In the production of small metal parts (screws, bolts, rivets, bushings, etc.), a single defective component can cause potentially catastrophic damage. Just think of the consequences of a broken wheel bolt or a batch of aircraft rivets. To avoid this risk, quality control systems applied to the entire volume of parts produced have been in place for decades. Dimac sorting machines (approximately 1,000 worldwide) annually inspect over 20 billion components in approximately half a million different geometries, separating the good parts from the rejects.

The controls include dimensional checks, production defects such as cracks, burrs, scratches, thread breaks, and molding defects.The process, automated and based primarily on computer vision, must be extremely reliable and fast to exclude all defective pieces from subsequent processes, impacting the final cost of the product as little as possible.

The fastener industry is highly competitive; to maintain its position, Dimac has invested in innovation and the transition to a service-based business model, which has been embodied in the project Dimac AI Tools: a proposal of Artificial Intelligence-based services to perform advanced quality controls by comparing pieces with synthetic source images.

With the introduction of Artificial Intelligence at the end of 2023, Dimac, the first in the world of metal components, is now able to replicate even the most complex human visual checks, extending quality control automation to sectors where computer vision could not provide reliable answers.

The result: more effective controls and reduction of production waste, to guarantee customers economic and energy savings and reduced time to market.

The project Dimac AI Tools saw the involvement of some strategic partners:

R2M, An international strategic consulting firm focused on the industrial implementation of research results has developed the AIQuS application, which generates hundreds of artificial images of potentially defective parts to train the neural network to distinguish good components from waste.

EURIX S.r.l., a Turin-based software house with expertise in IoT, Business Intelligence, and Big Data, developed the digital infrastructure that connects DIMAC machines to a central system for data collection and analysis. This has enabled the digitization of service delivery and remote management, making global expansion sustainable.

University of Eastern Piedmont (UPO) – Department of Science and Technological Innovation, Computer Science Section – contributed to industrial research, focusing on the application of artificial intelligence to computer vision in diagnostic imaging. The collaboration fostered the development of shared solutions and the creation of a permanent channel for exchange between industry and academia.

With AI Tools Dimac has taken a decisive step towards the digital transformation of its offering, establishing itself as provider of Artificial Intelligence solutions for industrial quality control.