H. Wei, D.F. Wang and Y.J. Zhai, “Car license plate feature extraction and recognition base on multi-stage classifier,” Proceedings of the International Conference on Machine Learning and Cybernetics, Vol. 1, pp. 128-132, Nov. 2003.
[39]Pandya, A. S., R.B. Macy, Pattern recognition with neural network in C++, CRC Press & IEEE Press, 1996.
[40]Parisi, R., E. D. Di Claudio, G. Lucarelli and G. Orlandi, “Car plate recognition by neural networks and image processing,” Proceedings of the 1998 IEEE International Symposium on Circuits and Systems, Vol. 3, pp. 195-198, 1998.
[41]Richard, O.D., E.H. Peter and G.S. David, Pattern classification, 2nd ed., John Wiley and Sons, New York, NY, USA, 2001.
[42]Raus, M. and L. Kreft, “Reading car license plates by the use of artificial neural networks,” Processings of the 38th Midwest Symposium on Circuits and Systems, Vol. 1, pp. 538-541, 1995.
[43]Ryu, Y.K. and H.S. Cho, “Visual inspection scheme for using in optical solder joint inspection system,” Proceedings of the IEEE International Conference on Robotics and Automation, Minneapolis, MN , Vol. 4, pp. 3259-3264, 1996.
[44]Tindall, D.W., “Application of neural network techniques to automatic license plate recognition,” European Convention on Security and Detection, Brighton, pp. 81-85, 1995.
[45]Tsai, D.M., and C.Y. Hsieh, “Automated surface inspection for directional textures, ” Image and Vision Computing, Vol. 18, pp. 49-62, 1999.
[46]Tsai, I.S., C.H. Lin and J.J. Lin, “Applying an artificial neural network to pattern recognition in fabbic detects,” Texible Research Journal, Vol. 65, No. 3, pp. 123-130, 1995.
[47]Wang, X.W., X.Q.Ding and C.S.Liu, “Optimized Gabor filter based feature extraction for character recognition, ” Proceedings of the International Conference on Pattern Recognition, Vol. 4, pp. 223-226, Aug. 2002.
[48]Yang, C.C., and S.C.Chi, “An ant-base self-organization feature maps algorithm, ”Proc.The 5th Workshop on Self-Organization Maps, Paris, France, pp.65-73, 2005.
[49]Zahn, C. T., and R. Z. Roskies, “Fourier descriptor for plane closed curves,” IEEE Transactions on Computer, Vol. 21, No. 3, pp. 269-281, 1972.
[50]Zhang, T.Y., and C. Y. Suen, “A fast parallel algorithm for thinning digital pattern,” Communication of the ACM, Vol. 27, No. 3, pp. 236-239, 1984.

二、中文文獻
[51]王進德、蕭大全(1994)，類神經網路與模糊控制理論入門，全華科技圖書公司。
[52]田榮雯(2001)，以FPGA實現倒傳遞類神經網路並應用於肌電圖分類，私立中原大學資醫學工程研究所，碩士論文。
[53]林東賦(2001)，應用影像處理技術與類神經網路理論於非織物瑕疵辨識，國立台灣科技大學纖維及高分子工程研究所，碩士論文。
[54]紀國鐘、鄭晃忠(2004)，液晶顯示器技術手冊，經濟部技術處台灣電子材料與元件協會，新竹。
[55]洪崇祐(2004)，應用一維傅立葉分析於TFT-LCD液晶顯示面板之瑕疵檢測，私立元智大學工業工程與管理研究所，碩士論文。
[56]侯東旭、陳健諭(1994)，「使用類神經網路於工件辨識之研究」，亞太工業工程研討會第八十三年論文集，第432-437頁。
[57]高祥益(2004)，液晶顯示器透明電極線缺陷型態辨識之研究，國立高雄應用科技大學機械與精密工程研究所，碩士論文。
[58]連國珍(2004)，數位影像處理，儒林，台北。
[59]陳一斌(2001)，「TFT彩色濾光片瑕疵檢測系統」，機械工業雜誌12月號，第225期，第204-209頁。
[60]郭正德(2003)，應用小波轉換作紋理影像之瑕疵檢測及合成，國立中央大學資訊工程研究所，碩士論文。
[61]陳進興(1993)，「圖形辨識方法對自動目標分類之分析比較」，國科會學術合作計畫報告，台南。
[62]陳緯達(2004)，類神經網路在手寫數字辨識之研究，國立中央大學資訊工程研究所，碩士論文。
[63]陳志忠(2001)，液晶顯示器的像素點缺陷與相對亮度均一性之自動化檢測，私立中原大學機械工程學研究所，碩士論文。
[64]陳世璋(2005)，使用類神經網路之自動化臉部表情辨識系統，國立成功大學電腦與通信工程研究所，碩士論文。
[65]陳鵬帆(2004)，以自適應共振理論網路為基礎建構彩色濾光片微觀瑕疵辨識系統之研究，國立成功大學工業與資訊管理學系，碩士論文。
[66]陳飛龍、謝詳文(2002)，「特徵為基礎之晶圓缺陷圖樣辨識與分類演算法」，中國工業工程學會期刊，第四期，第九卷，第17-19頁。
[67]許文豪(2000)，圖形識別概述與實作，國立清華大學資訊工程系，碩士論文。
[68]黃國源(2003)，類神經網路與圖形識別，維科，台北。
[69]黃哲韻(2001) ，應用機器視覺於隨機性紋路表面之瑕疵檢測，私立元智大學工業工程與管理研究所，碩士論文。
[70]曾彥馨(2003)，應用機器視覺於TFT面板之表面瑕疵檢測與分類，私立元智大學工業工程與管理研究所，碩士論文。
[71]溫福助(2000)，類神經網路樣板比對法於車牌字元辨識之研究，國立台灣大學電機工程學研究所，碩士論文。
[72]葉怡成(2000)，類神經網路模式應用與實作，儒林，台北。
[73]蔡英男(2003)，應用影像處理與類神經網路於偏光膜瑕疵分析，國立台灣科技大學纖維與高分子工程研究所，碩士論文。
[74]繆紹綱(2005)，數位影像處理―活用Matla，全華，台北。
[75]錢志豪(2002)，建構液晶顯示器(LCD)色彩偏差瑕疵之自動化視覺檢測系統之探討，私立朝陽科技大學工業工程研究所，碩士論文。
[76]顧鴻壽(2004)，光電液晶平面顯示器技術基礎及應用，新文京，台北。