摘要
单件小批量工件的自动化焊接中,焊接轨迹的生成过程操作复杂,为此以焊缝在平面内的焊接工件为研究对象,研发了一种异型工件自动焊接装置,提出了一种基于机器视觉的异型工件焊接轨迹在线生成的方法。通过图像识别并提取出异型工件外形轮廓,将最小二乘算法应用于NURBS曲线插补,实现了最小二乘NURBS曲线全局逼近拟合,在线生成工件轮廓的焊接运动轨迹。结果表明该算法具有较高的精度和效率,能快速在线生成各种工件外形轮廓的平滑焊接轨迹,避免了焊接过程中的抖动问题,保证了焊接质量。
During automatic welding for small batch,the generating process of welding path is complex. Aiming to solve this problem,the automatic welding equipment with the workpiece whose welding seam is in plane was developed,and a method was proposed that the welding path of shaped workpiece was generated on-line based on machine vision system. The contour figure of the shaped workpiece could be extracted through image recognition algorithms,and the welding path would be generated on-line after least square method applied to NURBS curve fitting. The results show that the algorithm can generate online the smooth welding path for any shaped workpiece quickly with high degree of precision and efficiency,and the welding jitter can be avoided and the welding quality can be ensured.
出处
《焊接》
北大核心
2017年第9期1-6,共6页
Welding & Joining
基金
国家04科技重大专项(2015ZX04005006)
广东省科技重大专项(2014B090921004
2015B010918002)
广州市科技重大项目(2014Y2-00014)
关键词
机器视觉
图像识别
曲线拟合
machine vision
image recognition
curve fitting
