摘要
磁粉探伤广泛应用于铁磁性工件表面和近表面的裂纹检测,但目前仍采用人工目检进行缺陷判别。因此,提出了一种运用机器视觉检测技术的圆柱形工件外表面裂纹自动检测方法,针对柱形轴承零件在车削加工过程中产生的纵向表面裂纹,使用工业相机拍摄工件柱面磁痕图像,运用数字图像处理技术提取出疑似裂纹区域的特征,然后基于这些特征,利用模式识别中的分类器技术完成轴承零件柱面裂纹的自动识别。试验结果表明:该方法不仅对柱面裂纹、纤维物和磁悬液滴等真、伪缺陷有较高的识别率,而且检测效率较高。
Magnetic particle inspection is widely used for detection of ferromagnetic workpiece surface cracks, but the result of detection depends on manual visual inspection. For a kind of longitudinal surface cracks occurre in the pro-cessing of turning, first magnetic image of bearing is gotten the by using industrial camera and the features of suspected crack areas is extracted through the technology of digital image processing. Based on these features, the system could recognize the cracks automatically with classifiers of pattern recognition. The experiment indicates that this method could not only discriminate real cracks and the false with higher recognition, such as fibers and magnetic droplets. ,but also improve the detection efficiency greatly with a sense of application research.
出处
《轴承》
北大核心
2017年第3期45-49,55,共6页
Bearing
基金
西南科技大学重点实验室基金(13zxtk05)
校创新团队基金(14tdtk01)
关键词
滚动轴承
无损检测
磁粉探伤
图像处理
表面裂纹
rolling bearing
non - destructive testing （NDT）
magnetic particle inspection
image processing
surface crack
