【作者】 尹辉；

【导师】 左建中；

【作者基本信息】 天津大学 ， 机械设计及理论， 2007， 硕士

【摘要】 螺纹连接件连接可靠,装配方便,被广泛应用在机械设备以及其他设备中。在工程应用中,对螺纹的互换性和可靠性要求很高,因此需要进行认真的检测。目前我国主要使用机械检测方法对螺纹进行检测,检测速度慢、检测精度低,已不能满足生产实际的需要。随着计算机技术和光电技术的发展,已经出现了一种新的检测技术——基于计算机视觉的检测技术。论文分析了计算机视觉检测系统的基本构成、特点及优越性,依据视觉检测系统硬件选用的一般原则,建立了计算机视觉检测实验系统。针对在图像采集过程中影响图像质量的干扰因素,并通过对滤波算子的实验研究,采用了中值滤波算法对图像进行了预处理。对一阶、二阶微分边缘检测算子进行了讨论研究。针对测量要求,提出了多种螺纹边缘轮廓的提取算法,在对各种算法进行具体分析的基础上,采用基于灰度直方图的阈值计算法来提取螺纹边缘轮廓。根据螺纹边缘特点提出了用于直线边缘的亚像素定位方法——最小二乘回归直线拟合法,该算法计算速度快,定位精度高,适应于直线边缘的整体定位。针对二维图像测量实际应用的特点,采用了不同方位多次标定取平均值的标定方法,可消除镜头畸变误差和标定过程引入的误差。根据螺纹检测要求,选择螺纹大径、螺距、牙型角和中径四个主要参数作为计算机视觉测量目标。以六角头螺栓为样件进行了检测实验,结果表明,能够满足工业检测的精度要求。更多还原

【Abstract】 The screw connection has been widely used in mechanical equipment and other devices because of its high reliability and convenient assembly characters. The interchangeability and reliability of the screw are required strictly in engineering application, so it must be measured seriously. Mechanical measurement is incapable of satisfying the actual demand of production because of its low measuring speed and low precision. With the development of computer technology, a new inspection technology based on computer visual has been presented.The paper analyzes the visual inspection system’s basic composing, characteristics and superiority; presents main principle of the system hardware selection; and establishes computer visual inspection experimental system. Focused on the interfere factors that affect image quality in the process of image collection, and through the experiments comparison we takes mean filtering operator to preprocess the image. First-order differential edge detection operator, second-order differential edge detection operators are studied and discussed. In the base of the character of screw thread and experiments comparison of the operators, many operators are presented, and gray-histogram threshold method is used to detect edge. In the base of the character of screw thread, a sub-pixel orientation method used to line edge—least-squares fitted regression line is proposed, which has faster speed of running and better orientation precision and is adaptive to line edge’s whole orientation. According to practical application character of 2-D image measurement, a calibration method which takes the average of the calibration value from different orientation is used. It can eliminate the aberration and error in calibration at one time. According to measure request of screw thread, four main characters-major diameter, screw pitch, angle of thread and simple pitch diameter are chosen to be measuring objects of computer vision. The experiment takes hex bolt as a sample, and the result shows that it can reach the precision requirement of industry measurement.更多还原