节点文献

无线组网式大尺寸光笔测量系统关键技术的研究

Research on Key Technologies of Wireless Network Large-scale Measurement System with a Light-pen

【作者】 张瑞

【导师】 刘书桂;

【作者基本信息】 天津大学 , 仪器科学与技术, 2017, 博士

【摘要】 便携性好、测量方便灵活、可以测量光学等方法无法探测的隐藏点、精度较高、能够实现十几米范围内的工业现场测量等优点,使得光笔式视觉坐标测量系统一直被国内外相关机构所高度关注。然而国内外相关研究工作的重点是如何提高光笔系统的测量精度与重复性,却很少关注如何突破光笔系统现有的测量范围,实现上百米的大尺寸测量这块领域。在对相关研究成果深入全面分析的基础上,本文从光笔结构的改进、控制点图像中心定位算法、探头中心标定算法、系统结构的优化设计等方面进行了相关的研究,以提高光笔系统的测量精度、测量重复性以及测量的范围。主要内容如下:1.基于直接线性变换模型,对第二代光笔的结构进行了改进。巧妙地利用三坐标测量机主轴移动的高精度、高重复性,构造了不同的光笔前六个控制点所在面与后四个控制点所在面距离的虚拟光笔,并通过虚拟光笔重复性实验与精度对比实验证明了改变此距离参数能够改善光笔系统的轴向测量重复性以及系统的测量精度。2.对第二代光笔特征点中心定位精度较差的原因进行了深入分析,认为特征点椭圆中心畸变误差影响了其中心定位精度。本文提出了一种基于交比不变性的特征点椭圆中心畸变误差求解模型,并通过同心圆实验证明了此项误差确实为第二代光笔特征点中心定位精度不理想的根本原因。3.探头中心在光笔坐标系中的定位精度直接影响了光笔系统的测量精度及重复性,本文提出了一种基于不变距离的光笔探头中心自标定算法,通过不同相机与光笔距离下的探头中心自标定重复性实验,证明了新的探头中心标定算法能够满足亚毫米级精度系统的测量要求。4.针对现有光笔式视觉坐标测量系统的不足之处,本文提出了一种新型的无线组网式大尺寸光笔测量系统结构。设计了集采集处理图像一体化的智能相机系统与无线通讯组网系统,并进行了实验验证。基于Zig Bee无线网络,可将计算机、可调光笔与智能相机组网为一个测量单元,而若干个测量单元可构成一个测量网络,从而实现上百米大尺寸目标的现场测量。

【Abstract】 The light-pen vision coordinate measurement system has always been concerned for recent decades,for that the system could realize the hiding spots in-situ measurement,which can hardly be done by other portable measurement systems.The light-pen system’s measurement accuracy and repeatability have been paid too much attention,however,its measurement range,by contrast,has been nearly ignored.In order to improve the measurement accuracy,repeatability and range of the light-pen system,this paper has concerned some technologies,such as the improvement of the light-pen structure,the feature point center positioning,the probe tip center positioning,the optimal design of the system structure and so on,on the basis of a deep analysis about related researches at home and abroad.The main work in this paper is as follows:1.Improve the the second-generation light-pen structure based on the direct linear transformation mathematical model.Several virtual light-pens,with different distances between the front plane where six control points lie and the back plane where other four ones lie,have been built skillfully by high accurate and repeatable movement of the coordinatge measuring machine’s main axis.And the results of the virtual light-pen repeatability experiment and the accuracy contrast experiment show that this distance parameter has a large impact on the system’s measurement repeatability along the optical axis and the whole system’s measurement accuracy.2.It is thought that the ellipse center distortion error of feature points influence the accuracy of their center positioning,after the deep analysis of the reason why the accuracy is so poor.A new model based on the cross-ratio invariance is given to solve the feature points’ ellipse center distortion error.And it is concluded that this error is what causes the unsatisfactory center positioning accuracy of the second-generation light-pen’s feature points by the concentric circle test.3.The accuracy of the light-pen’s probe tip center positioning has a direct influence on the whole system’s measurement accuracy and repeatability.Based on invariable distances,a new self-calibration algorithm of the probe tip center positioning is given.And the result of the probe tip center self-calibration repeatability test,in different distances between the camera and the light-pen,shows that the given algorithm could meet the requirement of the system with sub-millmeter accuracy.4.Aim at the deficiencies of existing light-pen vision coordianate measuring machines,a new structure of the wireless network large-scale measurement system with a light-pen is presented.A smart camera,which can not only collect the light-pen’s photos,but also process them,and a set of wireless network communication system are designed and verified by several tests.By the flexible Zig Bee wireless network,a computer,an adjustable light-pen and one or two smart cameras could make up a measurement unit.Furthermore,several measurement units can form a huge measurement network that could realize the in-situ measurement of large-scale objects with high accuracy.

  • 【网络出版投稿人】 天津大学
  • 【网络出版年期】2018年 08期
节点文献中: 

本文链接的文献网络图示:

本文的引文网络