摘要
对基于熔化极气体保护(Metal active gas,MAG)焊的管道打底焊过程中通过电荷耦合器件(Charge-coupled device,CCD)摄像机与复合滤光技术组成的光学系统实时采集的正面焊缝区域图像进行分析,实现熔池图像边缘提取。对获取的焊缝区域图像,经过高斯平滑后,考虑到初始轮廓对Chan-Vese主动轮廓模型提取边缘效率的影响,使用基于阈值的方法,用矩形标出熔池的初始区域,即实现熔池区域粗定位,求取矩形区域中心,以该中心设定一椭圆作为熔池初始轮廓,使用Chan-Vese主动轮廓模型提取熔池边缘。借助该模型实现了MAG管道打底焊焊缝区域图像的熔池边缘提取,与Sobel变换等方法比较,该方法提高了熔池边缘提取的精度。
The welding seam image gathered by an optical system composed of charge-coupled device(CCD) camera and composite filter technology for pipe-line back welding of metal active gas(MAG) is analyzed in a bid to extract the weld pool edge.Considering the effect of initial contour on Chan-Vese model,extract the rectangle representing the coarse location of molten weld pool by using the gray information after Gaussian smoother,and set the initial contour of Chan-Vese model at the center of the rectangle,then use the Chan-Vese model to extract the edge of the weld pool.Compared with the traditional methods,the experimental results show the precision of weld pool extraction is greatly improved.
出处
《机械工程学报》
EI
CAS
CSCD
北大核心
2011年第12期74-78,共5页
Journal of Mechanical Engineering
基金
江苏省科技成果转化专项基金资助项目(BA2007058)