摘要
在基于机器视觉的球焊金丝空间类缺陷检测中,针对传统算法提取的金丝中心线精度较低,不能用于三维重建这一问题,提出了一种基于场景信息先验的球焊金丝中心线提取方法。该方法充分利用了光器件管芯表面这一检测场景的先验知识,在通过语义分割获取金丝区域的基础上,先使用拓扑细化法初步获取骨架线,再根据Steger算法及最小二乘法对金丝重叠区域的骨架线进行校正以获取准确的中心线。实验证明,该方法在面对多种常见的金丝分布情况时均可准确地获取中心线,效果优于传统算法。
In the machine vision-based detection of spatial class defects of spherical welded gold wire,a method of extracting the centerline of spherical welded gold wire based on the scene information priori is proposed for the problem that the centerline of the wire extracted by the traditional algorithm has low accuracy and cannot be used for 3D reconstruction.The method makes full use of the priori knowledge of the inspection scene,which is the surface of the core of the optical device.On the basis of obtaining the gold wire region through semantic segmentation,the method first uses topological refinement to obtain the skeleton line initially,and then corrects the skeleton line in the overlapping area of the gold wire according to the Steger algorithm and the least squares method to obtain the accurate center line.It is proved that this method can accurately obtain the centerline in the face of many common filament distribution cases,and the results are better than the traditional algorithm.
作者
季远哲
唐立新
李斌
JI Yuan-zhe;TANG Li-xin;LI Bin(School of Mechanical Science and Engineering,Huazhong University of Science and Technology,Wuhan 430074,China)
出处
《组合机床与自动化加工技术》
北大核心
2022年第5期134-136,141,共4页
Modular Machine Tool & Automatic Manufacturing Technique
关键词
球焊金丝
中心线
骨架线
Steger算法
最小二乘法
spherical weld gold wire
centerline
skeleton line
Steger algorithm
least square method
