基于熔池图像特征的焊缝质量指标设计与分析  

Design and Analysis of Weld Quality Index Based on Image Features of Molten Pool

作　　者：徐梓惠 姚屏 宾坤 陈晔 何剑朋 陈飞昕 XU Zi-hui;YAO Ping;BIN Kun;CHEN Ye;HE Jian-peng;CHEN Fei-xin(Guangdong Polytechnic Normal University,Guangzhou 510630,China)

机构地区：[1]广东技术师范大学,广州510630

出　　处：《精密成形工程》2022年第10期98-104,共7页Journal of Netshape Forming Engineering

基　　金：国家自然科学基金(51805099);2021年省科技创新专项资金(SDZX2021020);广东省普通高校重点领域专项(2020ZDZX2019);广东省教育厅特色创新项目(2019KTSCX086)。

摘　　要：目的熔池是焊缝形貌的动态成形特征,建立熔池图像质量与焊缝形貌之间的映射关系能促进焊接质量实时智能检测。方法采用暗通道处理法进行滤波降噪,差分融合算法融合暗通道图和直方均衡化图。结果利用所述算法对5种焊缝缺陷对应的熔池图像进行处理能有效滤除点状噪声,并且图像的清晰度和对比度较原图增大了两倍及以上。结论设计融合图像及焊缝形貌评分指标,建立熔池–焊缝评分方程,验证了熔池图像预测焊缝缺陷等级的可行性。Weld pool is a dynamic forming feature of weld appearance.Establishing the mapping relationship between molten pool image quality and weld appearance can promote real-time intelligent detection of welding quality.In this paper,dark channel processing method was used for filtering and noise reduction,and differential fusion algorithm was used to fuse the dark channel map and the straight square equalization map.Using the algorithm to process the molten pool images corresponding to five kinds of weld defects can effectively filter the point noise,and the sharpness and contrast of image are increased by more than two times compared with the original image.In addition,the evaluation index of fusion image and weld appearance was designed,and the scoring equation of molten pool-weld seam was established,so as to verify the feasibility of predicting weld defect grade by molten pool image.

关 键 词：熔池图像 焊缝形貌 暗通道处理法 差分融合算法 评分方程 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]