基于形态学的双螺孔自动检测技术
|
摘要
为满足双螺孔准确、高速和客观的实时检测要求,搭建图像采集系统采集待检产品图像,研究了基于数字图像处理的视觉检测系统实现的关键技术。针对需同时检测双孔是否为螺纹孔及是否有烂牙和螺纹小径尺寸是否在公差范围内等多个检测特征,通过对图像依次作形态学预处理、最大类间方差阈值法分割,应用设定检测区域、图像比较和区域描述技术对螺孔进行分析,并定义检测目标的灰度值、像素个数作为螺孔特征识别的主要依据。开发了机器视觉自动检测系统,包括视觉检测软件及图像采集机构,试用结果表明该系统能准确实现双螺孔的自动检测,系统能满足检测的实时性与准确性要求。
For meeting the requirement of precise,high-speed and objective real-time inspection of the double screw holes,an image acquisition system was constructed to collect the product images to be inspected,during which the key technology for visual inspection system realization was studied. Due to the multiple inspection characteristics of double screw holes including the screw thread existence,the rotten tooth existence,and the size tolerance of small diameter,the image was processed through the morphology pretreatment and the maximum between-cluster variance threshold value method segmentation. The inspection zone,image comparison,and region description technology were used to analyze the screw hole. The gray value and pixels quantity were adopted as the main proof of screw hole characteristic recognition. A machine vision automatic inspection system was developed with the visual inspection software and the image capture mechanism. The test result shows that the dual screw holes can be successfully inspected using the proposed system.
For meeting the requirement of precise,high-speed and objective real-time inspection of the double screw holes,an image acquisition system was constructed to collect the product images to be inspected,during which the key technology for visual inspection system realization was studied. Due to the multiple inspection characteristics of double screw holes including the screw thread existence,the rotten tooth existence,and the size tolerance of small diameter,the image was processed through the morphology pretreatment and the maximum between-cluster variance threshold value method segmentation. The inspection zone,image comparison,and region description technology were used to analyze the screw hole. The gray value and pixels quantity were adopted as the main proof of screw hole characteristic recognition. A machine vision automatic inspection system was developed with the visual inspection software and the image capture mechanism. The test result shows that the dual screw holes can be successfully inspected using the proposed system.
引文
[1]陈挺,茅振华,周闻青,等.Lab VIEW平台的平面螺纹测量系统研制[J].中国计量学院学报,2015,26(2):145-150.
[2]王蕴哲,岳晓峰.内螺纹机器视觉检测系统[J].长春工业大学学报,2017,38(2):189-193.
[3]田野,叶兵,王连兵.基于机器视觉的内螺纹检测的实现方法[J].仪表技术与传感器,2018(1):64-70.
[4]李学聪.基于机器视觉指针表检测的关键技术研究[D].广州:广东工业大学,2014:23.
[5]BOUAYNAYA N,CHARIF-CHEFCHAOUNI M,SCHONFELD D.Theoretical foundations of spatially-variant mathematical morphology part I:binary images[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2008,30(5):823-836.
[6]RANQAYYAN R M,NQUYEN T M,AYRES F J,et al.Effect of pixel resolution on textures of features of breast masses in mammograms[J].Journal of Digital Imaging,2010,23(5):547-553.
[7]张银玲.基于图像的螺纹检测装置的研制[D].哈尔滨:黑龙江大学,2017:23.
[8]SONKA M,HLAVAC V,BOYLE R.图像处理、分析与机器视觉[M].艾海舟,苏延超,译.3版.北京:清华大学出版社,2014:124-224.
[9]张爽.列车轮对几何参数在线检测关键技术研究[D].吉林:吉林大学,2017:54.
[2]王蕴哲,岳晓峰.内螺纹机器视觉检测系统[J].长春工业大学学报,2017,38(2):189-193.
[3]田野,叶兵,王连兵.基于机器视觉的内螺纹检测的实现方法[J].仪表技术与传感器,2018(1):64-70.
[4]李学聪.基于机器视觉指针表检测的关键技术研究[D].广州:广东工业大学,2014:23.
[5]BOUAYNAYA N,CHARIF-CHEFCHAOUNI M,SCHONFELD D.Theoretical foundations of spatially-variant mathematical morphology part I:binary images[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2008,30(5):823-836.
[6]RANQAYYAN R M,NQUYEN T M,AYRES F J,et al.Effect of pixel resolution on textures of features of breast masses in mammograms[J].Journal of Digital Imaging,2010,23(5):547-553.
[7]张银玲.基于图像的螺纹检测装置的研制[D].哈尔滨:黑龙江大学,2017:23.
[8]SONKA M,HLAVAC V,BOYLE R.图像处理、分析与机器视觉[M].艾海舟,苏延超,译.3版.北京:清华大学出版社,2014:124-224.
[9]张爽.列车轮对几何参数在线检测关键技术研究[D].吉林:吉林大学,2017:54.