基于视觉传感的铆焊点处焊缝识别方法.pdfVIP

机械工程师

MECHANICALENGINEER

基于视觉传感的铆焊点处焊缝识别方法

王树强，陈光凯，郝振坤

（沈阳化工大学机械与动力学院，沈阳110000）

摘要：

针对组合件自动焊接时铆焊点对焊缝识别造成严重的干扰问题，提出一种基于视觉传感的薄板焊缝识别方法。首

先利用CCD工业相机采集带有激光条带的焊缝图像，再用中值滤波平滑噪声，OTSU自适应阈值分割图像，骨架提取细化

激光条纹，概率Hough变换拟合直线，采用基于二值矩阵的方法在ROI区域去除铆焊点对焊缝识别的干扰，最后利用直线

拟合得到的端点坐标找到焊缝中心点。试验结果表明，该方法误差小于0.6mm，能够快速提取焊缝特征点，满足焊接要求。

关键词：

视觉传感；焊缝识别；图像处理；铆焊点

中图分类号:TG447；TP391.41文献标志码:粤文章编号：员园园圆原圆猿猿猿（圆园25）01原园024原园4

IdentificationMethodofWeldingSeamatRivetingPointBasedonVisualSensing

WANGShuqiang,CHENGuangkai,HAOZhenkun

(CollegeofMechanicalandPowerEngineering,ShenyangUniversityofChemicalTechnology,Shenyang110000,China)

Abstract:Inordertosolvetheseriousinterferenceproblemofweldingseamidentificationcausedbyrivetedweldingpoints

inautomaticweldingofassemblyparts,amethodofweldidentificationbasedonvisualsensingisproposed.First,aCCD

industrialcameraisusedtocollecttheweldimagewithlaserstrip,thenthemedianfilterisusedtosmooththenoise,the

imagesaresegmentedwithOTSUadaptivethresholding,thelaserstripesarerefinedwithskeletonextraction,astraightlineis

fittedwithprobabilisticHoughtransform,andabinarymatrixbasedmethodisusedtoremovetheinterferenceofrivetedweld

jointsontherecognitionofweldseamintheROIregion.Finally,thecenterpointofweldisfoundbyusingtheendpoint

coordinatesobtainedbylinefitting.Theexperimentalresultsshowthattheerrorofthismethodislessthan0.6mm,which

canquicklyextracttheweldfeaturepointsandmeettheweldingrequirements.

Keywords:visualsensing；weldidentification；imageprocessing；weldingpoints

0引言中心算法，通过曲线拟合提取特征点获得焊缝二维