面向高速视觉检测的精确抓拍安全策略研究
|
摘要
为了解决高速流水线机器视觉检测相机精确抓拍问题,以香烟爆珠为例,提出了两层网络控制的精确抓拍安全策略。两层网络的底层网络实现对目标抓拍,通过线搜索获得感兴趣区域位置,并上送到高层网络进行控制;高层网络主要处理复杂算法为基于柔性专家控制的迭代学习控制和粒子滤波相结合的相机快门控制方法,不断更新香烟爆珠在图像中的位置偏移量,进一步控制相机快门延迟时间,实现精确抓拍高速运动的爆珠的图像。针对上述方法,建立了控制模型,并对模型进行仿真、分析。最后,通过实验平台测试,可以抵消平台运行中的干扰,保证了精确抓拍安全策略的可行性。
To precisely capture the object in the scene of high-speed pipeline by machine vision inspection,this paper takes accurate snapshots of cigarette pop bead as an example,and a two-layer network control algorithm to capture the details in the scene is proposed.The bottom structure of the two-layer network captures the object image and obtains the area of perceptive interests through line search.This area is transmitted to the upper layer to achieve the following control. The upper layer network mainly deals with the complex algorithm. The camera shutter control method is the combination of iterative learning control and particle filter based on the flexible expert control. The camera shutter delay time is adjusted by the continuously updated position of cigarette pop bead. In this way,the accurate position of pop bead under the high-speed movement condition can be achieved. Based on the above method,the control model is formulated,simulated and analyzed. Evaluated by the experimental platform,the results show that the interferences can be reduced and the safe strategy of accurate capture is feasible.
To precisely capture the object in the scene of high-speed pipeline by machine vision inspection,this paper takes accurate snapshots of cigarette pop bead as an example,and a two-layer network control algorithm to capture the details in the scene is proposed.The bottom structure of the two-layer network captures the object image and obtains the area of perceptive interests through line search.This area is transmitted to the upper layer to achieve the following control. The upper layer network mainly deals with the complex algorithm. The camera shutter control method is the combination of iterative learning control and particle filter based on the flexible expert control. The camera shutter delay time is adjusted by the continuously updated position of cigarette pop bead. In this way,the accurate position of pop bead under the high-speed movement condition can be achieved. Based on the above method,the control model is formulated,simulated and analyzed. Evaluated by the experimental platform,the results show that the interferences can be reduced and the safe strategy of accurate capture is feasible.
引文
[1]李新华,蔡大宇,彭辉,等.一种高速灌装自适应控制方法[J].仪器仪表学报,2017,38(7):1738-1746.LI X H,CAI D Y,PENG H,et al.Adaptive control approach for high-speed filling process[J].Chinese Journal of Scientific Instrument,2017,38(7):1738-1746.
[2]秦武旻,朱长婕.虹膜快速检测与精确定位的算法研究[J].电子测量技术,2017,36(4):25-28.QIN W M,ZHU CH J.Efficient and precise optimization algorithm for iris location[J].Electronic Measurement Technology,2017,36(4):25-28.
[3]LI P AN,LEI D ZH,SUN X J,et al.An investigation and application of high-speed measurement and precise control based on machine vision[J].Systems,Science and Technology,2016,doi:10.5013/IJSSST.a.17.04.13.
[4]周文举,费敏锐,王海宽,等.基于迭代学习和卡尔曼滤波的相机精确抓拍控制方法[J].仪器仪表学报,2014,35(5):1044-1051.ZHOU W J,FEI M R,WANG H K,et al.Accurate camera image capture control based on iterative learning and Kalman filter[J].Chinese Journal of Scientific Instrument,2014,35(5):1044-1051.
[5]YANG B,LI H G,KANG ZH B,et al.Hamiltonianbased binocular visual seroing of camera-in-hand robotic systems[J].International Conference on Modelling,2012:388-393.
[6]刘建,卜旭辉,梁嘉琪,等.迭代学习控制在GTAW焊接过程中的应用研究[J].电子测量与仪器学报,2017,31(7):1061-1066.LIU J,BU X H,LIANG J Q,et al.Application of iterative learning control in control of GTAW process[J].Journal of Electronic Measurement and Instrument,2017,31(7):1061-1066.
[7]曹伟,丛望,孙明.初态学习下时滞非线性系统的迭代学习控制[J].仪器仪表学报,2012,33(2):315-320.CAO W,CONG W,SUN M.Iterative learning control with initial state study for nonlinear time-delay system[J].Chinese Journal of Scientific Instrument,2012,33(2):315-320.
[8]杜大军,费敏锐,宋杨,等.网络控制系统的简要回顾及展望[J].仪器仪表学报,2011,32(3):713-721.DU D J,FEI M R SONG Y,et al.Brief survey and prospect of networked control systems[J].ChineseJournal of Scientific Instrument,2011,32(3):713-721.
[9]FEI M R,DU D J,LI K.Two-layer networked learning control of a nonlinear HVAC system[J].LSMS,2007,4688:697-709.
[10]HAN X,LI J,XIU D.Error analysis for numerical formulation of particle filter[J].Discrete and Continuous Dynamical Systems,2015,20(5):1337-1354.
[11]REKLEITIS I,DUDEK G,MILIOS E.Probabilistic cooperative localization and mapping in practice[C].IEEE International Conference on Robotics and Automation,2003,2(2):1907-1912.
[12]LIU Y J,PENG CH W,HUANG L.Expert-based multimode switching control for the temperature and humidity control system of archive establishment[J].Intelligent Control and Automation,2002,2(2):858-862.
[13]YAO Y Y,HU Y T.Recognition and location of solar panels based on machine vision[J].Intelligent Robot Systems(ACIRS),2017:7-12.
[14]ASCHEMANN H,WACHE A,KRAEGENBRING O.A discrete-time inversion-based approach to iterative learning control of an overhead crane[C].Chinese Control Conference,2017:3419-3424.
[15]BHOUNSULE P A,YAMANE K.Iterative learning control for accurate task-space tracking with humanoid robots[C].IEEE International Conference on Humanoid Robots,2015:490-496.
[16]MULLER T,ENDISCH C.An estimation based iterative learning control approach for cross-coupled vehicle test systems[C].IEEE International Conference on Advanced Intelligent Mechatronics,2017:1375-1381.
[17]苑晶,李阳,董星亮,等.基于运动模式在线分类的移动机器人目标跟踪[J].仪器仪表学报,2017,38(3):568-577.YUAN J,LI Y,DONG X L,et al.Target tracking with a mobile robot based on on-line classification for motion patterns[J].Chinese Journal of Scientific Instrument,2017,38(2):568-577.
[18]ZHAN R,QIAN X,WAN J.Modified unscented particle filter for nonlinear Bayesian tracking[J].Journal of Systems Engineering and Electronics,2008,19(1):7-14.
[19]SUN CH Y,HOU J,ZHANG A H,et al.An intelligent particle filter for state estimation a default detection[J].Intelligent Control and Information Processing,2016:203-209.
[20]BELLETTI F,HAZIZA D,GOMES G,et al.Expert level control of ramp metering based on multi-task deep reinforcement learning[J].IEEE Transactions on Intelligent Transportation Systems,2017(99):1-10.
[2]秦武旻,朱长婕.虹膜快速检测与精确定位的算法研究[J].电子测量技术,2017,36(4):25-28.QIN W M,ZHU CH J.Efficient and precise optimization algorithm for iris location[J].Electronic Measurement Technology,2017,36(4):25-28.
[3]LI P AN,LEI D ZH,SUN X J,et al.An investigation and application of high-speed measurement and precise control based on machine vision[J].Systems,Science and Technology,2016,doi:10.5013/IJSSST.a.17.04.13.
[4]周文举,费敏锐,王海宽,等.基于迭代学习和卡尔曼滤波的相机精确抓拍控制方法[J].仪器仪表学报,2014,35(5):1044-1051.ZHOU W J,FEI M R,WANG H K,et al.Accurate camera image capture control based on iterative learning and Kalman filter[J].Chinese Journal of Scientific Instrument,2014,35(5):1044-1051.
[5]YANG B,LI H G,KANG ZH B,et al.Hamiltonianbased binocular visual seroing of camera-in-hand robotic systems[J].International Conference on Modelling,2012:388-393.
[6]刘建,卜旭辉,梁嘉琪,等.迭代学习控制在GTAW焊接过程中的应用研究[J].电子测量与仪器学报,2017,31(7):1061-1066.LIU J,BU X H,LIANG J Q,et al.Application of iterative learning control in control of GTAW process[J].Journal of Electronic Measurement and Instrument,2017,31(7):1061-1066.
[7]曹伟,丛望,孙明.初态学习下时滞非线性系统的迭代学习控制[J].仪器仪表学报,2012,33(2):315-320.CAO W,CONG W,SUN M.Iterative learning control with initial state study for nonlinear time-delay system[J].Chinese Journal of Scientific Instrument,2012,33(2):315-320.
[8]杜大军,费敏锐,宋杨,等.网络控制系统的简要回顾及展望[J].仪器仪表学报,2011,32(3):713-721.DU D J,FEI M R SONG Y,et al.Brief survey and prospect of networked control systems[J].ChineseJournal of Scientific Instrument,2011,32(3):713-721.
[9]FEI M R,DU D J,LI K.Two-layer networked learning control of a nonlinear HVAC system[J].LSMS,2007,4688:697-709.
[10]HAN X,LI J,XIU D.Error analysis for numerical formulation of particle filter[J].Discrete and Continuous Dynamical Systems,2015,20(5):1337-1354.
[11]REKLEITIS I,DUDEK G,MILIOS E.Probabilistic cooperative localization and mapping in practice[C].IEEE International Conference on Robotics and Automation,2003,2(2):1907-1912.
[12]LIU Y J,PENG CH W,HUANG L.Expert-based multimode switching control for the temperature and humidity control system of archive establishment[J].Intelligent Control and Automation,2002,2(2):858-862.
[13]YAO Y Y,HU Y T.Recognition and location of solar panels based on machine vision[J].Intelligent Robot Systems(ACIRS),2017:7-12.
[14]ASCHEMANN H,WACHE A,KRAEGENBRING O.A discrete-time inversion-based approach to iterative learning control of an overhead crane[C].Chinese Control Conference,2017:3419-3424.
[15]BHOUNSULE P A,YAMANE K.Iterative learning control for accurate task-space tracking with humanoid robots[C].IEEE International Conference on Humanoid Robots,2015:490-496.
[16]MULLER T,ENDISCH C.An estimation based iterative learning control approach for cross-coupled vehicle test systems[C].IEEE International Conference on Advanced Intelligent Mechatronics,2017:1375-1381.
[17]苑晶,李阳,董星亮,等.基于运动模式在线分类的移动机器人目标跟踪[J].仪器仪表学报,2017,38(3):568-577.YUAN J,LI Y,DONG X L,et al.Target tracking with a mobile robot based on on-line classification for motion patterns[J].Chinese Journal of Scientific Instrument,2017,38(2):568-577.
[18]ZHAN R,QIAN X,WAN J.Modified unscented particle filter for nonlinear Bayesian tracking[J].Journal of Systems Engineering and Electronics,2008,19(1):7-14.
[19]SUN CH Y,HOU J,ZHANG A H,et al.An intelligent particle filter for state estimation a default detection[J].Intelligent Control and Information Processing,2016:203-209.
[20]BELLETTI F,HAZIZA D,GOMES G,et al.Expert level control of ramp metering based on multi-task deep reinforcement learning[J].IEEE Transactions on Intelligent Transportation Systems,2017(99):1-10.