【摘要】 目前国内外大部分公司采用人工检测法来检测玻璃件的缺陷,由于这种方法受检验人员主观因素影响较大,难以适应大规模自动化生产的需要,为此,提出一种基于机器视觉的气泡缺陷检测方法。分别介绍了几种经典边缘检测算子的锐化以及基于几何特征的气泡缺陷提取方法,并对检测结果进行了比较。实验结果表明,基于面积特征和圆形度的气泡缺陷提取方法能准确地检测出钢化玻璃绝缘子中的气泡缺陷。更多还原

【Abstract】 At present,most of the manufacturers of glass products are using manual detecting method to check the defects in the parts.This is definitely not suitable for large-scaled automatic production because of the influence from subjective factor of the inspectors.Thus the detection method for bubble defects in glass based on machine vision is proposed.Several classical edge detection algorithms sharpening and bubble defect extraction based on geometrical features are introduced,and the detecting results are compared.The experimental results indicate that the extraction methods based on area feature and roundness can precisely detect the bubble defects in toughened glass insulators.更多还原