Arc welding molten pool image recognition based on attention mechanism and transfer learning

被引：0
作者：
Hong, Yuxiang [1 ,2 ]
Ying, Qiluo [1 ]
Lin, Kai [1 ]
Wang, Kaiming [3 ]
Wang, Yaoqi [1 ]
机构：
[1] College of Mechanical and Electrical Engineering, China Jiliang University, Hangzhou
[2] Key Laboratory of Intelligent Manufacturing Quality Big Data Tracing and Analysis of Zhejiang Province, China Jiliang University, Hangzhou
[3] College of Mechanical and Vehicle Engineering, Changsha University of Science & Technology, Changsha
来源：
Hanjie Xuebao/Transactions of the China Welding Institution | 2025年 / 46卷 / 04期
关键词：
attention mechanism; molten pool image; residual network; semantic segmentation; transfer learning;
D O I：
10.12073/j.hjxb.20240112003
中图分类号：
学科分类号：
摘要：
Due to the influence of complex time-varying interference and variations in process conditions during the welding process, the boundary characteristics of the molten pool are easy to be blurred, and the scale information is complex and changeable, which poses significant challenges to the accuracy recognition and the robust segmentation of the molten pool. In this paper, a molten pool image recognition method combining attention mechanism and transfer learning is proposed. Firstly, the residual block(RB) is added to the UNet down-sampling process to extract multi-scale low-level features, and the coordinate attention block(CAB) is introduced in the down-sampling and up-sampling processes to improve the feature weight of the effective region. Secondly, the pre-trained deep convolutional neural network in Pascal VOC2012 is transferred to the UNet network to realize feature transfer and parameter sharing, so as to alleviate the over-reliance of training effect on datasets. The TL-RCUNet network proposed in this paper has achieved good recognition results on the untrained MAG and TIG cross-process datasets. The mean intersection over union(MIoU) reaches 96.21% and 79.55%, respectively, which is about 15% and 25% higher than the classical semantic segmentation network. The model provides a feasible solution to the problem that existing semantic segmentation methods rely on a large number of training samples and pixel-level annotations based on expert experience. © 2025 Harbin Research Institute of Welding. All rights reserved.
引用
收藏
页码：94 / 102
页数：8
相关论文
共 17 条
[1]  
HONG Y X, JIANG Y X, YANG M X, Et al., Intelligent seam tracking in foils joining based on spatial-temporal deep learning from molten pool serial images, Robotics and Computer-Integrated Manufacturing, 91, (2025)
[2]  
HONG Y X, YANG M X, JIANG Y X, Et al., Real-time quality monitoring of ultrathin sheets edge welding based on microvision sensing and SOCIFS-SVM, IEEE Transactions on Industrial Informatics, 19, 4, pp. 5506-5516, (2022)
[3]  
ZHANG Zhifen, CHEN Shanben, ZHANG Yuming, Et al., Research progress and prospect of welding intelligent monitoring technology, Transactions of the China Welding Institution, 45, 11, pp. 10-20, (2024)
[4]  
XU Yuanzhao, LUO Jiutian, FANG Naiwen, Et al., Image processing technology for ship plate melt pool based on MS-FCM algorithm, Transactions of the China Welding Institution, 45, 3, pp. 82-90, (2024)
[5]  
HONG Y X, YANG M X, CHANG B H, Et al., Filter-PCA-based process monitoring and defect identification during climbing helium arc welding process using DE-SVM, IEEE Transactions on Industrial Electronics, 70, 7, pp. 7353-7362, (2022)
[6]  
LI Wei, LI Taijiang, YANG Lue, Et al., Application of improved U-Net algorithm in image segmentation of pipeline inner weld defect, Welding & Joining, 2024, 11, pp. 73-80
[7]  
YANG D Y, DAI P H, CUI S W, Et al., Real-time recognition of molten pools based on improved DeepLabV3 + in keyhole tungsten inert gas welding applications, Electronics, 13, 2, (2024)
[8]  
QI J Y, WU Y F., Strip steel surface defect detection algorithm based on improved Faster R-CNN, China Welding, 33, 2, pp. 11-22, (2024)
[9]  
MA Xiaofeng, XIA Pan, LIU Haisheng, Et al., Depth learning detection method for all-position weld pool, Journal of Mechanical Engineering, 59, 12, pp. 272-283, (2023)
[10]  
LI T P, CAO Y, ZHANG Y M., Analysis of weld pool region constituents in GMAW for dynamic reconstruction through characteristic enhancement and LSTM U-Net networks, Journal of Manufacturing Processes, 127, pp. 573-588, (2024)
12