基于机器视觉的送纱管同轴度检测研究

Research on Coaxiality Detection of Bobbin Based on Machine Vision

对目前人工检测送纱管同轴度效率低、可靠性差等问题进行了研究,在分析了送纱管的结构及检测要求后,提出一种利用机器视觉技术对送纱管的同轴度进行高效检测的方法。首先使用自主搭建的实验平台获取送纱管内孔边缘图像,然后对图像进行预处理、并采用改进Canny算法得到送纱管的边缘特征,再对边缘应用RANSAC结合最小二乘法算法进行拟合及测量,得到两边缘圆心的距离,从而实现对送纱管同轴度的快速准确检测。实验结果表明,该方法检测的精度可以达到0.01mm,能够满足送纱管同轴度的测量精度要求。

In view of the problems of low speed and low efficiency in manual detection of coaxiality of feeding tube.Machine vision is used to detect the coaxiality of the feeding tube.Firstly,the image is acquired by using the self-made experimental platform.After preprocessing the image,improving Canny algorithm and RANSAC combined with least square algorithm,the edge features of the feeding tube are obtained,so as to realize the rapid detection of the coaxiality of the feeding tube.The experimental results show that the detection accuracy of the method can reach 0.01mm,which can meet the requirements of the measurement accuracy of the feeding tube.