基于机器视觉的机床冲孔检测方法研究

Research on Machine Tool Punching Hole Detection Method Based on Machine Vision

结合图像采集、图像预处理和摄像机标定等机器视觉技术以及真圆度算法,研究出一种基于图像处理的形态分析法和一种结合了最小二乘法拟圆的边缘拟合法,对冲孔机床上加工的制孔大小、个数和位置不固定的铁塔构件进行检测,并评价两者的检测结果和实用性。其中边缘拟合法的平均半径检测精度可达0.395mm,位置检测精度可达0.399mm,可同时满足检测构件制孔半径大小和中心位置两要素的要求,实现产品的自动化在线检测。并通过软件编程和辅助算法代替以往的图像拼接匹配方法和电机驱动控制板块,自动对进入钢联生产线传送区域的铁塔构件采集图像,为搭建机床冲孔视觉检测系统提供技术支持。

By combining true roundness algorithm and machine vision technology witch includes image acquisition,image preprocessing and camera calibration,developing a morphological analysis method based on image processing and an edge fitting method based on least square method(LSM),monitoring the hole making of iron-tower components processed on punching machine tool whose size,number and position are not fixed,and evaluates the monitoring results and practicability of the two methods.The average radius accuracy of edge fitting method could be reached at 0.395 mm,and the accuracy of position detection could be reached at 0.399 mm,which could meet the requirements of both the radius of hole making and the center position of the inspection component,realize automatic monitoring of products online and replace the previous image stitching matching method and motor driving control block by software programming and auxiliary algorithm.Automatically collect images of tower components entering steel joint production line conveying area,and provide technical support for building visual inspection system of machine tool punching.