An effective and efficient defect inspection system for TFT-LCD polarised films using adaptive thresholds and shape-based image analyses

Abstract

Defect inspection of polarised films is a crucial process in TFT-LCD (thin film transistor–liquid crystal display) production. In this paper we propose a defect inspection system for TFT-LCD film images that detects film defects in a real-time production environment and classifies them based on the type of a defect. The proposed system is designed to locate defects promptly using an adaptive threshold technique and determines the defect type through image analysis using various features, such as the geometric characteristics and the shape descriptor with intensity distribution called the shape + ID descriptor. The experimental results using a set of test images obtained in a real production line are quite promising. Compared with existing methods, our method identifies defects effectively and efficiently enough to be used in a real-time environment. It also achieves a high correctness in classifying the defect type for various types of defects. In addition, it demonstrates robustness with respect to scale and rotational transformation.