TFT-LCD模組製程的面板點燈檢測除基礎的光電響應量測之外、還包括壽命(aging)及畫質缺陷檢測，如點、線、異物、Mura…等缺陷，然而在諸多的畫質檢測中，尤其以Mura缺陷之瑕疵最不易檢測出來，而易招致客戶抱怨，不為消費者所接受。目前產業界的Mura缺陷檢測判定，皆於成品階段以人工目視之方式進行全面檢測，人工檢測除了費時之外也容易造成漏檢、誤判等問題，因此建立自動化產品的檢測機制是絕對必要的。
　　本研究深入探討TFT-LCD顯示原理，分析各種Mura缺陷的形成原因，依據缺陷之特徵進行分類，並蒐集相關文獻及學理，建立一套檢測程序。檢測方式分別以面掃描CCD攝影機模擬人工視覺檢測的情形，在約110公分的距離下，使用高解析度的鏡頭來檢測整塊面板的品質。並以自行撰寫的Mura缺陷檢測程式及建構檢測硬體之架構，完成Mura缺陷的自動化檢測系統。Mura檢測方法主要利用數位影像處理之技術，分別以線性迴歸診斷及預測偵測影像中異常值(Outliers)及影響點(Influential point)，以Niblack's閥值切割進行Mura缺陷區域的分割，再進一步量化評價備選Mura，並以0.5的辨識臨界值判斷出每一個真實的Mura缺陷。
　　本文實驗使用了13片17吋TFT-LCD面板產品，其中有10片panel經由人眼目視方式檢測出Mura缺陷(bad panel)，另外有3片panel經由人工檢測判定無任何缺陷之正常品(good panel)。經自動化檢測結果顯示本研究使用的方法皆可有效正確地分類出良品(G品)與不良品(NG品)的TFT-LCD面板產品。

　　The Liquid Crystal Display(LCD) nowadays will be the most important and promising technical product. The techniques of inspection of the LCD panel include aging, final display inspection , and final visual inspection. Line, poin , particle, and Mura etc. are those of the display defects of the LCD panel. However, the Mura defect on TFT-LCD is most difficult to inspect among all the known defects and thus it is prone to be complained by customer. Currently, most of TFT-LCD Mura defect inspection tasks are done by artificial inspecttion, which is time-consumption and low accuracy. Therefore it is essential to build up an automatic inspection scheme of the manufacture process and with the image processing techniques to attain a high accurate detection rate.

　　The presented method of the TFT-LCD Mura defects inspection is mainly with digital image process, which consists of three phases. The first phase, we employ regression diagnostics method to detect outliers and influential points, and then to estimate background image region. The second phase, we segment candidate region Mura from TFT-LCD display image using Niblack's threshold. In the third phase, we quantify Mura level for each candidate, which is used to identify real muras while the mura level threshold was set to be 0.5.

　　The experiment has been performed on 13 TFT-LCD panel samples consisting of 10 bad panels and 3 good panels. Each bad panel has at least one Mura defect. All Mura defects in our experiment were detected by human visual inspection in the field beforehand. Good panels are claimed to have no Mura defect. Finally, in the manner of the presented visual inspection technique of Mura defect in TFT-LCD, all Mura defects claimed by human inspection have been correctly detected and classify the good panels(pass) and the bad panels(fail) successfully.

[1] K. Nakashima, “Hybrid Inspaction System for LCD Color Filter Panels,” Instrumentation and Measurement Technology Conference, 1994.
[2] F. Saitoh, “Boundary Extraction of Brightness Unevenness on LCD Display
Using Genetic Algorithm Based on Perceptive Grouping Factors,” Proceedings of the International Conference on Image Processing, Vol. 2, pp 308-312, Kobe, Japan, 1999.
[3] S. Hecht, “The visual discrimination of intensity and the Weber-Fechner law,” J. Gen. Physiol., Vol. 7, pp. 241, 1924.
[4] Jae Y. Lee and Suk I. Yoo, “Automatic Detection of Region-Mura Defect in TFT-LCD”, IEICE TRANS, 2004.
[5] B. C. Jiang, C. -C. Wang and H.-C. Liu, “Liquid crystal display surface uniformity defect inspection using analysis of variance and exponentially weighted moving average techniques,” International Journal of Production Research,” Vol. 43, pp67–80, 2005.
[6] C. J. Lu and D. M. Tsai, “Automatic defect inspection for LCDs using singular value decomposition,” International Journal of Advanced Manufacturing Technology, Vol.25, pp53-61, 2004.
[7] P. J. Rousseeuw and A. M. Leroy,“Robust Regression and Outlier Detection.,” John Wiley & Sons, New York, 1987.
[8] R. D. Cook, “Detection of influential observation in linear regression.,” Technometrics, 19, pp15-18 , 1997.
[9] R. D. Cook and S.Weisberg, “Applied regression including computing and graphics,” Wiley, New York, 1999.
[10] D. A. Belsley, E. Kuh, and R. E. Welsch, “Regression Diagnostics,” John Wiley & Sons, USA, 1980.
[11] W. Niblack, “An Introduction to Image Processing,” Prentice Hall, 1986, pp115-116.
[12] Definition of measurement index (semu) for luminance mura in FPD image quality inspection, SEMI standard:SEMI D31-1102, http://www.semi.org
[13] VESA Flat Panel Display Measurements Standard Ver. 2.0, June 1, 2001.
[14] P. Moon and D. E Spencer, “The visual effect of nonuniform surrounds, " J. Opt. Soc., Vol. 35, pp233-248, 1945.
[15] D .M Tsai, C.-Y.Hsieh, “Automated surface inspection for directional textures,” Image and Vision Computing ,Vol.18, pp49-62, 1999.
[16] C. S. Lina, W. Z. Wub, Y. L. Layc and M. W. Chang, “A digital image-based measurement system for a LCD backlight module,” Optics & Laser Technology, Vol 33, pp499–505, 2001.
[17] K. Belkacem-Boussaid, A. Beghdadi, and H. Depoisot, “Edge detection using Holladay's principle,” Proc. IEEE Int. Conf. Image Processing, Vol.1, pp833-836, 1996.
[18] S. M. Sokolov and A. S. Treskunov, “Automatic vision system for final test of liguid crystal displays,” Robotics and Automation, 1992. Proceeding, IEEE International Conference on , Vol.2 , pp1578-1582, 1992.
[19] Lars Heucke, Mirko Knaak, and Reinhold Orglmester, “A new image segmentation method based on human brightness perception and foveal adaptation,” IEEE Signal Processing Letters, Vol.7, no. 6, pp129-131. June 2000.
[20] R. C. Gonzalez, R. E. Woodes, “Digital Image Processing”.2nd ed., Prentice
Hall, 2002.
[21] “Mura Defects Moving beyond Display Standards”, http://www.propii. com.tw/know.asp .[22] H. R. Blackwell, “Contrast thresholds of the human eye,”J. Opt. Soc., vol. 36, pp624-643, 1946.
[23] 錢志豪，“ 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管理研究所碩士論文，台中，2003.
[24] 曾彥馨，“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，2003.
[25] 陳志忠，“液晶顯示器的像素點缺陷與亮度均一性之自動化檢測”，私立中原大學機械工程研究，2001.
[26] 吳木杏，“Surface Defect Inspection and Analysis of Color Filter Panel”，國立台灣大學資訊工程研究所，2000.
[27] 劉祥吉，“液晶顯示器影像顯示不均檢測之研究”，私立元智大學工業工程與管理研究所碩士論文，2002.
[28] 應宗繼，“發展可靠度資料之異常值診斷方法“,國立中央大學工業管理研究所，1997.
[29] 曾健逢，“大量資料集之線性迴歸模式的估計問題“，國立成功大學統計學研究所，2002.
[30] J. J. Chang, C.Yao, H. C Cheng， ”Investigation of Defect Inspection Technology on TFT-LCD Manufacturing”，科技新儀，民國九十二年十月，pp23-35.
[31] 徐嘉彬，TFT-LCD檢測工程與設備，機械月刊，民國九十二年十二月，pp10-25.
[32] Toru TAMURA，認識限界コントラスト（JND）を用いたLCDの輝度ムラ評価法，Monthly Display，2004/12.
[33] 劉應興，“應用線性迴歸模型”，華泰書局, 1997.
[34] 唐麗英，王春合， “STATISTICA 與基礎統計分析”，滄海書局，台北， 初版，pp350-351，2003.
[35] 蕭金松，“數位影像處理”，全華科技圖書股份有限公司，1998.
[36] 陳連春，“彩色液晶顯示器原理與技術”，建興出版社，台北，2000.