基于机器视觉的零件同心度检测系统的设计

摘要

With the development of the advanced manufacturing technology,the technology of the part detection plays an important role in the field of industrial production.The concentricity of parts was taken as the research object,and the detection platform was built and then the new method for measuring the concentricity was proposed.Firstly,gray-scale images were segmented dynamically by Otsu method.Secondly,the edges of the outer circle and the inner circle were extracted using the method of binary morphology,such as filling treatment,XOR processing,and edge detection.Thirdly,the edges of the outer circle and the inner circle were fitted by iterative least square method based on the outlier detection.Finally,the value of the concentricity was calculated.Compared with the three-coordinate measuring machine,the proposed method is more suitable for large quantities and non-contact detection and its detection error is below 0.01mm.Experiments show that this method can effectively realize the detection of the concentricity of the parts and meet the actual needs of the enterprise.%随着先进制造技术的发展,零件检测技术在工业生产领域中占有重要地位.以零件的同心度参数为研究对象,搭建了同心度的检测平台并提出一种新的同心度检测方法.该方法首先采用Otsu法对灰度图像进行动态阈值分割,其次利用二值形态学中的填充处理、异或处理及边缘提取等运算获取外圆、内圆的边缘,然后使用基于外点剔除的迭代最小二乘法进行外圆、内圆的边缘拟合,最后计算零件的同心度值.与采用接触式检测的三坐标测量机相比,提出的检测方法更适合于大批量、非接触式零件同心度的检测,且检测误差在0.01mm以下.试验表明,该方法可有效地实现零件的同心度检测,满足当前企业的实际需求.