傳統製造共軛性高分子有機電激發光二極體 (PLED, Polymer Light-emitting Diode) 基板的旋轉塗佈 (spin-coating) 方式，製程上需要經過塗佈、曝光、顯影三次過程，才能達到紅、藍、綠的全彩效果。噴墨印製 (IJP, Ink Jet Printing) 技術可直接將紅、藍、綠三原色噴在畫素上，因此比傳統方式，可縮短製作程序，設備成本低，可以大面積製作，而且解析度高，可製作低成本的全彩PLED顯示器。 在噴墨法製作PLED基板的機器上，裝有彩色光藕合元件 (CCD, Charged Couple Device) 攝影機，擷取PF (Polyfluoren, 一種在PLED基板中有發光作用的材料) 溶液塗佈在基板上的影像。而這些PLED基板影像中包含有濺出液滴、塗佈不滿、漏噴、及溢出等瑕疵狀況。在本論文的研究中，我們使用了影像二值化、顆粒濾波與分析、影像增強、空間濾波、影像運算、型態學等常用的影像處理技術，並製作相對的遮罩 (mask) 影像，對這些瑕疵狀況做不同程序的檢測。我們使用LabVIEW IMAQ Vision Builder 6.0影像處理套裝軟體，一個步驟接一個步驟的，從實驗中決定各處理的恰當參數，以組合出合適的處理程序。 本研究要到達快速、準確、實用的階段，還有一段路要走，但是對未來PLED製程自動化檢測上，我們確實已做了一個先期的探索。 Ink Jet Printing technology is applied to the manufacturing of substrate of PLED (Polymer Light-Emitting Diode). Compared with the traditional spinning method, the Ink Jet Printing technology can make the huge area, full-color PLED display with lower cost and faster speed. In order to achieve high yield and high throughput of PLED, the defect detection of substrate is very important. After the ink is printed on the substrate, the defects on the substrate can be detected with the vision method. First, the image of the PF (Polyfluoren) solution coated on the substrate is captured with a CCD camera. Secondly, the image is processed offline to find out the defects such as drop-out, half empty cell and empty cell, over-printed. In this research, image processing technology such as bi-level thresholding, particle filtering and analysis, image enhancement, spatial filter and morphology are adapted to checking the detects. Mask image is also produced to detect these defects in the different procedures. LabVIEW IMAQ Vision Builder 6.0 is the tool for analysis.
