基于视觉传感器的钢管塔焊缝特征提取方法  

Weld Feature Extraction Method of Steel Pipe Tower Based on Vision Sensor

作　　者：王树强[1] 王旭 高元德 贺久洲 WANG Shuqiang;WANG Xu;GAO Yuande;HE Jiuzhou(School of Mechanical and Power Engineering,Shenyang University of Chemical Technology,Shenyang 110142,China;Dezhou Guangxin Steel Structure Co.,Ltd.,Dezhou 253000,China)

机构地区：[1]沈阳化工大学机械与动力工程学院,沈阳110142 [2]德州广鑫钢结构有限公司,山东德州253000

出　　处：《机械工程师》2024年第7期30-33,共4页Mechanical Engineer

基　　金：辽宁省教育厅基本科研面上项目(Z20220774)。

摘　　要：为了快速、准确地提取输电钢管塔纵向焊缝的特征点,提出一种基于视觉传感器的钢管塔焊缝特征提取方法。通过激光视觉传感器采集焊缝图像,利用中值滤波、顶帽变换去除焊缝图像中因坡口表面弧光反射所产生的干扰信息,采用基于灰度值累加法和最小距离划分感兴趣区域,采用灰度重心法提取中心线,采用坐标法和Hough变换相结合提取特征点。试验表明,该方法能够稳定、准确地提取钢管塔焊缝的特征点,提取的焊缝特征点与实际焊点的误差均不大于0.23 mm,为基于视觉传感器的焊缝跟踪系统研究奠定了基础。In order to quickly and accurately extract the feature points of the longitudinal weld of the transmission steel pipe tower,this paper proposes a method for extracting the characteristics of the steel pipe tower weld based on visual sensor.The weld image is collected by the laser vision sensor,the median filter and top hat transformation are used to remove the interference information caused by the arc reflection of the groove surface in the weld image,the area of interest is divided based on the gray value accumulation method and the minimum distance,the center line is extracted by the grayscale center of gravity method,and finally the coordinate method and Hough transform are combined to extract the feature points.Experiments show that the method can stably and accurately extract the characteristic points of the weld of the steel pipe tower,and the deviation between the extracted characteristic points and the actual weld joints is within 0.23 mm,which lays a foundation for the weld tracking system based on vision sensor.

关 键 词：钢管塔 最小距离 坐标法 中心线提取 特征提取 

分 类 号：TG409[金属学及工艺—焊接]