- 學位論文
液晶顯示器品質之互動分類資料庫
Interactive Classification Database for LCD Monitor Quality
摘要
由於液晶面板生產技術日趨成熟，品質檢測及衡量標準的重要性也相對提高。液晶顯示器的品質衡量包含動態影像品質與Mura兩部份，動態影像品質在於衡量動態影像的模糊程度，主要受到液晶面板反應時間的影響，可採用TrTf 與Pursuit N-BEW兩種模式來偵測面板的反應速度; Mura則偵測液晶面板在低亮度(灰階92、灰階128)背景時，所呈現亮度不均勻的區塊。廠商積極發展機器視覺的品質分析系統，動態影像品質與Mura通常被分開評等，實際上液晶顯示器品質應同時顧及兩者，故本研究欲結合動態影像品質與Mura，設計整合式品質分類資料庫系統，提供綜合分級結果及回饋資訊，協助廠商改良產品。在品質分類模組採用人因工程技術，導入人眼因素提供產品的分類規則，使機器視覺的分析趨近人眼辨識需求。互動查詢介面採用瀏覽式查詢QBB視覺查詢語言，結合關聯資料模式與視窗使用介面概念，提供使用者多樣化、具彈性的查詢功能。
並列摘要
As techniques of LCD manufacture are getting more and more superior, the criteria of display quality are stricter these years. The inspections and measuring standards for LCD quality are even more important. There are two categories of inspections for LCD quality which are motion image quality and Mura. The purpose of motion image quality inspection is to measure the blur degree during motion image display which is influenced by response time of LCD panel. There are two distinct methods for motion image quality inspection: TrTf and Pursuit N-BEW. The purpose of Mura inspection is to detect the defects as low brightness background (gray level 92, gray level 128). Manufacturers have developed several automatic inspection systems with machine vision. The evaluations of motion image and Mura are usually considered separately. However, the LCD quality should consider both categories. In this paper, we propose a composite quality classification database which combines motion image quality and Mura to provide composite classification results and feedback information. Ergonomics rules of human eyes is adopted as the classification rules in the quality classification module, which makes the inspection result form machine vision suit with the recognizing requirement of human eyes. The interactive query interface applies QBB (Query by Browsing), a visual query language, combining the concepts of relational model and desktop user interface paradigm. Therefore, the interactive query interface provides variety and flexible query functions.