研究生: |
劉宗憲 Zong-Xian Liu |
---|---|
論文名稱: |
自動化高亮度發光二極體瑕疵檢測系統之開發與研究 The Development and Research of an Automatic Defects Inspection System for High Brightness Light Emitting Diode |
指導教授: |
郭中豐
Chung-Feng Jeffrey Kuo |
口試委員: |
陳耿明
Keng-Ming Chen 蘇德利 none 高志遠 none 黃昌群 Chang-Chiun Huang |
學位類別: |
碩士 Master |
系所名稱: |
工程學院 - 材料科學與工程系 Department of Materials Science and Engineering |
論文出版年: | 2009 |
畢業學年度: | 97 |
語文別: | 中文 |
論文頁數: | 81 |
中文關鍵詞: | 正規化相關係數法 、K-means 、統計式門檻值 、倒傳遞類神經網路 、支援向量機 |
外文關鍵詞: | Normalized Correlation Coefficient, K-means, Statistical Threshold, Back Propagation Neural Network, Support Vector Machine |
相關次數: | 點閱:594 下載:0 |
分享至: |
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
本研究提出一套發光二極體自動瑕疵檢測系統,利用各種數位影像處理技術於晶粒瑕疵辨識中。檢測項目包含正常晶粒、破晶、電極區刮痕瑕疵、電極區殘金瑕疵、發光區刮痕瑕疵與發光區剝落瑕疵。
本研究主題分為三部分:發光二極體晶粒的定位與擷取、各個結構的特徵值抽取以及發光二極體瑕疵分類。在發光二極體晶粒擷取與定位方面,使用正規化相關係數法,將發光二極體每顆晶粒擷取出來。在取得晶粒各結構部分,利用K-means分群法將晶粒的外觀、電極區與發光區分開檢視。在晶粒外觀檢測方面,使用統計式門檻值二值化後,計算內部黑色區塊的個數,再以型態學區域填充,填補晶粒內不必要的雜點,接著取得外型的特徵值。在電極區瑕疵檢測方面,使用直方圖等化改變電極影像的對比度,利用統計式門檻值二值化後,運用型態學封閉運算修飾,之後計算電極區特徵值。最後在發光區瑕疵檢測方面,利用統計式門檻值三值化將發光區內的基板,發光區和背景的像素分隔開,使用型態學封閉運算修飾後,之後將發光區特徵值抽取出來。最後將所得到的各部分特徵值經過訓練之後利用倒傳遞類神經網路與支援向量機兩種分類器作分類,並將分類器拆成兩層來辨識。以本研究選取468顆晶粒影像進行辨識,在兩種分類器總辨識率都高達95%以上,證明本研究所提出的檢測方法為一種極適合於發光二極體瑕疵之檢測。
This study presents our proposed automatic defect inspection system for LEDs that applies various image processing methods on items including normal chips, chip fragment, scratch mark defect of the pad area, remained gold defect of pad areas, scratch mark defects of luminescent areas, and missing luminescent areas.
The inspection system included three steps: positioning and capture of LEDs, identification of LED components, and defect classification of LEDs. We first used the normalized correlation-coefficient method to capture every LED chip. We then applied K-means clustering method to separate each LED’s external surface, pad, and lighting area. We inspected the chip’s surface by counting black blocks after using statistical threshold binarization. Region filling of morphology was used to fill with unnecessary noises, and the features of externals were obtained. Regarding the inspection of pad areas, we used histogram equalization to change the contrast of each area’s image. After using statistical threshold binarization, we applied closing of morphology to modifying means and obtained features of pad areas. With inspecting lighting areas, we first used statistical multi-threshold to separate substrates, lighting areas, and backgrounds. We obtained features of lighting areas after using closing of morphology. We also adopted the back-propagation neural network and support vector machines to classify five defects.
Experimentally, we collected 486 LED chips for defect inspection. The total recognition rates of both classifiers reach 95%, indicating that we had successfully developed a set of defect inspection system that is applicable to LEDs.
[1] J. P. Lewis, “Fast Normalized Cross Correlation”, Vision Interface, pp. 120-123, 1995.
[2] B. Kai and D. H. Uwe, “Templete Matching using Fast Normalized Cross Correlation”, Proceedings of SPIE-The International Society for Optical Engineering, vol. 4387, pp. 95-102, 2001.
[3] D. M. Tsai and C. T. Lin, “Fast Normalized Cross Correlation for Defect Detection”, Pattern Recognition Letters, vol. 24, pp. 2625-2631, 2003.
[4] C. J. Huang, C. F. Wu and C. C. Wang, “Image Processing Techniques for Wafer Defect Cluster Identification”, IEEE Design & Test of Computers, vol. 19, no. 2, pp. 44-48, 2002.
[5] P. S. Windyga, “Fast Impulsive Noise Removal”, IEEE Transactions On Image Processing, vol. 10, No. 1, pp. 173-179, 2001.
[6] J. N. Kapur, P. K. Sahoo and A. K. C. Wong, “A New Method for Gray-level Picture Thresholding Using the Entropy of the Histogram”, Computer Vision, Graphics and Image Processing, vol. 29, no. 3, pp. 273-285, 1985.
[7] G. Ward, “Fast Robust Image Registration for Compositing High Dynamic Range Photographs from Handheld Exposure”, Journal of Graphics Tools, vol. 8, no. 2, pp. 17-30, 2003.
[8] C. Y. Chang, C. H. Chang, C. H. Li and M. D. Jeng, “Learning Vector Quantization Neural Networks for LED Wafer Defect Inspection”, IEEE Digital Object Identifier, pp. 230-230, 2007.
[9] J. Y. Wu, “Apply Image Processing Technology to the Appearance Defect Inspection of High Brightness Light Emitting Diode”, National Taiwan University of Science and Technology, Master thesis, 2007.
[10] Y. Y. Liao, “Apply Machine Vision Technology to Defect Inspection of High Brightness Light Emitting Diode”, National Taiwan University of Science and Technology, Master thesis, 2008.
[11] H. C. Kim, S. Pang, H. M. Je, D. Kim, and S. Y. Bang, “Pattern Classification using Support Vector Machine Ensemble”, Proceedings 16th IEEE International Conference on Pattern Recognition, vol 2, pp. 160-163, 2002.
[12] 余明興、吳明哲、黃世陽編著,“Borland C++ Builder 6 程式設計經典”,文魁資訊,2002.
[13] 吳上立、林宏敦編譯,“C語言數位影像處理”,全華科技圖書股份有限公司,2006.
[14] 陳隆建編著,“發光二極體原理與製造”,全華科技圖書股份有限公司,2006.
[15] C. Gonzalez and E. Woods, “Digital Image Processing”, Second Edition, Pearson Education Taiwan Ltd, 2003.
[16] 鍾國亮編著,“數位處理與電腦視覺第三版”,東華書局,2006.
[17] S. J. Ko and Y. H. Lee, “Center Weighted Median Filters and their Applications to Image Enhancement”, IEEE Transactions On Circuits and Systems for Video Technology, vol. 39, no. 9, pp. 984-993, 1991.
[18] A. K. Jain and R. C. Dubes, “Algorithms for Clustering Data”, Prentice Hall, New Jersey, 1988.
[19] N. Otsu, “A Threshold Selection Method from Gray-level Histograms”, IEEE Transactions on System, Man and Cybernetics, vol. SMC-9, no. 1, pp. 62-66, 1979.
[20] V. N. Vapnik, “Statistical Learning Theory”, Hardcover, John Wiley & Sons, Inc., New York, 1998.
[21] D. Gale, H.W. Kuhn, and A. W. Tucker, “Linear Programming and the Theory of Games”, Activity Analysis of Production and Allocation, pp. 317-329, 1951.
[22] C. C. Chang and C. J. Lin, “LIBSVM:A Library For Support Vector Machines ”, http://www.csie.ntu.edu.tw/~cjlin/libsvm, 2001.