顯示器的發展逐漸由大體積的陰極射線管(CRT),轉型為較為輕薄的液晶平面顯示器(LCD),並且朝著面板大型化與大面積化發展。由於高解析度與高畫面品質的要求,LCD的畫素越做越小,對於LCD面板成品的品質檢驗就愈加困難。目前大部份的生產廠商仍是依賴人工目視檢測,來檢驗畫面的品質,這對於檢測的速度、效能及產品的品質穩定性有相當大的影響。本論文重點在於建立一套自動化的數位光學影像分析系統,來取代人工做目視檢測的工作,其方法為利用CCD、影像擷取卡、光學尺、線性馬達移動平台作為系統工具,配合數位影像處理的光學非接觸性檢測之特性,針對LCD面板的畫素瑕疵與亮度均一性進行判斷,其中包含量測定位、LCD面板取像、影像處理與自行發展的十字標記數學演算法等技術。目前最高檢測速度為每秒30個畫面,已可達到快速檢測的目標。
The liquid crystal displays (LCD) are the leading contenders to replace the cathode ray tube (CRT) displays. Larger displays with higher resolutions will only be possible if the pixel can be kept smaller and brighter. In order to achieve a higher picture performance, the technology of the inspection and the measurement is more important and more difficult for LCD. In most manufactories, some performances on LCD are generally examined by human eyes to check the chromaticity, luminance, viewing-direction, uniformity, etc. ,and this cause much time and higher cost. In this study, an automatic inspection system with a digital image analysis was developed to replace the manual inspection, which include CCD, image interface, linear encoders and linear moter. The main function is to measure the dot defect for checking the pixels under control and to calculate the brightness uniformity of LCD panel. The positioning technique, the grabbing skill on LCD, the image processing ,and a virtual “+” marking algorithm to search the dot defect were constructed in this measuring system. The highest inspection speed on LCD panel could be successfully achieved thirty frames per second.