基于机器视觉的衔铁组件质量检测系统设计

Design of Vision-based Armature Assembly Quality Detection System

针对衔铁组件几何参数人工测量精度低、一致性差的问题,为进一步改善工件的品质管控,设计一种基于机器视觉的衔铁组件品质检测系统。采用2套视觉系统对工件正面、侧面分别成像用于不同尺寸参数的测量。在A尺寸测量中,通过特征定位点建立相对坐标系,从原始图像中截取包括小轴凸台和参考边感兴趣区域并对该区域图像进行预处理,在ROI区域内采用Canny算法进行凸台边缘检测以降低粉尘等带来的边缘检测干扰。针对工件厚度待测区域远超视觉系统的景深,通过在夹爪上布置参考块间接测量厚度值。采用不同批次的工件用于试验,试验结果显示品质管控准确率达100%,与图像尺寸测量仪的误差不超过5μm,可以满足工件品质管控的应用需求。

Aiming at the problems of low accuracy and poor consistency of manual measurement of geometric parameters of armature components, a quality detection system of armature assembly based on machine vision was designed to improve the quality control of workpiece.Two sets of vision systems were used to image the front and side of the workpiece respectively for the different dimensional parameters measurement.In the A dimension measurement, the relative coordinate system was established through the feature positioning points, the region of interest including small axis boss and reference edge was intercepted from the original image, and the image of this region was preprocessed.In the ROI region, Canny algorithm was used for boss edge detection to reduce the edge detection interference caused by dust.As the thickness of the workpiece to be measured is far beyond the depth of field of the vision system, indirect thickness measurement by arranging reference blocks on the gripper jaws was adopted.Different batches of workpieces were used for the test.The test results show that the quality control accuracy rate is 100%,and the error with the image size measuring instrument does not exceed 5 μm, which can meet the application requirements of workpiece quality control.