| dc.description.abstract |
The main purpose of this study is to construct a non-contacted automated optical inspection and using this system to inspect the surface structure and to quantify the defect of grinding disc. This inspection system includes LED illumination device, industrial camera and a personal computer.
Using digital image process with C++ program based on OpenCV designed by ourselves, Grinding disc surface defects inspection are divided into four parts, dimension, mesh, density and crack respectively, Second, each experiment that simulate automated manufacture for fit problem of manufacturing carried out repetitive experimental, Third, measure and quantify the defect of grinding disc, Fourth, set the threshold to distinguish which grinding disc is good.
Cause grinding disc’s surface are complicated, this system focus on how to reduce high frequency noise, and using Fourier transform to solve this problem, and According to each experimental needs, use image process technique like space filter, median filter, Fourier transform, edge detection, feature detection...and so on, to achieve the automated optical inspection.
This study will be used on quality control of grinding disc of automated manufacture, to reduce the waste of human resource and to cost down, and to avoid the damage when someone is using their eye to inspect the quality of grinding disc, thus, developing this kind of automated optical inspection system into automated manufacture will increase accuracy of inspection and bring in lots of revenue.
Keyword: Computer vision, Automated Optical inspection(AOI), Grinding disc, Defect inspection | en_US |