Development of Sealant Detection Methods using Auto-Finding Algorithm in TFT-LCD Glass Bonding Process

Abstract

Thin Film Transistor-Liquid Crystal Displays (TFT-LCDs) is most worldwide used device in display industry. In bonding process Color Filter (CF)-substrate and TFT-substrate, they are bonded by sealant which has various overlapped patterns. Measurement of sealant-width which provides evaluation of well-bonded condition is important. Various overlapped patterns and substrate-glass make hard to detect sealant and measure sealant-width. In order to detect sealant, the proposed method automatically sets Area of Interest (AOI) for sealant-width measurement. The modified Otsu method is proposed to eliminate substrate-glass area except sealant area. Modified application of gamma correction is used to select threshold for extracting standard of sealant intensity. In these methods, sealant can be detected automatically and measured in pixel level. For the precise measurement, quartic Facet model is proposed to convert pixel intensity to sub-pixel intensity. The Discrete Orthogonal Polynomials (DOP) are used to verify quartic Facet model. The experimental result shows that new method has higher detection rate than conventional method and good repeatability of measurement than pixel level.