基于机器视觉螺栓类零件参数在线质检系统

Online Quality Inspection System for Bolt Parts Parameters Based on Machine Vision

针对传统测量螺栓类零件各几何参数存在精度低、分拣低效等问题,开发了一种基于机器视觉的系统用于提高螺栓类零件多几何参数在线测量和分拣的效率。该系统以Halcon联合VisualStudio2017的开发环境,为了完整提取螺栓类零件的轮廓图像,图像采集系统采用背光源的方式,使用Blob分析图像处理、Shape-Matching模板匹配定位、缩小检索区域、亚像素拟合轮廓边缘等方法。运用边缘轮廓对灰度值梯度求导的算法,通过设置阈值剔除干扰点,对符合要求的测量点进行排序,创建边缘对,实现实时测量螺栓类零件的各几何参数,且能够快速的分拣出螺栓类零件的合格件等功能。实验结果表明:该系统可以达到工业高精度测量要求且具有良好的稳定性、时效性和高效性,测量重复性精度达到0.005 mm。

Aiming at the problems of low accuracy and inefficient sorting in the traditional measurement of geometric parameters of bolt parts, a system based on machine vision is developed to improve the efficiency of online measurement and sorting of multiple geometric parameters of bolt parts. The system uses the development environment of Halcon and Visual Studio 2017. In order to completely extract contour images of bolt parts, the image acquisition system adopts a backlight source, and uses the methods of Blob analysis image processing, shape-matching template matching positioning, narrowing the search area, sub-pixel simulation contouring edges and other methods. By the algorithm of using the edge contour to derive the gradient of gray value, removing the interference points by setting the threshold, sorting the measurement points that meet the requirements and creating edge pairs, real-time measurement of geometric parameters of bolt parts, quickly sorting ualified parts of bolt parts and other functions are realized. The experimental results show that the proposed system can meet the requirements of industrial high-precision measurement, and has good stability, timelinessand high efficiency. The measurement repeatability accuracy reaches 0.005 mm.