弧焊机器人焊缝图像处理及识别方法的研究  

Seam recognition research based on laser vision in welding robot

作　　者：熊震宇[1] 姜建锋[1] 

机构地区：[1]南昌航空工业学院材料科学与工程学院,江西南昌330034

出　　处：《焊接技术》2006年第3期8-11,共4页Welding Technology

基　　金：航空科学基金资助项目(02H56007);江西省材料科学与工程中心基金资助项目(ZX200301018)

摘　　要：针对电荷耦合器(CCD)所获取焊缝图像的特点,研究了相应的图像处理方法,有效地消除了焊接过程中的飞溅和弧光对焊缝图像的干扰。通过识别激光带在焊接坡口处的变形,准确识别出图像中焊缝中心位置并计算出焊枪的偏差。采用多线程技术编写了应用程序,充分利用计算机系统资源,大大提高了图像采集、处理和存储的速度。识别算法对每幅图像的平均运算时间为16ms,识别精度达到±0.2ms。试验证明:图像处理的过程和焊缝中心识别的方法是有效的,该系统满足焊缝跟踪控制的实时性要求。According to the characteristic of images gathered by charge coupled device(CCD),the method of image processing that is high effective in eliminating the noise of spatters and welding arc light is studied in the article.The welding seam center is recognized and the value of error between torch and welding seam is figured out.Multi-thread technology is used to program the software which make fully use of computer source and increase the speed of image gathering,image processing and image saving.The average time of the recognition process is about 16 ms and the absolute error is less than 0.2 mm.The experiences show that the image processing and image recognizing are available,and the system meets the real-time requirement of welding track control.

关 键 词：视觉传感 图像处理 焊缝跟踪 

分 类 号：TP242.2[自动化与计算机技术—检测技术与自动化装置] O29[自动化与计算机技术—控制科学与工程]