Imago Vision - The role of technology: how evolution depends on algorithms, hardware, software and research.
11/02/2022

The role of technology: how evolution depends on algorithms, hardware, software and research.

 

THANKS TO TECHNOLOGICAL PROGRESS, INDUSTRY HAS CHANGED MUCH MORE THAN YOU THINK, BOTH IN THE PRODUCTS THAT CAN BE REALIZED AND IN THEIR PRODUCTION PROCESSES. WE RECEIVED THE OPINION OF FRANCESCO MAGRI, HEAD OF RESEARCH AND DEVELOPMENT OF ARTIFICIAL VISION IN IMAGO.

 

We could define it as the engine of development, the indispensable tool for the evolution of society and industry in all its fields.

Technology is an irreplaceable resource: hardware, software and algorithms are the basis of the functioning of many every day objects, but also of the same production processes that led to their creation.

Francesco Magri, research and development manager of Imago’s artificial vision section, helps us to understand the role of technology in the industrial sector.

 

How has the way of producing changed in recent years and wha thas been the fundamental role of technology?

"For some time we have been living in an era of over production: there are many players on the market, quality standards are growing and profit margins are getting lower and lower. Added to this are the scarcity of raw materials and the need to reduce the impact on the environment. In such a scenario, in order to survive a company cannot waste resources and the difference between success and failure may depend on minimal precautions. The technology in use is based on a scientific approach, aimed at optimizing every aspect of production, reducing costs and maintaining high quality standards. Basically, it's not just about making better products, but about producing more and more efficiently ”.

 

Is this the job of your type of company?

"Yes. We deal with computer vision and our goal has always been to overcome the concept of a simple good / reject selector in the quality control of parts. Sometimes it is a question of translating the knowledge of expert staff into algorithms (and therefore into software) capable of replicating their work; other times, to go beyond the capabilities of the human eye in terms of resolution, speed and repeatability of the result, intercepting very small defects or even those outside the visible spectrum. Using special optics, polarized waves, ultraviolet light or thermal cameras it is possible to detect imperfections that would otherwise go unnoticed. Our task therefore is not limited to defining whether a part is compliant or not: the information collected often allows us to understand what the problem depends on and which point of the production line is most responsible. In addition, by analyzing the trends we can also identify the presence of a discrepancy before a defective product is generated, allowing us to intervene in a targeted manner and with preventive maintenance ".

 

What are the necessary skills to operate in this sector?

“Oursis a multidisciplinary profession, which deals with optics, mechanics, physics and information technology. It starts by listening to the customer and identifying whath is realneeds are.

The following step is to design the optimal solution, choosing the most suitable technologies among the best available and finding the most performing configuration, in terms of both effectiveness and costs. The role of the engineeris to bring together the best existing technologies to create an innovative solution. It is not so much a question of inventing from scratch, as of applying, experimenting and testing, with out ever stopping to do research ”.

 

This is for the hardware component. What about the software?

"Once the devices have been chosen, the most difficult part remains: teaching a computer to do human work. To do this, we must convert human reasoning into operations that can be replicated by an electronic brain, translating the operator's experience into machine language. This is not at all obvious, which requires an in-depth comparison to acquire all the skills that will then have to be transferred into an algorithm. This algorithm will subsequently be perfected, made more efficient and written in such a way that it can be easily and quickly adapted if necessary ".

 

What should we expect for the near future?

"In the field of artificial intelligence, today we talk a lot about “deep learning”: in practice it is no longer the programmer who translates and writes the code, but the machine it self that develops the algorithm starting from the collected data. We are only at the beginning, but we are talking about a huge evolutionary leap, the overcoming of human intelligence it self. We have already begun to experiment and analyze the results: even if today there are still relatively few sectors in which these techniques find a real application space, the potentialis exceptional, in any field. And it's just a matter of time”.