基于改进主动轮廓模型的强噪声焊缝识别方法.docxVIP

窑PAGE

窑PAGE30窑

窑PAGE

窑PAGE31窑

第54卷 第3期2025年2月

热加工工艺

HotWorkingTechnology

ISSN1001-3814CN61-1133/TG

DOI:10.14158/j.cnki.1001-3814. rjggy@

基于改进主动轮廓模型的强噪声焊缝识别方法

周毅,章国宝

(东南大学自动化学院,江苏南京211189)

摘要院基于结构光视觉的焊缝跟踪系统被广泛应用于机器人焊接之中袁而保障此类焊接作业质量的关键在于焊缝识别的精度遥传统的焊缝识别算法往往难以解决强焊接噪声的干扰问题袁为此提出了一种基于改进主动轮廓模型的焊缝识别算法遥首先为了克服传统主动轮廓模型手动获取初始轮廓的缺点袁设计了一种基本的分割算法来自动提取初始轮廓遥接着根据强噪声焊缝图像的灰度分布特征袁采用该算法对传统主动轮廓模型的能量函数进行了优化袁并将拓展的结构张量添加至外部能量项袁确保收敛结果的稳定与准确遥最后袁基于多段直线拟合和RANSAC直线拟合完成焊缝特征点的提取遥实验表明院在强噪声干扰下袁该算法具有良好的适准确性遥

关键词院焊缝识别曰抗干扰曰主动轮廓模型曰结构张量曰直线拟合

中图分类号院TG409 文献标识码院A 文章编号院1001-3814(2025)03-0030-05

WeldSeamRecognitionMethodunderStrongNoiseBasedonImprovedActiveContourModel

ZHOUYi,ZHANGGuobao

(SchoolofAutomation,SoutheastUniversity,Nanjing211189,China)

Abstract院Theseamtrackingsystembasedonstructuredlightvisioniswidelyusedinrobotwelding,andthekeytoensurethequalityofsuchweldingistheaccuracyofseamrecognition.Traditionalweldseamrecognitionalgorithmscanhardlysolvetheinterferenceproblemofstrongweldingnoise,soaweldseamrecognitionalgorithmbasedonimprovedactivecontourmodelwasproposed.Firstly,inordertoovercometheshortcomingsofmanuallyobtainingtheinitialcontourbytraditionalactivecontourmodel,abasicsegmentationalgorithmtoautomaticallyextracttheinitialcontourwasdesigned.Then,accordingtothegrayscaledistributioncharacteristicsoftheweldimagewithstrongnoise,thealgorithmwasusedto

optimizetheenergyfunctionofthetraditionalactivecontourmodel,andtheextendedstructuretensorwasaddedtotheexternalenergytermtoensurethestabilityandaccuracyoftheconvergenceresult.Finally,theweldfeaturepointswereextractedbasedonmulti-segmentlinefittingandRANSAClinefitting.Theresultsshowthatthealgorithmhasgoodadaptiveaccuracyunderstrongnoiseinterference.

Keywords院weldseamrecognition;anti-interference;activecon