Virtual services
Overview
Improving quality control in welding processes requires new ways to capture and interpret production data. Irida Labs, a technology provider of software IPs for Edge Vision AI sensors and solutions, focused on applying vision AI to better assess welding quality while addressing increasing production demands and the need for more efficient, digitalised processes.
By collaborating with AI-MATTERS, the company was able to test and validate its approach in a realistic manufacturing context, accelerating the development of its vision AI solutions for quality assurance and control.
Impact
With the support of AI-MATTERS, Irida Labs achieved tangible results:
- High-speed prediction of expulsion defects and diameter estimation in under 2 seconds.
- Around 10% reduction in robotic automation costs when detecting defective processes.
- Improved product quality through AI-supported predictive maintenance approaches.
- Upskilling of 10 employees in AI-based manufacturing technologies.
These outcomes highlight how vision AI can support faster and more informed quality control processes while contributing to cost efficiency and skills development.
The Challenge
Irida Labs faced the challenge of scaling its vision AI-powered solutions and positioning them within a global manufacturing market that was not yet fully explored.
A key limitation was the lack of access to resistance spot welding (RSW) equipment and appropriate testing facilities needed to run experiments and validate models in realistic conditions. In addition, specific manufacturing know-how, particularly related to defect identification, was required to reduce the time needed for data labelling.
Without these elements, progressing from development to validated industrial application remained complex.
The Solution
Through AI-MATTERS, Irida Labs gained access to advanced testing facilities, technical guidance, and a realistic industrial environment to experiment with and validate new vision AI models.
This support enabled the company to test its solutions on welding processes, refine model performance, and better understand production conditions. Access to both infrastructure and expertise helped reduce experimentation barriers and supported the integration of AI into Irida Labs’ vision AI product portfolio.
As a result, the company was able to move forward in validating its approach and strengthening its applications in manufacturing quality control.
