Good resolution #2: Automate your industrial control using the power of computer vision

[#TogetherUpgrade2023] A new year under the sign of challenge, sharing and innovation! To start the year well, our experts have lent themselves to the game of good resolutions, a challenge that they will help you meet with their expertise and advice.

Guillaume Fortier, Head of Computer Vision Inetum, wants to challenge you ??

First of all, let's define together what visual industrial control is: it is a process that uses images on a production line to automatically check the conformity of a product.?

Within its innovation teams, Inetum is committed to disseminating and pushing disruptive Artificial Intelligence and automation technologies to its various industrial clients.

When we talk about industrial inspection we can mention :

• Visual non-destructive testing and quality control on industrial sites,
• Compliance on production campaigns,
• Monitoring and alarm triggering,
• Detection of defective products,

Let's discover together a case study: Computer Vision for automation and process monitoring??

Inetum's Computer Vision teams have carried out various automation projects aimed at making better use of human time and resource consumption at the industrial production sites of one of their clients.??

The systems developed were able to move from manual control by an operator to automated control using deep learning algorithms.??

Use case 1: Automated bubble detection in an industrial glue pack production line.

As part of the automation of the visual inspection of its products, a manufacturer in the chemical industry commissioned Inetum to develop a programme for the visual detection of defects on a production line.

Products with air bubbles were to be detected by the deployed system. Each product is inspected in less than a second, with a percentage confidence rating applied to the presence of a potential defect.?

Use case 2: Detecting anomalies on nails based on shape deformation

Visual inspection based on shape deformation has been tested on several types of objects. For example, deformed and non-compliant nails are detected and isolated.

Thanks to the automatic visual inspection, our customers' field operators are now able to be instantly alerted if faults are detected and to act accordingly.?

So what's next for you? Do you have a specific use case you would like us to work on together? Let's discuss it!

Guillaume Fortier , Head of Computer Vision Inetum

Did you know? The Computer Vision Unit has 10 people in France, Marocco and Spain working on Image and videos analysis , Tailor-made solutions and Expert advices. It is also 6 research topics in progress: Few Shot Learning , Keypoints, Clustering , SyntheticData , Avatar and BIM.?