摘要
研究了一种基于数字图像处理的玻璃缺陷在线检测系统,实现了对玻璃缺陷的实时检测、识别和分类,在缺陷提取上运用的是一种改进的归一化互相关算法,该算法可以显著提高图像的匹配速率和缺陷检测效率,在检测效率上比传统的算法提高了25.7%;在缺陷的识别分类上研究采用的是一种改进的支持向量机算法,该算法在检测精度和效率上都有较大的提高,在缺陷种类的识别上,其准确率可达到95%。
The on-line glass defect inspection system based on digital image processing is introduced in this passage. It could realize detection,distinguish and classified on-line. This system uses an improved algorithm of normalized cross correlation,it could improve the image matching rate and the efficiency of defect detection. It was over 25.7% more efficient than the traditional one. An improved support vector machine algorithm is used on the identification and classification of defects. The detection precision and efficiency are greatly improved.In the identification of defects types, its accuracy can reach 95%.
出处
《电子技术应用》
北大核心
2013年第12期90-92,共3页
Application of Electronic Technique
基金
山东省高校科技计划项目(J10LD09)
关键词
玻璃缺陷
数字图像处理
缺陷检测
glass defect
digital image processing
defect detection
