基于机器视觉的刀具轮廓在线检测方法的研究

超精密单点金刚石车削过程中需要对刀具参数进行检测。基于机器视觉原理,采用新型的机械结构设计了新型刀具在线检测系统。通过垂直放置的两组由同轴远心镜头和CMOS相机组成的光学系统来获取车刀刀尖正面及侧面的图像,再经图像处理系统来获取刀具的轮廓及位置信息,完成对刀具轮廓及位置的在线检测,实验验证可实现测量的重复性定位精度达1μm。在线检测的方式借助了超精密车床自身的高精度和运动机构,可以保证相机焦点和机床主轴之间的相对位置关系,弥补了目前刀具检测系统稳定性差和重复性精度低的不足,提高了加工的整体效率。

Research on on-line detection method for tool profile base on machine vision

Single point diamond turning has become an important method in production of high-precision optical devices and high-precision molds. A new on-line tool detection system is designed by using a new mechanical structure that based on the principle of machine vision. The hardware of the system is mainly composed of a telecentric coaxial microscope, a CMOS camera, a light source, telecentric coaxial lens, CMOS camera, computer and circular chuck. Cylindrical chuck with diameter of 120 mm is used in the system. When measuring, it is directly mounted on the spindle sucker of ultra-precision lathe. The control system of machine tool is used to realize the movement of tool stand and lathe spindle. Two sets of image acquisition systems which are perpendicular to each other are used to obtains the front and side images of turning tool tips. In the design software part, after preprocessing and contour fitting of the collected image, the cutter contour and position information are obtained. From then on, the on-line detection of tool profile and position is completed, and the repetitive positioning accuracy of the measurement can reach 1 um. With the help of the high precision and motion mechanism of the ultra-precision lathe, this measurement method makes up for the deficiencies of the poor stability and low repeatability accuracy of the current tool detection system. It avoids tool wear in previous measurement, shortens tool setting time, and improves the overall efficiency of machining.