| DC 欄位 |
值 |
語言 |
| DC.contributor | 資訊工程學系 | zh_TW |
| DC.creator | 謝承恩 | zh_TW |
| DC.creator | Cheng-En Shie | en_US |
| dc.date.accessioned | 2008-7-15T07:39:07Z | |
| dc.date.available | 2008-7-15T07:39:07Z | |
| dc.date.issued | 2008 | |
| dc.identifier.uri | http://ir.lib.ncu.edu.tw:444/thesis/view_etd.asp?URN=955202065 | |
| dc.contributor.department | 資訊工程學系 | zh_TW |
| DC.description | 國立中央大學 | zh_TW |
| DC.description | National Central University | en_US |
| dc.description.abstract | 近年來，薄膜電晶體液晶顯示器(Thin-Film Transistor Liquid-Crystal Display) 相關裝置及設備在市場中愈來愈受歡迎，且競爭也愈來愈激烈。因此檢測薄膜電晶體液晶顯示器品質的議題也愈受到注意，這個議題也同時考驗著生產者的能力。薄膜電晶體液晶顯示器主要是由彩色濾光片玻璃基板、薄膜電晶體玻璃基板、液晶、和背光模組所組成。這些組成元件在製程中，會因某些因素而造成產品的瑕疵。為了維護薄膜電晶體液晶顯示器的品質及降低檢測的成本，我們發展一套針對彩色濾光片玻璃基板的Mura瑕疵檢測系統。
我們的方法是透過累加同一彩色濾光片玻璃基板在連續影像序列中的灰階，藉以擴大Mura瑕疵與背景的差異，才得以準確的檢測Mura瑕疵。我們的Mura瑕疵檢測系統主要包含三個部份。一、藉由平均背景影像與未校正影像相除，以校正不均勻的照度。二、藉由相關性 (correlation) 數學模式來校正連續影像之不均勻的位移。三、累加校正影像並以多重解析度影像來分析背景以檢測Mura瑕疵。
我們的檢測影像大小有1024?1024或256?256。在實驗中我們的提供測試影像包含真實影像及人工影像，人工影像中有region Mura、line Mura、cloud Mura、twill line Mura、和ring Mura。從實驗結果中，我們驗證了我們的方法可以穩定地檢測出與背景相差大約5個灰階以上的Mura瑕疵。 | zh_TW |
| dc.description.abstract | As TFT-LCD (Thin-Film Transistor Liquid-Crystal Display) devices get more and more popular and competitive in display market; the inspection of TFT-LCD quality has received increasing attention and become a more critical issue for manufacturers. TFT-LCD is composed of CF (Color Filter) glass substrate, TFT glass substrate, liquid crystal, and back light module. To maintain the quality of TFT-LCD, we develop a Mura inspection method to detect defects on CF glass substrates.
In the proposed method, we accumulate the difference between the Mura defect in an image sequence and background; the background was constructed from multiresolution images to adapt the variant of the acquired images. Then according to analyzing the characters of Mura defects and production lines, we propose the automatic inspection system with three processes. First, we calibrate non-uniform illuminance by dividing an average background. Second, we calibrate non-uniform displacement by using correlation. Finally, we detect Mura defect on the calibrated accumulated images. Our input data is an image sequence (1024?1024 or 256?256). Both real Mura defects and several artificial Mura defects like region Mura, line Mura, cloud Mura, twill line Mura, and ring Mura are used for testing in our experiments. From the experimental results, we find the proposed method can stably detect unclear Mura defect. | en_US |
| DC.subject | 液晶薄膜顯示器 | zh_TW |
| DC.subject | Mura 瑕疵 | zh_TW |
| DC.subject | 多重解析度 | zh_TW |
| DC.subject | TFT-LCD | en_US |
| DC.subject | Mura defect | en_US |
| DC.subject | multiresolution | en_US |
| DC.title | 以累積差異及多重解析度背景相減偵測 LCD Mura 瑕疵 | zh_TW |
| dc.language.iso | zh-TW | zh-TW |
| DC.title | LCD Mura Detection using Accumulated Differences and Multisolution Background Subtraction | en_US |
| DC.type | 博碩士論文 | zh_TW |
| DC.type | thesis | en_US |
| DC.publisher | National Central University | en_US |