有機發光二極體顯示面板(OLED)為現階段具有自發光性質等多項優點之最新平面顯示元件(Flat Panel Device)，各大廠針對各自不同的製造過程所製作出的OLED做檢測，且大多以人工目視的方式檢測。一片OLED多達上千點的發光點，易視覺疲勞而造成誤判，準確性大大降低。本論文針對在發光(Light On)階段有機發光層表面發生單一及同時瑕疵發生之狀況，開發一套自動影像檢測與模糊決策樹分類器系統，瑕疵種類含正常以外，有暗點、斷線及亮度不均，也考量了多種瑕疵同時存在，故增加了另外四種瑕疵情況-暗點與斷線、亮度不均與暗點、亮度不均與斷線及亮度不均與暗點與斷線。影像處理過程中，將影像經過色彩重建，利用九宮格式比較平均亮度值來判定亮度量測；接著針對目標影像輪廓特性，利用幾何學上的尤拉數連通物與洞判斷，標計出大洞數目與小洞數目來判別暗點發生性及線段完整性。瑕疵樣本方面，蒐集210筆瑕疵樣本，由實驗得知，前兩層經由規則式的決策樹分類，辨識率達100%。在暗點發生狀況中利用模糊決策樹，能更準確且合乎人性判斷的分類，辨識率也達100%，此時驗證了瑕疵特徵選取的適用性及分類器合乎人性化的準確性，也明顯的看出此檢測系統成功的被應用於OLED有機發光層的檢測工作上。

Nowadays, the latest flat panel device with autoluminescence and a number of advantages is organic light-emitting diode display panel (OLED). Each individual factory focuses on testing and inspecting OLED produced by different manufacturing processes. There are over thousands of light points on a plate of OLED, causing the mistakes and poor rates of correctness due to worker’s visual fatigue. According to this thesis, in light On phase, the organic light-emitting layer brings about single and multiple defects. We develop an automatic image detection and fuzzy decision tree (FDT) System for three categories of defects-normal, dark dots, broken, and uneven brightness. We also consider the situation that multiple defects exist at the same time and provide another four cases of defects. In image processing, after the reconstruction of colors of images, we can compare the different average values of brightness to decide the brightness through nine palaces of dose algorithm. For the features of contours of target images, by means of the Euler number of connected objects involved in geometry and the determination of holes, we label the number of large hole and small hole respectively to verify the possibility of dark spots as well as line integrity. Otherwise, the experiment results from 210 defect samples, show that recognition rate is 100% by regular classification for the first two phase. While adopting FDT in the state of emergence of dark spots, we find that the method is more accurate and appropriate for human, even the recognition rate is 100%. As a result, we believe the applicability of defect features and the classifier which both are quite accurate, and it is also obvious that this detection system is used in the certain testing way of OLED organic light-emitting layer successfully.

[1] M. Eritt, C. May, K. Leo, M. Toerker and C. Radehaus,“OLED Manufacturing for Large Area Lighting Applications,”Thin Solid Films, Vol. 518, pp. 3042-3045, 2010.
[2] L. YouJong, K. JooHyung and W. G. Jang,“Development of Inverted OLED with Top ITO Anode by Plasma Damage-Free Sputtering,”Thin Solid Films, Vol. 517, pp. 4019-4022, 2009.
[3] L. Eunghi,“Simulation of the Thin-Film Thickness Distribution for an OLED Thermal Evaporation Process,”Vacuum, Vol. 83, pp. 848-852, 2009.
[4] 姚文珍，應用於TFT-LCD隔板之低複雜度缺陷檢測架構，大同大學通訊工程研究所論文，2007。
[5] 陳志忠，液晶顯示器的像素缺陷與相對亮度均一性之自動化檢測，中原大學機械工程研究所論文，2001。
[6] 李孟坤，高分子有機發光二極體瑕疵自動光學檢測系統，國立交通大學工業工程與管理學研究所論文，2005。
[7] 陳志強，影像處理技術應用於PLED噴墨基板瑕疵檢測，國立中央大學資訊工程研究所論文，2003。
[8] R. Kinnel,“Demosaicing: image reconstruction from color CCD samples,”IEEE Trans. on Image Processing, Vol. 8, pp. 1221-1228, 2002.
[9] N. Otsu, “A Threshold Selection Method from Gray Level Histogram,” IEEE Trans. on Systems, Man, and Cybernetics, Vol. SMC-8, pp. 62-66, 1978.
[10] S. J. Ko and Y. H. Lee,“Center Weighted Median Filters and Their Applications to Image Enhancement,”IEEE Trans. on Circuits and Systems, Vol. 38, no. 9, pp. 984-993, 1991.
[11] J. Yang,“Topology Description for Data Distributions Using a Topology Graph With Divide-and-Combine Learning Strategy,”IEEE Trans. on Systems, Man, and Cybernetics, Part B: Cybernetics, Vol. 36, pp. 1296-1305, 2006.
[12] B. C. Jiang, S. L. Tasi and C. C. Wang“Machine Vision-Based Gray Relational Theory Applied to IC Marking Inspection,”IEEE Trans. on Semiconductor Manufacturing, Vol. 15, pp. 531-539, 2002.
[13] G. Fang and W. Gao,“Large Vocabulary Sign Language Recognition Based on Fuzzy Decision Trees”IEEE Trans. on Systems, Man and Cybernetics, Part A: Systems and Humans, Vol. 34, pp. 305-314, 2004.
[14] X. Wang,“A Comparative Study on Heuristic Algorithms for Generating Fuzzy Decision Trees”IEEE Trans. on Systems, Man, and Cybernetics, Part B: Cybernetics, Vol. 31, pp. 215-226, 2002.
[15] D. S. Yeung,“An Initial Comparison of Generalization-Capability between Crisp and Fuzzy Decision Trees,”Machine Learning and Cybernetics, Vol. 4, pp. 1846-1851, 2002.
[16] C. Z. Janikow, “Fuzzy Decision Trees: Issues and Methods,” IEEE Trans. on Systems, Man, and Cybernetics-Part B: Cybernetics, Vol. 28, pp.1-14, 1998.
[17] 蔡孟凱編著，C++ Builder6完全攻略，上奇科技股份有限公司，2003。
[18] 余明興編著，Borland C++ Builder程式設計精典，文魁資訊股份有限公司，2002。
[19] 簡金雄，陳金鑫，二十一世紀的明星產業-有機發光二極體平面顯示器技術，光電科技，2001。
[20] P. Pope, H. P. Kallmann and P. J. Magnante,“Electroluminescence in Organic,”Journal of Chemical Physics, Vol.38, pp. 2042-2043, 1963.
[21] C. W. Tang and S. A. Van Slyke, “Organic Electroluminescent Diodes,” Applied Physics Letters, Vol. 51, pp. 913-915, 1987.
[22] 陳志強編著，OLED有機發光二極體顯示器技術，全華圖書股份有限公司，2006。
[23] 顧鴻壽，有機電激發光顯示器技術與應用，新文京出版公司，2006。
[24] M. A. Baldo, D. F. O’Brien, Y. You, A. Shoustikov, S. Sibley and M. E. Thompson, S. R. For. Forrest, Nature(London), pp. 395, 151, 1998.
[25] 陳光榮、林彥仲，光訊第98期，p.23，2002。
[26] 城戶淳二著，圖解有機EL，世茂出版社，2004。
[27] 鐘國亮編著，影像處理與電腦視覺第三版，東華書局，2005。
[28] N. Bouaynaya,“Theoretical Foundations of Spatially-Variant Mathematical Morphology Part II: Gray-Level Images,”IEEE Trans. on Pattern Analysis and Maxhine Intelligence, Vol. 30, pp. 831-850, 2008.
[29] Y. Du, C. I. Chang and P. D. Thouin,“Unsupervised Approach to Color Video Thresholding,”Optical Engineering, Vol. 43, NO. 2, pp. 282-289,2004.
[30] R. T. Chin,“Analysis of Thinning Algorithms Using Mathematical Morphology,”IEEE Trans. on Pattern Analysis Machine Intelligence, Vol. 12, pp. 541-551, 1990.
[31] Z. Zhang, H. Moss and W. V. Stoecker,“Border Detection on Digitized Skin Tumor Images,”IEEE Trans. on Medical Imaging, Vol. 19, pp. 1128-1143, 2000.
[32] C. Z. Janikow, “Fuzzy Decision Tree: Issues and Methods,” IEEE Trans. on Systems, Man, and Cybernetics-Part B: Cybernetics, Vol. 28, NO.1, pp.1-14, 1998.
[33] J. Zeidler and M. Schlosser, “Continous-valued attributes in fuzzy decision trees,” Proceedings of Information Processing and Management of Uncertainty in Knowledge-Based Systems, pp. 395-140, 1996.
[34] L. A. Zadeh, “Fuzzy Sets,” Information and Control, Vol. 8, pp. 338-353, 1965.
[35] 王文俊，認識Fuzzy，第二版，全華科技，2001。
[36] 蕭芳智，應用階層式粒子群方法於模糊決策樹之研究，元智大學工業工程與管理研究所，2005。
[37] J. Buckley, "Fuzzy Complex Analysis II: Integration", Fuzzy Sets and Systems, vol. 49, pp. 171-179, 1992 .
[38] 段麗莉，以多目標基因演算法設計高效能模糊分類器，逢甲大學資訊工學系研究所論文，2003。
[39] 陳麗君，應用資料探勘技術於信用卡黃金級客戶之顧客關係管理，元智大學工業工程與管理研究所碩士論文，2003。
[40] Y. Peng, P. Flach ,“Soft Discretization to Enhance the Continuous Decision Tree Induction,”Integrating Aspects of Data Mining, Decision Support and Meta-Learning, pp. 109–118, 2001.
[41] J. R. Quinaln,“Induction of Decision Trees,”Machine Learning, Vol. 1, pp. 81-106, 1986.
[42] Y. Yuan and M. J. Shaw, “Induction of Fuzzy Decision Trees,” Fuzzy Sets and System, Vol. 69, pp. 125-139, 1995.