基于机器视觉的玻璃磨边缺陷检测

Detection of Defects on Ground Glass Edge based on Machine Vision

玻璃边部的磨削质量的快速检测是保证玻璃品质的重要措施,本文对现有的玻璃磨边缺陷进行了分类与成因分析,选用了一套合理的玻璃磨边缺陷检测光源,设计了一套基于机器视觉的玻璃边部缺陷检测装置,提出一种结合了快速傅立叶变换、高斯滤波、亚像素边缘阈值分割、数学形态学运算、频域处理和最小二乘法的检测方法,实验证明所提出的检测方法能够快速有效地检测出亮斑、白线和爆边3种缺陷.

The rapid detection of glass edge grinding quality is an important measure to ensure the quality of glass. The elassifieation of existing glass edge defects is analyzed. The cause of formation of defect is concisely intro- duced. A kind of reasonable light source for detecting defect on ground glass edge is selected. The detection device for glass edge defects based on machine vision has been established. Detection methods based on Gauss filter, sub- pixel threshold segmentation, mathematical morphology operation, frequency domain processing and the least square method are proposed. It is shown that the methods can effectively detect the bright spot, white line and edge breakage.