Descriptor: Intelligent Laser Welding State Detection Continuous Frame Image Dataset (MoltenPool2D)
Year: 2025
Type: article
Abstract: The study aims to construct a high-quality continuous frame image dataset of laser welding to support the research of automatic welding state detection in intelligent laser welding. The dataset is collected by our self-built welding platform and a high-resolution industrial camera. The dataset not only makes up for the shortage of existing laser welding video data, but also provides rich molten pool and welding process context information, which facilitates the development and verification of condition detection algorithms. The article first introduces the collection method of the dataset, and then analyzes the quality of the dataset, including the distribution law of various parameters of the data, the statistical analysis of the molten pool characteristics, etc. In order to verify the effectiveness of the dataset in tasks such as welding state recognition, corresponding experiments are designed and the record and storage method of the dataset are introduced. The public release of the MoltenPool2D dataset will provide an important resource for intelligent welding research and promote the development of intelligent welding technology. In the future, we will continue to update and expand the dataset to further enhance its scientific research value and application potential.IEEE SOCIETY/COUNCILIndustrial Electronics SocietyDATA TYPE/LOCATIONMovie/Video; Shanghai, ChinaDATA DOI/PID10.21227/qj7e-bw83
Source: IEEE data descriptions.
Authors: Jiawei Fan, Songlin Li, Haonan Zhang, Qi Hu, Zhuguo Li, Jiuchao Qian
Institution: Shanghai Jiao Tong University
Language: English
Subfield: Mechanical Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $600
OpenAlex ID: W4413120797