New Image Processing Technology to Evaluate Welding Skills
In a significant advancement for the welding industry, the Tokyo Metropolitan Industrial Technology Research Institute has collaborated with CreativeWorks, Konno Seisakusho, and AIS Corporation to develop an innovative image judgment system. This system allows for the quantitative assessment of welding skills by analyzing the external images of welding products.
The primary goal of this cutting-edge technology is to generate an objective evaluation of completed welding products, comparing them against exemplary data from experienced professionals. By automatically measuring various elements that constitute the welding piece, the system provides not only an assessment of the final product’s quality but also offers training guidelines aimed at improving skills among new welders.
This highly beneficial aspect aims to visualize the mastery traditionally reliant on the intuition and experience of skilled workers, thereby aiding in the transfer of skills to the younger generation of welding professionals. The system is set to be showcased at the EdgeTech+ 2026, taking place from November 18 to 20 at Pacifico Yokohama.
Background of Development
The initiative traces its roots to a project funded in the fiscal year 2019, where a manual welding training system was created to visualize welding tasks. However, the evaluation of welding quality was not yet achievable at that time. In addressing the pressing skills shortage in the welding industry, two major issues surrounding training methodologies were identified:
1.
Quantitative Assessment: Previous methods relied heavily on qualitative evaluations based on the subjective observations of experienced welders, often leading to ambiguous judgments.
2.
Reproducibility: Evaluation outcomes varied significantly depending on the assessor, often resulting in inconsistent judgment results.
A detailed comparison of images between trainers and skilled welders highlights these discrepancies and the necessity of a standardized measurement system.
Technical Features of the New System
The developed system boasts several advanced technical features:
- - Automatic Measurement: The system can automatically assess the various components of welding products.
- - Standardized Criteria: It formalizes and quantifies the benchmarks for assessment, allowing for clear cut evaluations.
- - Automatic Judgment: The system provides automatic assessments on the quality of completed welding tasks, including good or bad ratings.
Value Provided by the System
The introduction of this system promises to enhance training effectiveness in two critical ways:
- - Guidelines for Skill Improvement: It offers actionable feedback that assists in skill enhancement for trainees.
- - Immediate Feedback: Instantaneous responses to trainees allows for rapid adjustments and corrections in their technique.
Collaboration Framework
In this collaborative endeavor, the Tokyo Metropolitan Industrial Technology Research Institute focused on developing the image processing and AI-based evaluation technology. The partner companies contributed their insights from skilled workers and provided essential training data for rigorous validity testing, enabling the digitization of traditional craftsmanship.
Future Plans
Following the unveiling of this collaborative project at the EdgeTech+ 2026, additional plans are set for the system’s deployment. Starting in 2027, WELDPORT and the Miyamoto Welding School will initiate training services utilizing this innovative system.
For more information, visit the official WELDPORT website:
WELDPORT