FANG Jimi, WANG Kehong, HUANG Yong, et al. Image edge extraction of GMA-AM weld pool based on FCM-CV[J]. 2018,48(6):19-23.DOI:
基于FCM-CV的GMA-AM熔池图像边缘提取
摘要
针对熔化极气体保护电弧(Gas Metal Arc,GMA)增材制造(Additive Manufacturing,AM)熔池图像,提出了一种基于模糊C-均值聚类(fuzzy C-Means,FCM)协作主动轮廓(Chan-Vese,CV)模型的熔池图像分割方法。该算法利用FCM粗分割理论设定CV模型的初始化位置,然后利用CV主动轮廓模型提取GMA-AM熔池轮廓。结果表明,FCM-CV算法消除了CV模型对初始位置敏感的问题,能够有效地提取不同工艺条件下的熔池轮廓,与传统分割方法相比,提取到的熔池轮廓准确、边缘光滑封闭,同时该算法能够准确提取电弧区、浮渣区和熔池尾部的轮廓,具有很好的适应性和稳定性。
Abstract
A weld pool image segmentation method based on FCM cooperative CV model is proposed for the weld pool image of GMA-AM.The algorithm uses the FCM segmentation theory to set the initial position,and then uses the CV active contour model to extract the weld pool profile.The results show that the FCM-CV algorithm eliminates the problem that the CV model is sensitive to the initial position and can effectively extract the contour of the weld pool under different process conditions.Compared with the traditional segmentation method,the weld pool profile of FCM-CV is accurate,smooth and closed.Also,the algorithm can accurately extract the contours of the arc,scum and the tail of the weld pool,and it has good adaptability and stability.
关键词
主动轮廓模糊聚类边缘提取增材制造视觉传感
Keywords
active contourfuzzy clusteringedge extractionadditive manufacturingvisual sensing