基于视觉的圆筒内外壁均匀度在线检测系统
|
摘要
圆筒在机械、能源、军事等领域有着大规模的应用。但在圆筒的生产过程中,一直采用传统的人工离线检测方法进行检测。为提高圆筒检测的速度,目前工厂急切需要能够实现圆筒在线检测的方法和系统。但由于圆筒的一端封闭,且生产现场温度高,现有的机器视觉检测方法和系统难以实现圆筒的实时在线检测。本文提出了一种基于视觉的一端封闭圆筒内壁几何参数在线检测方法,设计了一种特殊结构的成像机构,通过对激光反射所形成的反映圆筒内壁几何尺寸变化的光斑图像进行处理,从而获取圆筒内壁几何参数变化信息。
根据机器视觉原理,对圆筒内壁进行检测,本文提出以激光三角测距方法为基础,使用新的成像方法的检测方法。该方法采用反射镜对激光光路进行变向,使激光垂直照射到圆筒内壁表面,形成反映圆筒内壁几何尺寸变化的光斑图像,采用三棱镜对激光光斑图像成像。该方法能够实现激光垂直射入,图像垂直传出。一方面,该方法可以完整地检测圆筒内壁,即使是接近圆筒封闭一端的远端也可以实现内壁检测;另一方面,该方法将成像所需的电子设备远离被测圆筒设置,可以满足高温生产环境的长时间工作的要求。
根据新的圆筒检测方法,设计了圆筒检测系统。该系统由包括激光器、反射镜和三棱镜在内的成像系统、包括镜头、相机和采集卡在内的图像采集系统和机械系统组成。文中详细地分析了光路反射系统的组成,以半五角棱镜作为最终成像装置。文中详细地分析了镜头与相机之间的关系,并根据实际检测要求仔细地计算了各个参数,包括视角、视野、工作距离。根据计算,选择标准PAL制式的相机,配合图像采集卡可以达到768×576的分辨率,需要镜头视角小于7.84°。设计制作了能够固定激光器、反射镜、三棱镜和相机,且反射镜、三棱镜位置角度可调的机械装置。将各个组件与机械装置进行组装,能够实现实时在线检测的演示模型样机。
就激光光斑中心定位算法进行了研究,设计了将检测系统采集到的光斑图像自动转化成光斑位置数据的程序。经过演示模型样机的实验表明,该系统对于内壁尺寸检测的分辨率可达0.1毫米/像素,内壁尺寸变化检测范围可达30毫米,完成单点的检测时间小于40毫秒。
Cylinder has a large-scale application in machinery, energy, military and other fields. But in the process of cylinder production, traditional off-line methods of detection is using. To improve the speed of cylinder detection, cylinder on-line detection methods and systems are urgently needed to achieve at the factory. However, for the cylinder has one end closed and the temperature at the production scene is very high, the real-time on-line detection can't be achieved with the existing testing methods and machine vision systems. A vision-based one end closed cylinder wall geometric parameters on-line measurement method is presented in this paper, an imaging agency with special structure is designed, by processing the spot image which is formed by the laser reflector and reflects the cylinder wall geometry change, cylinder wall geometric parameters change information is accessed.
The principle of machine vision is adopted, to test the inner wall of the cylinder. The method proposed in this paper is based on the laser triangulation method, with a new imaging method designed. Reflector is used to change the light path of the laser in this detection method, to make laser beam to the cylindrical surface of the inner wall, a spot image which reflects cylinder wall geometry changes formed, prism is used to imaging the laser spot on the image. The vertical in of laser and vertical out of image can be realized by this method. On the one hand, the method can detect cylinder wall perfectly, even the detection to the wall near the cylinder closed distal end can be achieved. The other hand, required electronic imaging equipment is setup away from the measured cylinder, the requirements which is need for long time working in the high temperature production environment is met.
According to the new cylinder detection method, the cylinder detection system is designed. The system consists of imaging systems, which includes laser, mirror and prism, image acquisition system, which includes the lens, cameras and image acquisition card, and the mechanical system. The composition the optical reflector system is analyzed in detail, eventually a half pentagonal prism is taken to be the imaging devices. In this paper, the relationship between the camera and the lens is analyzed in detail, various parameters including the angle of view, field of view and working distance in accordance with the actual requirements is carefully calculated. According to the computation standard PAL camera which can reach 768×576 resolution when working with image acquisition card is selected, camera angle is need to be less than 7.84°. The mechanical devices has been designed and produced to the fixed laser, mirror, prism and cameras, with the position and angle of the mirror and prism adjustable. Demonstration model prototype which can achieve real-time on-line cylinder detection is made by assembling all components and mechanical devices.
Laser spot positioning algorithm is studied in this paper, and program which can automatically change the spot images collected by the detection system to spot location data is designed. After the experiments of demonstration prototype model, it is showed that the resolution of the detection system of the inner wall can be 0.1 mm/pixel, wall size change detection range is not less than 30mm, and the detection time of one spot is less than 40ms.
根据机器视觉原理,对圆筒内壁进行检测,本文提出以激光三角测距方法为基础,使用新的成像方法的检测方法。该方法采用反射镜对激光光路进行变向,使激光垂直照射到圆筒内壁表面,形成反映圆筒内壁几何尺寸变化的光斑图像,采用三棱镜对激光光斑图像成像。该方法能够实现激光垂直射入,图像垂直传出。一方面,该方法可以完整地检测圆筒内壁,即使是接近圆筒封闭一端的远端也可以实现内壁检测;另一方面,该方法将成像所需的电子设备远离被测圆筒设置,可以满足高温生产环境的长时间工作的要求。
根据新的圆筒检测方法,设计了圆筒检测系统。该系统由包括激光器、反射镜和三棱镜在内的成像系统、包括镜头、相机和采集卡在内的图像采集系统和机械系统组成。文中详细地分析了光路反射系统的组成,以半五角棱镜作为最终成像装置。文中详细地分析了镜头与相机之间的关系,并根据实际检测要求仔细地计算了各个参数,包括视角、视野、工作距离。根据计算,选择标准PAL制式的相机,配合图像采集卡可以达到768×576的分辨率,需要镜头视角小于7.84°。设计制作了能够固定激光器、反射镜、三棱镜和相机,且反射镜、三棱镜位置角度可调的机械装置。将各个组件与机械装置进行组装,能够实现实时在线检测的演示模型样机。
就激光光斑中心定位算法进行了研究,设计了将检测系统采集到的光斑图像自动转化成光斑位置数据的程序。经过演示模型样机的实验表明,该系统对于内壁尺寸检测的分辨率可达0.1毫米/像素,内壁尺寸变化检测范围可达30毫米,完成单点的检测时间小于40毫秒。
Cylinder has a large-scale application in machinery, energy, military and other fields. But in the process of cylinder production, traditional off-line methods of detection is using. To improve the speed of cylinder detection, cylinder on-line detection methods and systems are urgently needed to achieve at the factory. However, for the cylinder has one end closed and the temperature at the production scene is very high, the real-time on-line detection can't be achieved with the existing testing methods and machine vision systems. A vision-based one end closed cylinder wall geometric parameters on-line measurement method is presented in this paper, an imaging agency with special structure is designed, by processing the spot image which is formed by the laser reflector and reflects the cylinder wall geometry change, cylinder wall geometric parameters change information is accessed.
The principle of machine vision is adopted, to test the inner wall of the cylinder. The method proposed in this paper is based on the laser triangulation method, with a new imaging method designed. Reflector is used to change the light path of the laser in this detection method, to make laser beam to the cylindrical surface of the inner wall, a spot image which reflects cylinder wall geometry changes formed, prism is used to imaging the laser spot on the image. The vertical in of laser and vertical out of image can be realized by this method. On the one hand, the method can detect cylinder wall perfectly, even the detection to the wall near the cylinder closed distal end can be achieved. The other hand, required electronic imaging equipment is setup away from the measured cylinder, the requirements which is need for long time working in the high temperature production environment is met.
According to the new cylinder detection method, the cylinder detection system is designed. The system consists of imaging systems, which includes laser, mirror and prism, image acquisition system, which includes the lens, cameras and image acquisition card, and the mechanical system. The composition the optical reflector system is analyzed in detail, eventually a half pentagonal prism is taken to be the imaging devices. In this paper, the relationship between the camera and the lens is analyzed in detail, various parameters including the angle of view, field of view and working distance in accordance with the actual requirements is carefully calculated. According to the computation standard PAL camera which can reach 768×576 resolution when working with image acquisition card is selected, camera angle is need to be less than 7.84°. The mechanical devices has been designed and produced to the fixed laser, mirror, prism and cameras, with the position and angle of the mirror and prism adjustable. Demonstration model prototype which can achieve real-time on-line cylinder detection is made by assembling all components and mechanical devices.
Laser spot positioning algorithm is studied in this paper, and program which can automatically change the spot images collected by the detection system to spot location data is designed. After the experiments of demonstration prototype model, it is showed that the resolution of the detection system of the inner wall can be 0.1 mm/pixel, wall size change detection range is not less than 30mm, and the detection time of one spot is less than 40ms.
引文
[1]唐辉,叶静.国标《包装用于发货、运输和收货标签的一维条码和二维条码》的研究与制定.交通标准化,2006(10):42-46.
[2]肖朝晖,纪钢.条码技术及其在包装防伪中的应用.包装工程,2005,26(3):68-70.
[3]孙旭,杭柏林.错位散斑技术及其在轮胎检测中的应用.轮胎工业,2007,27(7):441-443.
[4]陈亚军,张二虎.基于机器视觉的印品质量在线检测.今日印刷,2005(11):12-14.
[5]苑玮琦,张永辉.基于DSP5402的指纹识别系统实现问题研究.微计算机信息,2007,23(32):195-196,187.
[6]田莹,苑玮琦.人耳识别技术研究综述.计算机应用研究,2007,24(4):21-25.
[7]苑玮琦,林忠华,徐露.一种快速而准确的虹膜识别方法.计算机应用研究,2007,24(6):197-200.
[8]王坚,任现君,苏建刚.基于激光投影法的火炮身管内膛疵病深度测量.中国测试技术,2004,30(5):16-17,54.
[9]窦益华,贾连璧,郑跃辉等.油管内壁损伤检测及评判方法.1998,26(4):28-31.
[10]王坚,苏建刚,任现君.身管内壁形貌检测技术研究.中国测试技术,2004,30(4):52-54.
[11]李莉,邸旭,马宏.炮管内膛型廓自动光电检测系统.仪器仪表学报,2006,27(6):1544-1546.
[12]孙吉红,杨清文,战延谋等.基于激光测量的身管磨损量检测系统研制.计算机测量与控制,2004,12(12):1131-1133.
[13]吴恩启,李江雄,柯映林.微细管道内壁三维测量技术研究.计算机辅助设计与图形学学报,2005,17(1):151-154.
[14]Manabe T,Tsujimura T,TomitaS,et al.Acoustic inspection system for underground telecommunication conduit.Instrumentation and Measurement Technology Conference,1994:761-764.
[15]Okada T,Sanemori,T.MOGRER:A vehicle study and realization for in-pipe inspection tasks.IEEE Journal of Robotics and Automation,1987,3(6):573-582.
[16]Basu A,Southwell D.Omni-directional sensors for pipe inspection.IEEE International Conference on Intelligent Systems for the 21st Century,1995:3107-3112.
[17]Kuntze H B,Haffner H.Experiences with the development of a robot for smart multisensoric pipe inspection.IEEE International Conference on Robotics and Automation,1998:1773-1778.
[18]Duran O,Althoefer K,Seneviratne L D.Laser profiler model for robot-based pipe inspection.Proceedings of World Automation Congress,2004,15(28):353-358.
[19]Duran O,Althoefer K,Seneviratne L D.Experiments using a laser-based transducer and automated analysis techniques for pipe inspection.Proceedings of IEEE International Conference on Robotics and Automation,2003:2561-2566.
[20]Duran O,Althoefer K,Seneviratne L D.A sensor for pipe inspection:model,analysis and image extraction.Proceedings of International Conference on Image Processing,2003:597-600.
[21]Duran O,Aithoefer K,Seneviratne L D.Automated Pipe Defect Detection and Categorization Using Camera/Laser-Based Profiler and Artificial Neural Network.IEEE Transactions on Automation Science and Engineering,2007,4(1):118-126.
[22]Costa R R,dosReis N R S,Liu H,et al.Robotized system for in-pipe inspection.IEEE International Symposium on Industrial Electronics,Brazil,2003:1021-1025.
[23]Orghidan R,Salvi J,Mouaddib E M.Accuracy estimation of a new omnidirectional 3D vision sensor.Processing of IEEE International Conference on Image Processing,Italy,2005:Ⅲ-365-8.
[24]庄葆华,王少清,张吉华等.高精度激光三角法位移测量被测表面倾斜影响研究.计量技术,1996(2):2-4.
[25]张文伟,庄葆华,刘迈等.飞机表面划痕非接触测量仪的研究.计量学报,1998,19(4):264-269.
[26]庄葆华,张文伟,刘迈等.飞机表面划痕的非接触测量.应用激光,1997,17(2):49-53.
[27]李兰君,喻寿益.单点激光三角法测距及其标定.研究与开发,2003(10):49,51.
[28]林小倩,林斌,潘泰才.基于CMOS单点激光三角法测距系统设计.光学仪器,2006,28(2):27-30.
[29]冯俊艳,冯其波,匡萃方.高精度激光三角位移传感器的技术现状.应用光学,2004,25(3):33-36.
[30]林小倩.基于CMOS单点激光三角法测距系统研究:(硕士学位论文).杭州:浙江大学,2006.
[31]宋开臣,张国雄.激光三角法扫描测头特性的研究.中国机械工程,2000,11(4):385-388.
[32]贾永红.数字图像处理.武汉:武汉大学出版社,2003.
[33]徐亚明,邢诚,刘冠兰等.几种激光光斑中心检测方法的比较.海洋测绘,2007,27(2):74-76.
[34]杨耀权,施仁,于希宁等.用Hough变换提高激光光斑中心定位精度的算法.光学学报,1999,19(12):1655-1660.
[35]王芳荣,赵丁选,廖宗建等.激光光斑中心空间定位方法的研究.激光技术,2005,29(1):87-89.
[36]孔兵,王昭,谭玉山.基于圆拟合的激光光斑中心检测算法.红外与激光工程,2002,31(3):275-279.
[37]富容国,常本康,刘淑荣.1.06 μm激光光斑中心的计算机分析.光电子技术与信息,2004,17(2):26-29.
[38]黄浩,詹玲.一种光斑中心定位算法的研究与实现.小型微型计算机系统,2005,26(7):1196-1199.
[39]陈家凤,彭其圣.图像采集卡开发库函数调用及二次开发探讨.计算机应用与软件,2005,22(6):84-86.
[40]闰晓霞,程荫杭.基于DH-CG300图像采集卡的图像采集平台的开发.交通与计算机,2005,23(5):49-52.
[41]孙玉杰,汪岳峰,凌军.图像采集卡对光学仪器性能数字化检测的影响研究.光学仪器,2003,25(2):8-11.
[2]肖朝晖,纪钢.条码技术及其在包装防伪中的应用.包装工程,2005,26(3):68-70.
[3]孙旭,杭柏林.错位散斑技术及其在轮胎检测中的应用.轮胎工业,2007,27(7):441-443.
[4]陈亚军,张二虎.基于机器视觉的印品质量在线检测.今日印刷,2005(11):12-14.
[5]苑玮琦,张永辉.基于DSP5402的指纹识别系统实现问题研究.微计算机信息,2007,23(32):195-196,187.
[6]田莹,苑玮琦.人耳识别技术研究综述.计算机应用研究,2007,24(4):21-25.
[7]苑玮琦,林忠华,徐露.一种快速而准确的虹膜识别方法.计算机应用研究,2007,24(6):197-200.
[8]王坚,任现君,苏建刚.基于激光投影法的火炮身管内膛疵病深度测量.中国测试技术,2004,30(5):16-17,54.
[9]窦益华,贾连璧,郑跃辉等.油管内壁损伤检测及评判方法.1998,26(4):28-31.
[10]王坚,苏建刚,任现君.身管内壁形貌检测技术研究.中国测试技术,2004,30(4):52-54.
[11]李莉,邸旭,马宏.炮管内膛型廓自动光电检测系统.仪器仪表学报,2006,27(6):1544-1546.
[12]孙吉红,杨清文,战延谋等.基于激光测量的身管磨损量检测系统研制.计算机测量与控制,2004,12(12):1131-1133.
[13]吴恩启,李江雄,柯映林.微细管道内壁三维测量技术研究.计算机辅助设计与图形学学报,2005,17(1):151-154.
[14]Manabe T,Tsujimura T,TomitaS,et al.Acoustic inspection system for underground telecommunication conduit.Instrumentation and Measurement Technology Conference,1994:761-764.
[15]Okada T,Sanemori,T.MOGRER:A vehicle study and realization for in-pipe inspection tasks.IEEE Journal of Robotics and Automation,1987,3(6):573-582.
[16]Basu A,Southwell D.Omni-directional sensors for pipe inspection.IEEE International Conference on Intelligent Systems for the 21st Century,1995:3107-3112.
[17]Kuntze H B,Haffner H.Experiences with the development of a robot for smart multisensoric pipe inspection.IEEE International Conference on Robotics and Automation,1998:1773-1778.
[18]Duran O,Althoefer K,Seneviratne L D.Laser profiler model for robot-based pipe inspection.Proceedings of World Automation Congress,2004,15(28):353-358.
[19]Duran O,Althoefer K,Seneviratne L D.Experiments using a laser-based transducer and automated analysis techniques for pipe inspection.Proceedings of IEEE International Conference on Robotics and Automation,2003:2561-2566.
[20]Duran O,Althoefer K,Seneviratne L D.A sensor for pipe inspection:model,analysis and image extraction.Proceedings of International Conference on Image Processing,2003:597-600.
[21]Duran O,Aithoefer K,Seneviratne L D.Automated Pipe Defect Detection and Categorization Using Camera/Laser-Based Profiler and Artificial Neural Network.IEEE Transactions on Automation Science and Engineering,2007,4(1):118-126.
[22]Costa R R,dosReis N R S,Liu H,et al.Robotized system for in-pipe inspection.IEEE International Symposium on Industrial Electronics,Brazil,2003:1021-1025.
[23]Orghidan R,Salvi J,Mouaddib E M.Accuracy estimation of a new omnidirectional 3D vision sensor.Processing of IEEE International Conference on Image Processing,Italy,2005:Ⅲ-365-8.
[24]庄葆华,王少清,张吉华等.高精度激光三角法位移测量被测表面倾斜影响研究.计量技术,1996(2):2-4.
[25]张文伟,庄葆华,刘迈等.飞机表面划痕非接触测量仪的研究.计量学报,1998,19(4):264-269.
[26]庄葆华,张文伟,刘迈等.飞机表面划痕的非接触测量.应用激光,1997,17(2):49-53.
[27]李兰君,喻寿益.单点激光三角法测距及其标定.研究与开发,2003(10):49,51.
[28]林小倩,林斌,潘泰才.基于CMOS单点激光三角法测距系统设计.光学仪器,2006,28(2):27-30.
[29]冯俊艳,冯其波,匡萃方.高精度激光三角位移传感器的技术现状.应用光学,2004,25(3):33-36.
[30]林小倩.基于CMOS单点激光三角法测距系统研究:(硕士学位论文).杭州:浙江大学,2006.
[31]宋开臣,张国雄.激光三角法扫描测头特性的研究.中国机械工程,2000,11(4):385-388.
[32]贾永红.数字图像处理.武汉:武汉大学出版社,2003.
[33]徐亚明,邢诚,刘冠兰等.几种激光光斑中心检测方法的比较.海洋测绘,2007,27(2):74-76.
[34]杨耀权,施仁,于希宁等.用Hough变换提高激光光斑中心定位精度的算法.光学学报,1999,19(12):1655-1660.
[35]王芳荣,赵丁选,廖宗建等.激光光斑中心空间定位方法的研究.激光技术,2005,29(1):87-89.
[36]孔兵,王昭,谭玉山.基于圆拟合的激光光斑中心检测算法.红外与激光工程,2002,31(3):275-279.
[37]富容国,常本康,刘淑荣.1.06 μm激光光斑中心的计算机分析.光电子技术与信息,2004,17(2):26-29.
[38]黄浩,詹玲.一种光斑中心定位算法的研究与实现.小型微型计算机系统,2005,26(7):1196-1199.
[39]陈家凤,彭其圣.图像采集卡开发库函数调用及二次开发探讨.计算机应用与软件,2005,22(6):84-86.
[40]闰晓霞,程荫杭.基于DH-CG300图像采集卡的图像采集平台的开发.交通与计算机,2005,23(5):49-52.
[41]孙玉杰,汪岳峰,凌军.图像采集卡对光学仪器性能数字化检测的影响研究.光学仪器,2003,25(2):8-11.