Proyecto Multi AI

Izertis begins to work on a defect detection and anomaly prediction project with international partners

Our company leads the consortium of the MULTI-AI project, which consists of the development of multimaterial and multidefect detection and anomaly prediction system, in which the Romanian companies Beia Consult International SRL and PETAL SA, the Belgian company Phoenix AI SA and the Belgian industrial research center Sirris also participate.

The project is co-financed by the Institute of Economic Development of the Principality of Asturias (Spain), the Executive Unit for Financing Higher Education, Research, Development and Innovation of Romania and the Public Service of Wallonia (Belgium), within the framework of the program MANUNET (reference: MNET20 / ICT3798).

“Coordinating this project is an exciting challenge, not only because of its ambitious goal, but also because of the possibility of being able to work in a multicultural team with entities from very different fields, but at the same time complementary. In addition, it allows us to continue developing technology related to artificial vision and apply it to solve the real needs of companies”, has highlighted Raquel García, our firm's innovation consultant and project coordinator.

The system to be developed within the framework of this project will allow the simultaneous analysis of multiple types of defects in different materials, without the need to have different inspection stations for each type of defect. This scalable and customizable system will allow quality control of any part manufactured on a production line.

The application of artificial intelligence will allow establishing a correlation between the variables of the production process and the anomalies that the manufactured parts could present. Likewise, the use of IoT technologies for monitoring and real-time sensing the manufacturing lines will be essential to guarantee the delivery of adequate datasets in terms of size, quality and labelling, and thus lead to a robust and reliable system.

There are several objectives set in this project, such as the design and development of a new system based on machine vision and deep learning for the multi-material, multi-defect detection and anomaly prediction in a single station, as well as the development of a semi-supervised labelling procedure that will allow the generation of knowledge on the new anomalies detected by the system itself. This will allow the categorization of defects in a flexible way and, thus, meeting the needs of each productive process in terms of quality control.

Through this project, a quality control system that can be integrated with the rest of the data from the industrial plant 4.0 at an edge on-device perspective or on-premises will be achieved. This characteristic leads to the development of a root cause detection system by connecting the detected data with the rest of the parameters of the production line.