Lunch lecture ROSEN: using machine learning to detect corrosion

Corrosion is an unwanted and costly phenomenon. It will deteriorate the steel construction of your bike or car and may even cause a fracture. In the same way, it is also deteriorating storage tank bottom plates. To assist our clients in detecting these anomalies, ROSEN uses magnetic flux leakage techniques. This technique enables the detection of these spots and makes early repairs possible. However, not every single sensor peak represents a real corrosion spot. For example, sensor vibrations may occur. To distinguish between real and false calls, we use the pattern of the input information. Machine learning techniques are implemented with help of ground truth data as gathered by domain experts. The computer, therefore, saves hundreds of manhours and it is also easily repeated. But what to do when the computer makes a mistake in classification; how to manage that risk?

During this lecture, we will inform you on the possibility to manage risks involved with the machine learning approach by using a section based classification. We will also hint at our next challenge on boosting the software’s speed performance and inform about possible cooperation projects.

Enrol for free delivered lunch within Enschede and Hengelo, to enjoy while participating in the lunch lecture. Due to logistics issues, the number of enrolments for free lunch has been limited, but attending the lunch lecture without sign-up is, of course, allowed.

During sign-up choose one sandwich of your choice. A list of available sandwiches & allergen information is available on

Link to the lunch lecture:

Unfortunately, you cannot enroll for this activity anymore, as the enrollment period has ended.

Who, what, where, when...

Committee for Lectures and Excursions
Lunch lecture ROSEN: using machine learning to detect corrosion
Tuesday 11 May 2021 from 12:50 until 13:35 iCal-link