摘要
光切法是实现准确测量工件表面粗糙度的可行方法之一。针对光切法存在的取样长度与测量精度的矛盾,通过优化光切法,提出了一种基于缝合线驱动图像拼接的表面粗糙度测量方法。利用线性激光发生器、CCD工业相机、精密线性位移平台及计算机等搭建测量实验平台,采集三种加工样块(分别对应车、立铣、刨三种处理)的光切图像,利用上述拼接算法获得拼接图像,利用Freeman链码跟踪算法提取图像光带边缘轮廓,利用最小二乘法求取轮廓中线,建立粗糙度轮廓算术平均偏差计算模型。对比单光切法、触针法及光切法拼接前后表面粗糙度的测量值,结果显示,光切法比触针法的结果更为准确;单光切法相较触针法的平均相对误差降低了2.75%,但对刨削工件的测量误差较大;所提方法相较单光切法的测量误差平均降低了1.42%,对于微观不平度间距大的切削工件尤为明显,测量误差降低了2.57%。所提方法能够解决传统光切法的取样长度小、精度低的问题,能更精确地测量切削工件的表面粗糙度。
Objective Surface roughness is an important index to characterize the surface quality of workpieces.With the improvement of automation in machining,the requirement of roughness on-line measurement is put forward.Compared with the traditional contact method,the non-contact and fast light section method does not cause scratches on the workpiece surface,and it is one of the effective methods to measure the workpiece surface roughness.Theprinciple of the light section method is as follows:a linear beam of light is projected on the workpiece surface at an angle of 45”and intersects with the surface contour to form a stripe beam and the camera captures the image magnified by the microscope lens and transmits it to the computer,which processes the image to calculate the roughness value.This method can obtain the height information of the uneven surface within the length range of the reflected light.Different from other non-contact optical measurement methods such as light scattering,speckle,and interferometry,the light section method is used to measure the workpiece surface roughness directly.Due to the limitation of instrumental structure and microscopic field of view,the light-section method has the problems of small measuring range and low precision.In order to solve these problem,this paper presents an optimized light section methodology to calculate the surface roughness based on the seam-driven image stitching algorithm.Methods The measurement test platform(Fig.7)is built by a linear laser generator,a CCD camera,a precision linear displacement platform,a computer,and others.The light section images of the turning,end-milling and planing samples are captured and stitched.The seam-driven image stitching algorithm consists of the following steps:1)calculating the SIFT feature points of two mosaic images;2)filtering the matching feature points by the optimized RANSAC method;3)computing a series of homographic matrices and calculating the seam-cuttings;4)evaluating each seam-cutting to select the stitched image that
作者
赵馨雨
屈盛官
吕继亮
徐爱民
姚添文
李小强
Zhao Xinyu;Qu Shengguan;Lv Jiliang;Xu Aiming;Yao Tianwen;Li Xiaoqiang(School of Mechanical&Automotive Engineering,South China University of Technology,Guangzhou 510640,Guangdong,China;Guangdong Key Laboratory for Processing and Forming of Advanced Metallic Materials,Guangzhou 510640,Guangdong,China)
出处
《中国激光》
EI
CAS
CSCD
北大核心
2022年第9期30-38,共9页
Chinese Journal of Lasers
基金
广东省自然科学基金(2021A1515010550)。
关键词
测量
表面粗糙度
光切法
图像拼接
measurement
surface roughness
light section method
image stitching
