Welding Operation Action Recognition Dataset

The current industrial welding sector faces challenges such as insufficient automation, high complexity of processes, and safety hazards. Existing welding monitoring systems can only provide limited real-time data and cannot accurately recognize specific actions and paths in welding operations, leading to low production efficiency and frequent safety hazards. This dataset aims to assist researchers in developing smarter welding monitoring systems by providing high-quality welding action videos, enabling more efficient production processes and safety assurance. The dataset includes video data under various welding processes, collected using high-definition camera equipment in a standard laboratory environment, ensuring the video content is rich and representative. Data annotation involves multiple rounds of review and consistency checks to ensure the accuracy and consistency of annotation information. The data is stored in MP4 format, organized by welding type and action labels, facilitating subsequent training and analysis.