TEC Eurolab
Context & Problem
TEC Eurolab is Italy’s leading industrial laboratory in the aerospace and automotive sectors, offering high-quality defect analysis using industrial tomography.
Although tomography is capable of detecting hidden defects, it requires time and money. Specialized and trained technical personnel are necessary to avoid errors and ensure reliable results.
Despite previous attempts at automation with image analysis software, the variability of defects posed a challenge for the in-house team.
Goal
Significantly reduced investment of time and money.
Solution
Our computer vision algorithm can automatically identify crucial anomalies of various shapes, sizes, and locations. We used an Agile project management approach, collaborating frequently with the development teams and the client and making continuous software releases.
During the first, crucial phase, we worked closely with the client to ensure AI compatibility with the quality and format of the acquired tomographic images. To simplify testing and reduce manual data preparation time, advanced pre-processing filters were added to the images.
Next, we designed a Deep Learning algorithm and created a simple user interface for training the AI prototype. TEC Eurolab staff can now work independently, validating the anomalies identified by the AI and adding new ones using the provided software. The algorithm processes these anomalies and retrains itself regularly to continuously improve performance.
Applied Technology
Computer Vision
Results
- Artificial intelligence can identify anomalies that may be missed by specialized technicians. This enables objective inspection, reduces time by 50%, and increases reliability.