【作者】 周兴林；

【导师】 叶声华； 曲兴华；

【作者基本信息】 天津大学 ， 测试计量技术及仪器， 2007， 博士

【摘要】 大尺寸工件的三维截面几何形状及直线度的测量主要由应用经纬仪或者激光跟踪仪组成三维坐标测量系统,通过对表面点坐标的测量与评价完成横截面几何形状及轴线直线度等的测量。这些测量方法精度高但是测量的速度很慢,人工参与过多,实现起来较为繁琐。本论文基于点结构光视觉测量原理,研究一种大尺寸截面几何形状测量在线测量系统。系统组建后能快速检测大型工件的横截面几何形状、端面方位、轴线方位及同轴度,测量的精度高于千分之一。论文的主要研究内容及创新性成果有:1.深入分析和比较了大型工件的各种三维测量原理,介绍了国内外主要的大尺寸三维测量技术及设备,并对激光视觉测量技术在工业测量中的应用进行了总结概括;2.借助固定转轴建立了点结构光三维扫描测量的数学模型,推导了从计算机像面坐标到世界坐标之间的映射计算关系,实现传感器的旋转扫描测量。利用立体圆孔靶标实现了传感器局部参数标定,并结合双经纬仪空间坐标定位系统,完成对系统全局参数的标定;3.介绍各种分布式网络化测量技术和分布式网络测量的几个关键技术,基于CAN总线技术在工业网络测量中的成功应用,将CAN总线技术引入本系统中,实现分布式网络化测控信息在主计算机与传感器之间的传递;由此,给出了CAN总线网络控制软件的实现,完成了基于CAN总线的系统硬件实现框图;4.详细分析系统所采集的图像特点,运用椭圆拟合等图像处理方法完成对目标中心的准确提取;针对大尺寸工件横截面多为圆形或者椭圆形状的特点,比较了各种椭圆拟合算法;利用图像处理的各个横截面离散点,完成横截面椭圆中心拟合;运用直线度拟合方法,计算出待测件轴线方位。并将离散点投影到轴线的垂直平面上,再利用椭圆拟合最终完成横截面几何形状的评定。5.对传感器的结构进行了深入的分析,优化系统的结构参数;通过实验验证了本系统有着出色的性能和广阔的应用前景。更多还原

【Abstract】 Section geometric shape and straightness measurement for large-scale part is performed by 3D measurement system consisted of theodolites or laser trackers, which evaluate the cross section geometric shape and the straightness of axis by measuring the coordinates of many characteristic points. Although the high precision result can be obtained, these systems are hard, tiring and time-consuming to be used by operators. In this dissertation, a large-scale cross section geometric shape vision system based on point structured-light scanning measurement is developed to accurately evaluate the cross section geometric shape and the straightness of axis. The major contributions and innovations of the dissertation are as follows:1. Various large-scale part 3D measurement techniques are deeply analyzed and large-scale 3D measurement instruments at home and abroad are introduced in detail. Applications in industry measurement of laser vision measurement techniques are summarized as well, such as white car body vision system, on-line measurement of seamless steel pipe straightness by laser vision etc.2. The mathematic model for point structured-light scanning measurement fixed on a rotation axis is built when the rotation and translation which relates the world coordinate system to the camera coordinate system is computed. The sensor’s intrinsic unknown parameters are calibrated using the three-dimensional circularity drone and the measurement system parameter also is calibrated using a dual theodolites coordinate system.3. Various distributed network measurement techniques and system and relevant key techniques are introduced and analyzed. Because of the advantage in industry measurement, CAN field-bus technique is introduced into the large-scale cross section geometric shape vision system to transfer the signals between the computer and sensors. From this, the network control software of CAN field-bus and the hardware figure of system are shown in this paper.4. The laser point images are analyzed in detail and the center of light point is obtained by the method of such as ellipse fitting. All kinds of algorithm about ellipse fitting are compared to find out the good one for fitting the light point center into the section geometric shape. Straightness of axis is evaluated by using the line-fitting on the ellipse centers above computed. The section geometric shape can be acquired when the ellipse is projected to the plane perpendicular to the axis.5. The structure of sensor is analyzed in depth and the parameters are optimized excellently. The experiment is shown that the system has a good performance and deep potential for industry application.更多还原