变极性等离子弧穿孔立焊双面图像处理与特征提取  

Double-Side Image Processing and Feature Extraction of Variable Polarity Vertical-Up Plasma Arc Welding

作　　者：陈路路[1] 胡明华[1] 陈云霞[1] 吴頔[1] 陈华斌[1] 林涛[1] 

机构地区：[1]上海交通大学材料科学与工程学院,上海200240

出　　处：《上海交通大学学报》2015年第3期361-365,370,共6页Journal of Shanghai Jiaotong University

基　　金：国家自然科学基金项目(51275301)资助

摘　　要：针对变极性等离子弧穿孔立焊的双面图像进行特征信号提取,提出了一种适应性强的图像处理算法来获取焊接参数.开发了基于组合部件模型匹配和边缘窗口线性扫描的组合算法,并通过筛选法则提取特征信号,实现对焊接双面特征的有效提取.结果表明,所提出的方法能获得可靠的正面熔池及背面小孔的几何特征信息,为视觉多信号融合提供基础.In order to obtain the double-side image feature signal for variable polarity vertical-up plasma arc welding, an adaptable image processing method was proposed. The tree part-based model image processing algorithm and linear scanning edges algorithm were used simultaneously, and the appropriate one was se- lected. The result shows that reliable geometric features of front pool and back keyhole puddle can be ob- tained with the proposed method, which enables it to integrate muti-visual signals.

关 键 词：变极性等离子弧焊 双面图像 小孔熔池图像 组合部件模型 

分 类 号：TG456.2[金属学及工艺—焊接]