Research Of Auto-focus And Defect Classification In The FPD Lighting Detection System

The FPD (Flat Panel Display) with high resolution, high brightness, low power consumption, small size and light weight, etc, are widely used in the display area. In order to guarantee the display quality and improve the production of FPD, defect detection is essential during FPD manufacturing process. There is a urgent need of fully automated defect detection system to reduce costs and improve production efficiency. FPD module manufacturing process panel lighting detection based on the photoelectric response measurement, including life expectancy and quality of defect detection, such as points, line, mura and other defects. However, due to industrial cameras in the detection system generally does not auto-focus and the best method of detection of mura need to make the camera lens to adjust to the shooting of the FPD image with the specified texture cycle. The processing of making FPD which includes more than100working procedures is very complicated, so there must be many kinds of defects and all the manufactures want a software that can quickly classification all these defects. In this paper, due to these questions here do the following work:1. Tested the airspace, entropy and frequency domain algorithm separately and designed the focusing process in order to achieve the FPD imaging system of auto-focus.2. Calculated the texture cycle of the FPD image with Fourier transform as feedback to further adjust the camera lens and auto focus to meet the specified value of texture cycle.3. Developed the classification program,make sure it can classify the defects with any number of pattern, any kind and any grade of defects.