Data for Wire Arc Additive Manufacturing Video Dataset (WAAM-ViD)
Date published
Free to read from
Supervisor/s
Industry supervisor/s
Journal Title
Journal ISSN
Volume Title
Publisher
Department
Course name
Type
ISSN
Format
Citation
Abstract
The Wire Arc Additive Manufacturing Video Dataset (WAAM-ViD) is a dataset designed for multi angle melt pool dimension analysis in WAAM. It contains 110 videos recorded from five different orientations (as shown in Fig. 1), each averaging 26.5 seconds in duration and captured at 25–30 frames per second. The videos were filmed using two types of cameras: a passive camera (n = 80) and an active laser-illumination camera (n = 30). In addition to the videos, the dataset includes detailed metadata such as the camera matrix, distortion coefficients, rotation matrix, translation vector, presence of external lighting, arc current, travel speed, and weld seam dimensions. A subset of annotated frames is provided for training and testing the DeepLabv3 model used for melt pool segmentation. All metadata are also included.
