基于线结构光的钢领内表面圆度测量方法

Measurement method of steel ring roundness based on line structured light

针对钢领内表面圆度在人工接触式测量中存在主观干扰及自动化程度低等问题,提出一种基于线结构光的钢领内表面圆度测量方法。通过构建图像掩模提取钢领内表面线结构光条纹的感兴趣区域,采用基于主成分分析的改进Steger方法,通过计算条纹法线方向的二阶泰勒展开得到亚像素级条纹中心,提升算法效率。由线结构光条中心坐标相机内外参数、光平面方程及旋转轴方程重构钢领内表面三维点云数据,根据圆度评价方法建立基于点云数据的钢领内表面圆度的测量模型。实验结果表明,本文方法所测最大偏差为5.8μm,与文献中投影法所测结果相比更加精确,重复性测量标准差为0.725,且算法运行时间小于40 ms。

Objective Aiming at the background that the roundness of the inner surface of steel ring is still mainly measured manually,an automatic solution based on machine vision measurement is proposed.According to the characteristics of textile steel ring,a method of measuring the inner surface roundness of steel collar based on line structured light was proposed.The method proposed in this paper aims to solve the problems of subjective judgment interference,cumbersome operation and low degree of automation in traditional manual contact measurement.Method Firstly,the camera calibration,optical plane calibration and rotation center calibration of the measurement system were completed by the calibration plate.The point cloud data of the inner surface of the steel collar was obtained based on the calibration coefficient and the extracted optical strip center coordinates.The roundness information of the inner surface of the steel ring was obtained by constructing a virtual plane.Results Roifill function is used to fill the specified region of interest polygon in the image,and the pixel values on the polygon boundary are smoothly interpolated inwards to extract the line structure light fringe area,and then the maximum inter-class variance method is used for binary processing.The hole and edge burr of optical strip were solved by morphological closed operation(Fig.6).The principal component analysis method matrix was used to solve the covariance matrix instead of Hession to obtain the normal direction of fringes.By calculating the second-order Taylor expansion of the normal direction of fringes,the sub-pixel fringe center was obtained,and the extraction time was increased from 0.0184 s to 0.0029 s,which greatly shortened the time of the traditional algorithm(Fig.7).The three-dimensional point cloud data on the inner surface of the steel collar were reconstructed from the center of the linear structure light fringe and the internal and external parameters of the camera,the optical plane equation and the rotation axis equation(Fig.8).Points on the same plane were screened out by means of zero distance between points and the plane on the virtual plane established on the vertical axis of rotation,and the circularity value was obtained by using the minimum region method.The point cloud data obtained in the experiments of three types of steel collar were calculated respectively(Fig.10),and the measurement experiments were carried out with the traditional manual method and three-dimensional measuring instrument(Tab.1)and repeatability measurement experiments(Tab.2).Conclusion Based on the principle of line structure,the parameter calibration is completed.The image mask method is combined with the maximum inter-class variance method,and the optical strip image is successfully obtained from the original image.The center of the optical strip is extracted by the improved Steger method based on principal component analysis,which improves the efficiency of the algorithm.Based on camera calibration parameters,optical plane calibration equation and rotation axis calibration equation,three-dimensional point cloud data on the inner surface of the steel collar were obtained from the fringe center of the line structure light.Combined with the roundness evaluation method,a measurement model of the inner surface roundness of the steel ring was established based on the point cloud data.Through the measurement experiment and repeatability measurement experiment of the three types of steel collar respectively,compared with the traditional manual method and three-dimensional measuring instrument,the maximum deviation measured by the method in this paper is 5.8μm,which is more accurate than the results measured by the projection method in the literature,and the repeatability measurement standard deviation is 0.725,and the running time of the algorithm is less than 40 ms,which proves that the method has practical value for the roundness measurement of textile steel ring.