图像处理方法在焊缝自动跟踪系统中的应用研究  

Study on application of image processing technique in welding seam automatic tracking system

作　　者：唐刚[1] 王龙飞[2] 王璐 

机构地区：[1]南昌工程学院机械与电气学院,南昌330099 [2]上海交通大学微纳科学技术研究院微米/纳米加工技术国家级重点实验室,薄膜与微细技术教育部重点实验室,上海200240 [3]利渤海尔机械(大连)有限公司,大连116000

出　　处：《机械设计与制造》2011年第12期87-89,共3页Machinery Design & Manufacture

摘　　要：以基于视觉传感器的焊缝自动跟踪系统为研究对象,论述了相应的焊缝图像处理方法,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。主要包括转灰度图、图像滤波、图像分割、图像边缘检测、图像细化以及HOUGH变换等。其中采用卷积滤波和中值滤波相结合的方法,提高了中值滤波能力和处理速度;采用Sobel算子,较好地保留了焊缝图像的边缘。并将这些方法综合应用成功地识别出焊缝的两条边缘,进而提取出焊缝中心,并仿真出焊枪在纠偏过程中的运动轨迹。试验表明方法处理效果好,处理速度快,能够满足跟踪系统的实时性要求。It studies the image processing technique in welding based on visual sensor.The discussion is mainly focused on the techniques age processing,including converting to gray image ,image filterlng,image tion , image thinning seam automatic tracking system used in the whole course of imsegmentation,image edge detecand HOUGH transform,which can effectively eliminate the interference of spatter and arc light in welding seam image.And the combination of convolution and media filter can solve the effect of median filter and the processing speed.Moreover,the edge of welding seam image can effectively reserved by using Sobel operator, With the above method,two edges of welding seam are successfully indentified. Then,the center of welding seam is extracted and the welding torch movement in the process of rectifying deviation is also simulated.The tests indicate that the method is effecive with fast processing speed,which can meet the red-time requirement of welding seam automatic tracking system.

关 键 词：视觉传感器 图像处理方法 焊枪纠偏仿真 

分 类 号：TH16[机械工程—机械制造及自动化]