基于机器视觉的物体阶跃表面检测算法研究与实现
|
摘要
在光学精密测量中,机器视觉应用广泛,其中算法的选择会直接影响测量的精度。针对具有阶跃表面的检测物体,采用变频相移干涉测量的方法。在变频相移干涉测量中,相位提取的精度会直接影响测量的精度,常用相位提取算法只能测量简单的光滑表面,测量具有阶跃表面的复杂形面时容易引入相移误差,影响测量的精度。运用一种基于迭代的相位提取算法,通过交替迭代法求解出被测相位。在仿真和实验中,选用既包括光滑曲面和阶跃表面的被测面形,对传统的二步相移算法、四步相移算法和基于迭代的相位提取算法的相位提取精度进行了比较。结果表明,基于迭代的相位提取算法测量误差最小,并且在复杂阶跃表面的测量中其精度优于传统的相位提取算法。
In optical precision measurement,machine vision is widely used,and the choice of algorithm will directly affect the accuracy of measurement. For the detection object with step surface,the precision of phase extraction will directly affect the measurement accuracy in the frequency-conversion phase-shift interferometry. However,the commonly used phase extraction algorithm can only measure simple smooth surface,and it is easy to introduce phase shift error when measuring complex surface with step surface. Error,which produces measurement error,affects the accuracy of measurement. An iterative phase-finding algorithm is used,which does not need to calibrate the phase of the test system. The phase can be obtained by alternating iteration method when the measured phase and phase shift are unknown. In the simulation and experiment,the measured surface including smooth surface and step surface is selected,and the phase extraction accuracy of traditional iterative algorithm,four-step phase-shifting algorithm and iterative phase-solving algorithm are compared. The results show that the iterative phase extraction algorithm has the smallest measurement error,and its accuracy is better than the traditional phase extraction algorithm in the measurement of complex step surface.
In optical precision measurement,machine vision is widely used,and the choice of algorithm will directly affect the accuracy of measurement. For the detection object with step surface,the precision of phase extraction will directly affect the measurement accuracy in the frequency-conversion phase-shift interferometry. However,the commonly used phase extraction algorithm can only measure simple smooth surface,and it is easy to introduce phase shift error when measuring complex surface with step surface. Error,which produces measurement error,affects the accuracy of measurement. An iterative phase-finding algorithm is used,which does not need to calibrate the phase of the test system. The phase can be obtained by alternating iteration method when the measured phase and phase shift are unknown. In the simulation and experiment,the measured surface including smooth surface and step surface is selected,and the phase extraction accuracy of traditional iterative algorithm,four-step phase-shifting algorithm and iterative phase-solving algorithm are compared. The results show that the iterative phase extraction algorithm has the smallest measurement error,and its accuracy is better than the traditional phase extraction algorithm in the measurement of complex step surface.
引文
[1]范金坪,徐小飞,吕晓旭,等.基于最小二乘迭代的多波长同时相移干涉测量方法[J].中国激光,2016,43(03):217-226.
[2]田爱玲,马迪,王大森,等.波长调谐干涉测量的相移标定方法[J].光子学报,2016,45(07):156-160.
[3]吴飞斌,唐锋,王向朝,等.Ronchi相移剪切干涉仪及其相位提取误差分析[J].光学学报,2015,35(06):142-152.
[4] REENA D,TAKAMASA S. Simultaneous measurement of thickness and refractive index using phase shifted Coherent Gradient Sensor[J]. Optics and Laser Technology,2016,86.
[5]张晨皓,万新军,杨波,等.基于选择采样的高效迭代相位提取算法[J].应用光学,2016,37(01):64-68.
[6] SHUHEI Y. Measurement of moving objects with phaseshifting digital holography using liquid crystal retarder[J].Optics Communications,2018,420.
[7]徐媛媛,时刻,王亚伟,等.相移干涉技术及相移获取方法研究进展[J].激光与光电子学进展,2018,55(02):36-49.
[8]高芬,倪晋平,李兵.多步相移中被测件径向相移不均匀引入误差比较及校正[J].光学学报,2018,38(04):215-221.
[9]刘锋伟.随机相移干涉测量关键技术研究[D].中国科学院光电技术研究所,2017.
[10] MAO J. Target detection and recognition algorithm for moving UAV based on machine vision[J]. Cluster Computing,2018:1-7.
[11] HANAYAMA R,ISHII K. 1P1-K05 DC Servo Motor Driven Phase Shifting Equipment for a Phase Shift Shearing Interferometer(Future of Optronics)[J]. The Proceedings of JSME annual Conference on Robotics and Mechatronics,2012,2012(0).
[12]刘剑,田爱玲,刘丙才,等.一种变频相移干涉测量的相位提取算法[J].光学学报,2014,34(03):120-124.
[13]郭振黎,邢飞,娄小平,等.基于机器视觉的启动器外观缺陷检测[J].电子测量与仪器学报,2018(2).
[14]陈广锋,管观洋,魏鑫.基于机器视觉的冲压件表面缺陷在线检测研究[J].激光与光电子学进展,2018(1):335-341.
[15]徐蕾,彭月平,贺科宁.基于改进双边滤波与随机共振的图像去噪算法研究[J].激光杂志,2018,39(08):142-148.
[16] RAJU R S U,RAMESH R,RAJU V R. Curvelet transforms and flower pollination algorithm based machine vision system for roughness estimation[J]. Journal of Optics,2018(1-2):1-8.
[17]李超,孙俊.基于机器视觉方法的焊缝缺陷检测及分类算法,2017年3月6日[J].计算机工程与应用,2018(6).
[2]田爱玲,马迪,王大森,等.波长调谐干涉测量的相移标定方法[J].光子学报,2016,45(07):156-160.
[3]吴飞斌,唐锋,王向朝,等.Ronchi相移剪切干涉仪及其相位提取误差分析[J].光学学报,2015,35(06):142-152.
[4] REENA D,TAKAMASA S. Simultaneous measurement of thickness and refractive index using phase shifted Coherent Gradient Sensor[J]. Optics and Laser Technology,2016,86.
[5]张晨皓,万新军,杨波,等.基于选择采样的高效迭代相位提取算法[J].应用光学,2016,37(01):64-68.
[6] SHUHEI Y. Measurement of moving objects with phaseshifting digital holography using liquid crystal retarder[J].Optics Communications,2018,420.
[7]徐媛媛,时刻,王亚伟,等.相移干涉技术及相移获取方法研究进展[J].激光与光电子学进展,2018,55(02):36-49.
[8]高芬,倪晋平,李兵.多步相移中被测件径向相移不均匀引入误差比较及校正[J].光学学报,2018,38(04):215-221.
[9]刘锋伟.随机相移干涉测量关键技术研究[D].中国科学院光电技术研究所,2017.
[10] MAO J. Target detection and recognition algorithm for moving UAV based on machine vision[J]. Cluster Computing,2018:1-7.
[11] HANAYAMA R,ISHII K. 1P1-K05 DC Servo Motor Driven Phase Shifting Equipment for a Phase Shift Shearing Interferometer(Future of Optronics)[J]. The Proceedings of JSME annual Conference on Robotics and Mechatronics,2012,2012(0).
[12]刘剑,田爱玲,刘丙才,等.一种变频相移干涉测量的相位提取算法[J].光学学报,2014,34(03):120-124.
[13]郭振黎,邢飞,娄小平,等.基于机器视觉的启动器外观缺陷检测[J].电子测量与仪器学报,2018(2).
[14]陈广锋,管观洋,魏鑫.基于机器视觉的冲压件表面缺陷在线检测研究[J].激光与光电子学进展,2018(1):335-341.
[15]徐蕾,彭月平,贺科宁.基于改进双边滤波与随机共振的图像去噪算法研究[J].激光杂志,2018,39(08):142-148.
[16] RAJU R S U,RAMESH R,RAJU V R. Curvelet transforms and flower pollination algorithm based machine vision system for roughness estimation[J]. Journal of Optics,2018(1-2):1-8.
[17]李超,孙俊.基于机器视觉方法的焊缝缺陷检测及分类算法,2017年3月6日[J].计算机工程与应用,2018(6).