基于机器视觉的圆形零件尺寸参数测量

Dimension Measurement of Circular Parts Based on Machine Vision

针对圆形零件尺寸传统测量方法存在测量效率低、一致性差及同心度参数不易测量的问题,设计一种基于机器视觉的圆形零件特征参数测量系统。采用阈值分割法对灰度图像进行阈值分割以提取特征目标,利用数学形态学方法对二值图像进行腐蚀、膨胀操作,避免纤细、重叠的噪声干扰;通过最小二乘法对内、外圆弧轮廓点进行拟合,得到圆心和半径参数,通过欧式距离计算出同心度参数。实验测试显示,系统精度达到0.01mm,与采用测量仪相比,视觉测量方法更适合大批量非接触式测量。

Aiming at the problems of low measurement efficiency,poor consistency and difficulty measurement of concentricity parameter in the traditional measurement methods of circular parts,a dimension measurement system of circular parts based on machine vision is designed.The gray image is segmented by threshold segmentation method to extract feature objects,and the mathematical morphology method is used to avoid the interference of thin and overlapping noise by corrode and expand the binary image.The center and radius parameters of the inner and outer arc contour points are fitted by the least square method,and the concentricity parameter is calculated by Euclidean distance.The experimental results show that the accuracy of the system reaches 0.01mm.Compared with the projector,the visual measurement method is more suitable for large-scale,non-contact measurement and meets the actual needs of industrial applications.