Research and Development of Intelligent On-Line Real-time Defect Inspection System for Polymer Polarizer

Abstract

This paper aimed at developing a set of on-line real-time defect inspection system for Polymer polarizer. It intended to help producers make different improvements for the spot or line defects in manufacturing on-line polarizer. This paper used line-scan charged couple device (CCD) to capture the defect images of polarizer: firstly, to scale down the captured images by 64 times with downsampling compression; then, we utilized Laplace operator to find out the edges of defects and the shield with isotropic results for 45° and 90° incremental rotation. Subsequently, we used the statistical decision method of threshold value to divide up the edges of defects that were discovered with Laplace operator from the image. Eventually, we differentiated the spot defects, like dust, foreign objects, scars of hit and air bubbles in the production process from the line defects, like scratches in the follow-up handling process with the straight line detection characteristic of Hough Transform. This paper captured 200 pieces of defect images of polarizer samples in all and then input them to the inspection system developed by ourselves for test. The differentiation rate of defects amounted to 98%, which proved that we had successfully developed a set of on-line real time automatic optical inspection (AOI) system suitable to be applied to polymer polarizer.