摘要
针对孔径自动化检测的问题,提出了一种基于Roberts改进的多梯度方向插值孔径检测方法。通过归一化模板匹配将输入图像拆分成圆孔ROI图像组,分别对每个ROI进行图像增强处理,改进了Roberts算子并进行边缘特征粗定位,通过多梯度方向插值的亚像素算法在边缘粗定位的基础上进一步精细定位,利用最小二乘法进行圆拟合,得到圆孔曲线和半径,搭建视觉检测实验平台并进行实验,将实验结果与三坐标测量结果进行对比验证。实验表明,该方法的检测精度为10μm,重复测量最大偏差为10μm,实现了对孔径的高精度、高效率、高稳定性检测。
Aiming at the problem of automatic detection of aperture,a multi-gradient direction interpolation hole detection method based on Roberts improvement is proposed.The input image is split into a group of aperture ROI images by normalized template matching,image enhancement is performed on each ROI respectively,Roberts operator is improved and edge features are coarsely located,the sub-pixel algorithm of multi-gradient direction interpolation is further finely located on the basis of edge coarsely located,the circle fitting is performed by using the least squares method to obtain the curve and radius of the aperture,and the vision detection experimental platform is built and experimented,and the experimental results are compared and verified with the CMM measurement results.The experiments show that the detection accuracy of the method is 10μm,and the maximum deviation of repeated measurements is 10μm,which realizes high-precision,high-efficiency and high-stability detection of the aperture.
作者
刘方吉
茅健
赵嫚
LIU Fangji;MAO Jian;ZHAO Man(School of Mechanical and Automotive Engineering,Shanghai University of Engineering Science,Shanghai 201620,China;Sichuan Research Institute,Shanghai Jiaotong University,Chengdu 610213,China)
出处
《组合机床与自动化加工技术》
北大核心
2025年第5期110-114,共5页
Modular Machine Tool & Automatic Manufacturing Technique
基金
四川省自然科学基金项目(2023NSFSC0366)。
