基于MLS检测的一维绝对位移的误差分析与标定
|
摘要
基于最大长度序列(MLS)的原理设计了一种图像靶标,采用单目视觉和激光位移传感器开展了一维绝对位移的检测实验。在较为理想的环境下,系统重复定位检测数据的标准差小于1μm。针对实际应用环境和系统安装状况,从镜头失焦、照明不均匀、靶标的偏转及镜头倾斜等可能造成检测误差的情况,开展全面实验和分析,探讨了该一维绝对位移检测系统的重复定位,并提出了改进方法。实验结果表明,采用最大长度序列所设计的靶标在镜头失焦情况下,一维位移检测的重复定位测量数据的标准差小于1μm;通过改进算法使得照明不均匀测量数据的标准差小于1μm;靶标偏移角度不大于5°的情况下,重复定位测量数据的标准差不超过1μm;而镜头倾斜角不大于3°的情况下,重复定位测量数据的标准差小于1μm。最后,采用激光位移传感器对该一维绝对位移检测进行了系统标定,经标定,在以上环境和安装误差条件下,该系统测量数据的均方根误差限制在了2μm。测试实验和数据分析结果表明,该基于最大长度序列的原理的一维绝对位移检测系统,可实现高精度非接触精密位移检测,且具有系统结构简单、对实施工艺要求低和便于安装调试的优点。
An image target was designed based on the principle of maximum length sequence(MLS), using monocular vision and laser displacement sensor carried out One-dimensional absolute displacement test. In a relatively ideal laboratory environment, It's show the standard deviation of the system's repeated location detection data was less than 1 μm. According to the actual application environment and system installation status, for loss of focus from the lens, uneven illumination, deflection of the target and tilt of the lens may cause the detection error, conduct comprehensive experiments and analysis, The repetitive positioning of the one-dimensional absolute displacement detection system is discussed and proposed the method of improved. The experimental results show the standard deviation of the repeated positioning measurement data detected by one-dimensional displacement is less than 1 μm when the lens is out of focus; That the standard deviation of the non-uniform measurement data of lighting was less than 1 μm by improving the algorithm; When the targets under the condition of not greater than 5°, The standard deviation of the repeated positioning measurements is no more than 1 μm; And the camera angle is not greater than 3°, the standard deviation of the repeated positioning measurements is less than 1 μm. The last, the one-dimensional absolute displacement measurement is systematically calibrated by using laser displacement sensor. Under the above environment and installation error conditions, the root mean square error of the system measurement data is limited to 2 μm after the calibration.Test experiment and data analysis show the one-dimensional absolute displacement detection system is based on the principle of maximum length sequence can achieve high precision non-contact precision displacement detection, the system structure is simple and low requirement to implement process, advantages of easy installation and debugging.
An image target was designed based on the principle of maximum length sequence(MLS), using monocular vision and laser displacement sensor carried out One-dimensional absolute displacement test. In a relatively ideal laboratory environment, It's show the standard deviation of the system's repeated location detection data was less than 1 μm. According to the actual application environment and system installation status, for loss of focus from the lens, uneven illumination, deflection of the target and tilt of the lens may cause the detection error, conduct comprehensive experiments and analysis, The repetitive positioning of the one-dimensional absolute displacement detection system is discussed and proposed the method of improved. The experimental results show the standard deviation of the repeated positioning measurement data detected by one-dimensional displacement is less than 1 μm when the lens is out of focus; That the standard deviation of the non-uniform measurement data of lighting was less than 1 μm by improving the algorithm; When the targets under the condition of not greater than 5°, The standard deviation of the repeated positioning measurements is no more than 1 μm; And the camera angle is not greater than 3°, the standard deviation of the repeated positioning measurements is less than 1 μm. The last, the one-dimensional absolute displacement measurement is systematically calibrated by using laser displacement sensor. Under the above environment and installation error conditions, the root mean square error of the system measurement data is limited to 2 μm after the calibration.Test experiment and data analysis show the one-dimensional absolute displacement detection system is based on the principle of maximum length sequence can achieve high precision non-contact precision displacement detection, the system structure is simple and low requirement to implement process, advantages of easy installation and debugging.
引文
[1] 郭瑞峰,袁超峰,杨柳,等.基于OpenCV的机器视觉尺寸测量研究[J].计算机工程与应用,2017,53(9):253-257.
[2] 许楠楠,刘缠牢.光学元件外径的机器视觉测量技术研究[J].光学仪器,2016,38(4):292-296.
[3] 龚阶.基于机器视觉的卡簧测量算法研究[J].中国计量,2016(11):98-101.
[4] 刘昶,韦飞云,孙维广.基于双平行平面相机模型的钢板尺寸视觉测量[J].光学精密工程,2016,24(4):714-725.
[5] 胥磊.机器视觉技术的发展现状与展望[J].设备管理与维修,2016(9):7-9.
[6] 黄松梅,毕远伟,刘殿通,等.双目立体视觉非接触式测量研究[J].烟台大学学报(自然科学与工程版),2017,30(4):323-327.
[7] 关瑞芬,杨凌辉,王丽君,等.基于正交柱面成像的空间物体位姿精密测量[J].光学学报,2016,36(11):126-134.
[8] 徐刚,杨世模,龚雨兵.大型光学望远镜副镜位姿精调机构的优化设计[J].光学精密工程,2008(7):1181-1189.
[9] 李强,杨德华,费飞,等.基于最大长度序列的绝对位移精密检测方法[J].计算机测量与控制,2017,25(8):32-35.
[10] 杨德华,费飞,李开宇,等.一种基于视觉的位置检测编码靶标及系统:201510201768.X[P].2015-04-24.
[11] 袁小翠,吴禄慎,陈华伟.基于Otsu方法的钢轨图像分割[J].光学精密工程,2016,24(7):1772-1781.
[12] 余慧杰,韩平畴.双目视觉系统的测量误差分析[J].光学技术,2007(S1):157-159.
[13] 王兴龙,周志成,曲广吉.视觉测量误差对空间机械臂捕获目标卫星控制精度的影响分析[J].航天器工程,2017,26(3):31-37.
[14] 霍炬,崔家山,王伟兴.基于共面特征点的单目视觉位姿测量误差分析[J].光子学报,2014,43(5):150-156.
[15] 邓辉,谢俊,孟广月,等.基于机器视觉的重复定位精度测量技术[J].电子测量技术,2014,37(12):45-48.
[16] 吕虹,段颖妮,管必聪.一种非线性最大长度伪随机序列发生器的设计[J].电子器件,2008(3):898-900.
[17] 郭皓然,邵伟,周阿维,等.全局阈值自适应的高亮金属表面缺陷识别新方法[J].仪器仪表学报,2017,38(11):2797-2804.
[2] 许楠楠,刘缠牢.光学元件外径的机器视觉测量技术研究[J].光学仪器,2016,38(4):292-296.
[3] 龚阶.基于机器视觉的卡簧测量算法研究[J].中国计量,2016(11):98-101.
[4] 刘昶,韦飞云,孙维广.基于双平行平面相机模型的钢板尺寸视觉测量[J].光学精密工程,2016,24(4):714-725.
[5] 胥磊.机器视觉技术的发展现状与展望[J].设备管理与维修,2016(9):7-9.
[6] 黄松梅,毕远伟,刘殿通,等.双目立体视觉非接触式测量研究[J].烟台大学学报(自然科学与工程版),2017,30(4):323-327.
[7] 关瑞芬,杨凌辉,王丽君,等.基于正交柱面成像的空间物体位姿精密测量[J].光学学报,2016,36(11):126-134.
[8] 徐刚,杨世模,龚雨兵.大型光学望远镜副镜位姿精调机构的优化设计[J].光学精密工程,2008(7):1181-1189.
[9] 李强,杨德华,费飞,等.基于最大长度序列的绝对位移精密检测方法[J].计算机测量与控制,2017,25(8):32-35.
[10] 杨德华,费飞,李开宇,等.一种基于视觉的位置检测编码靶标及系统:201510201768.X[P].2015-04-24.
[11] 袁小翠,吴禄慎,陈华伟.基于Otsu方法的钢轨图像分割[J].光学精密工程,2016,24(7):1772-1781.
[12] 余慧杰,韩平畴.双目视觉系统的测量误差分析[J].光学技术,2007(S1):157-159.
[13] 王兴龙,周志成,曲广吉.视觉测量误差对空间机械臂捕获目标卫星控制精度的影响分析[J].航天器工程,2017,26(3):31-37.
[14] 霍炬,崔家山,王伟兴.基于共面特征点的单目视觉位姿测量误差分析[J].光子学报,2014,43(5):150-156.
[15] 邓辉,谢俊,孟广月,等.基于机器视觉的重复定位精度测量技术[J].电子测量技术,2014,37(12):45-48.
[16] 吕虹,段颖妮,管必聪.一种非线性最大长度伪随机序列发生器的设计[J].电子器件,2008(3):898-900.
[17] 郭皓然,邵伟,周阿维,等.全局阈值自适应的高亮金属表面缺陷识别新方法[J].仪器仪表学报,2017,38(11):2797-2804.