It’s Time to Talk About Vision

The first edition of the Smart Vision Forum, the exhibition-conference dedicated specifically to machine vision and created thanks to the commitment of AIdAM, ANIE Automazione and Messe Frankfurt Italia, will be held in Bologna on June 25th. We have involved most of the companies that will exhibit at the event in a long consideration on their expectations, trends and market awareness as well research developments – in terms of artificial intelligence, but not only – regarding a truly enabling technology for the factory of the future.

di Fabrizio Dalle Nogare

“In nature there is no effect without a cause, you understand the cause and you do not need any experience”. Giovanni Genovese, Omron’s Vision Solution Specialist, uses the words of Leonardo da Vinci, who “looked at experience as the weapon used by those who had not yet perceived the importance of the method and scientific representation of phenomena”, to introduce the world of machine vision. “As Omron, we are trying to promote scientific culture behind machine vision solutions through the many innovative solutions we implement thanks to our extensive network of field technicians, and thanks to the help of our R&D department closely linked to the automation market. All this leads us to think that the first expectation is to increase the cultural background of our customers well beyond the sales prospects”.
Expectations about the Smart Vision Forum (SVF) cannot but be discussed with those who promoted it and are preparing to liven up the first edition. “My expectations about the Forum are quite high”, confirms Lorenzo Benassi, Product Specialist Position & Vision at ifm Italia. “The SVF is certainly a new way to spread culture in the field of vision products. The debate and sharing of technical and scientific know-how is supposed to be the basis of the next technological innovations”.
Elio Bolsi, General Manager of wenglor sensoric italiana, sees a significant growth of interest as for machine vision. “This is partly due to the progress made in recent years by technology, which has allowed an extension of the use of image acquisition and processing systems that goes beyond the ‘traditional’ inspection and control tasks. At the same time, we must consider that an increasing number of companies are moving towards the adoption of Industry 4.0 paradigms in order to increase production efficiency and the ability to react to the demands of an ever dynamic market that requires speed and quality, at the same time”.

An Industry 4.0 enabling factor
The project that led to the SVF started from afar, that is to say from the work carried out in the latest months by AIdAM and ANIE Automazione along with Messe Frankfurt Italia. “As one of the Forum’s promoters – says Nicola Lo Russo, Managing Director at Vision – the expectations on the success in terms of visibility, participation and number of visitors are high. The requests for participation have been numerous since the first meetings, further demonstrating the importance of machine vision systems in the field of automation and beyond. The objectives we set ourselves are manifold: dissemination, communication, discussion but also reference in a world that from being a ‘market niche’ has now become an ‘enabling factor’ for Industry 4.0”.
iMAGE S has always believed in the importance of promoting and spreading the ‘culture of vision’”, says co-founder Marco Diani. “For this reason, we periodically organize seminars and open houses, in addition to collaborating with universities and associations, both Italian and international. We therefore participate in the SVF, a very important event for us as the first in Italy entirely dedicated to artificial vision technologies, a key part of the industry of the future.
A meeting point for all machine vision professionals in Italy. This is how Giorgia Calzolari, Product Specialist at Advanced Technologies, sees the SVF, which “will strengthen the image and visibility of the entire sector as well as helping companies through the exchange of ideas and experiences, to grow and improve. Machine vision is experiencing a time of growth, both in terms of industrial applications and out-of-factory applications in the retail, safety and life science sectors. Despite this, the events and workshops devoted to vision are few and far between”.

A time for information and discussion
As a new player in the field of vision, after the acquisition of Matrix Vision, Balluff relies heavily on the success of the SVF, “both as an opportunity to bring together supply and demand – says Fabio Rosso, Head of Service Center – and for the possibility of debating between experts in various disciplines and technologies that contribute to find applications compliant with the needs of the market. Image processing is now indispensable both in industry and in other sectors, such as healthcare, traffic control, safety, analysis, research, etc.”.
“The expectations we have for the Forum are essentially to continue along the path of innovation that began several years ago in the increasingly strong spread of vision systems in the manufacturing industry”, explains Eugenio Meregalli, Technical Department Senior Manager at Keyence. “Although the use of artificial vision is now well established in every industrial sector, it is still essential to exchange information and educate end users of vision systems on new technologies and continuous innovations in this field. The Forum certainly has an important role to play in this regard”.
The Forum, according to Michele Leoni, Global Product Sales Specialist Traceability at Datalogic – is aimed at users/integrators of machine vision solutions. “On the one hand, for those who are already familiar with machine vision issues, it can be a networking and learning opportunity about the most interesting technological innovations and market trends. On the other hand, for those who are approaching this field for the first time, it can be an excellent opportunity to get to know the technology better and evaluate its application potential, as well as getting in touch with possible products and solutions suppliers“.
“Thanks to machine vision – explains Fabio Rosi, Head of R&D at VEA – many industries have significantly increased the quality of their products, but above all they are discovering new ways to produce, unthinkable until a few years ago. The SVF is a starting point to promote culture in the field of vision, an aspect that has been underestimated in the past, and I think that the idea of creating a Forum dedicated to vision is a winning one, especially because it is more focused than a traditional trade fair”.

Big data, Industry 4.0 and digital factory
“We felt the need, in Italy, to create an opportunity to update on this issue”, says Serena Monti, Product Manager Vision Systems at SICK, one of the promoters of the SVF. “We aim to create a small community where we can talk extensively about machine vision. The event is supposed to be a benchmark for operators and, we hope, the first of a long series of events of this kind, also because companies are increasingly asking for innovative tools, so the time is really ripe to make this issue explode”.
According to Mahmood Talouzi, Technical Support Engineer at Beckhoff, “the event can certainly be a starting point. The current technological challenges dictate a path of continuous improvement and are necessary to adapt quickly to market needs, offer innovation and at the same time achieve concrete benefits not only in terms of efficiency, but also in terms of economy”. The same trust expressed by Alessandro Liani, CEO of Videosystems. “The SVF will give us an opportunity to debate with other national realities in the sector with a view to mutual growth, given that machine vision, along with artificial intelligence, is about to become increasingly crucial in production processes”.
“Machine vision is a key part of industry, generating data and their processing. Big data, Industry 4.0 and digital factory are all issues related to vision: a Forum and a dedicated workshop are therefore important initiatives both in terms of topics covered and audience involved,” said Michele Giannoni, Senior District Sales Manager Northern Italy at Cognex.

Are we really aware of the potential of vision?
Thinking of the audience of users and potential users of machine vision technologies, we asked the experts if, according to their experience, there is a widespread awareness of the crucial role of vision systems in the transition to the digital factory. Let’s start with those who see the glass half full. “Surely – says Eugenio Meregalli (Keyence) – today we no longer have to ask ourselves whether or not to use a vision system in line, but rather how to use the vision system itself. The awareness of our audience of customers and potential customers is quite strong. Companies that rely on technologies like these are truly innovative and believe in investing for the growth of quality”. “Of course, yes – answers Nicola Lo Russo (Vision) -. In a world where global competition is fierce, only those productions that manage to combine productivity, safety and low costs will be able to survive. Machine vision plays a key role as it provides solutions to meet the challenges of the production of the future. Machine vision has now entered our lives: from autonomous driving to traffic management; from agriculture to healthcare; from identification and separation of materials to sustainable recycling, thanks to machine vision we can say that we migrate from product quality to a better quality of life”.
“Vision is no longer a specialized subject – argues Elio Bolsi (wenglor) – but has become familiar to manufacturers and end-users working in many different sectors. The factors that have contributed to these new openings are the evolution of the imaging hardware, as well as the development of software, which has opened up application areas until a few years ago far from these technologies because often accompanied by not trivial issues related to integration. The advent of the smart industry also requires the availability of components with small size and features that can be easily integrated on board the machine, also able to transmit the information collected in the field through standard protocols, with a view to the IoT”.
“The industrial fabric is now experiencing a competitive environment that requires not only an increase in productivity and plant availability, but also and above all an increase in the quality of its products”, says Mahmood Talouzi (Beckhoff).

More vision, higher quality
Michele Giannoni (Cognex) is also optimistic. “The huge amounts of data generated by machine vision systems, once analysed and processed, make production lines ever faster, more precise and more effective; they suggest applications, ideas, evolutions, new frontiers and ways of producing”. “Through the image processing – adds Michele Leoni (Datalogic) – it is possible to get information on products and processes, thus opening up countless application scenarios in which the data collected becomes the key to improving the efficiency and throughput of all company activities, from actual production to logistics flows, up to product quality and traceability”.
“Awareness of the crucial importance of machine vision systems in automation processes has grown considerably. Vision can undoubtedly be one of the major factors in the transition to the digital factory, supporting the work of machines and man in order to make all production processes more efficient. Vision systems also make quality control in production easier, an increasingly important factor indeed for companies wishing to maintain high quality standards of their products”, says Lorenzo Benassi (ifm). Marco Diani (iMAGE S) is also convinced that “companies are becoming aware of the key role of vision systems in the transition to the digital factory. As a result, machine vision solutions are now gaining ground in all industrial sectors: from manufacturing to electronics, from automotive to packaging, from food to healthcare”.
Giorgia Calzolari (Advanced Technologies) talks about an extremely heterogeneous audience in terms of knowledge. “Those who have a superficial knowledge of vision systems are not fully aware of the advantages brought by the use of ‘Smart Factory’ vision systems, such as the use of augmented reality technologies to support assembly activities or the use of classification data for analysis activities on the correct machine operation”.

Training and education are needed
The path to greater awareness is far from over, that’s for sure. “There is a lot of confusion and a lot of false beliefs about what vision technologies can and cannot do. In-depth knowledge is well preserved in the areas of integration and it is difficult for manufacturers or end users to approach such solutions without the support of experts. Therefore, on the one hand, we need high-performance products that are easy to integrate and, on the other hand, we need to be more aware of the role that machine vision can play”, explains Fabio Rosso (Balluff).
Referring to vision, Alessandro Liani (Videosystems) is convinced that “this branch of technology still needs training and education to fully understand the great potential for improvement that can also introduce in the view of the digital factory. In the past, controls were limited by the available computing powers; these limits are quickly zeroing and therefore, thanks to vision and artificial intelligence, new goals can be achieved also in terms of reducing waste and pollution”.
“In the meetings with our potential users I often come across a brief and varied knowledge of vision systems, sometimes seen as a sort of magic wand to solve any production problem, other times relegated to the role of smart sensor, but not too much. Our mission is to create awareness among users, generating an objective and professional vision of the pros and cons that these systems can bring to companies”, says Fabio Rosi (VEA).

An “accessory” to automation?
“Artificial vision is still perceived, albeit wrongly, as an ‘accessory’ to automation,” says Serena Monti (SICK). “I say ‘wrongly’ because, if used in all its potential, it can really be an Industry 4.0-oriented solution. First of all, it is a versatile solution: even if it is mainly used for quality control and product traceability, it does so in the most diverse situations and work environments. Such versatility, combined with an extension of the inner intelligence of vision instruments, is making it the ideal solution, for example, in robotics, where cameras are used to move safely and easily robots and cobots within the work areas and to carry out picking operations with maximum precision, even at high speeds”.
“The digital factory is artificial vision: the importance of artificial vision systems within the digital factory is equal to the ratio of the weight of sight to the total weight of the five senses. I am thinking of the development in recent times of real networks of cameras to monitor industrial environments rather than for ‘classic’ or ‘manual’ applications such as quality or process control. Artificial vision is the digital factory: I am also thinking of artificial vision systems that today carry out the job of scales, smoke detectors, spectroscopy and chemical compositions, analysis far beyond the visible. The digital factory is information: the information that can be generated, interpreted and aggregated by a generic artificial vision system goes far beyond the capabilities of any system immersed in an industrial environment”, summarizes Giovanni Genovese (Omron).

The advent of artificial intelligence
As for innovation, there is a lot of talk about research developments compared to artificial intelligence (AI). What will be the most disruptive innovations in process automation? The answers do not lack concreteness. “With regard to vision, the next frontiers of innovation can ideally be divided into two main strands: the first concerns vision technology itself, the second concerns software processing. Among the emerging technologies, structured light has the best potential”, explains Elio Bolsi (wenglor). “With regard to software, AI is undoubtedly one of the most promising frontiers. Its use in industry is still very limited, almost experimental, but it does not fail to evolve day after day, making significant progress. However, it is difficult to make predictions”.
“We are starting to use AI to increase the effectiveness of vision systems – says Fabio Rosi (VEA) -, especially in the fields of measurement and qualitative analysis. For example, until a few years ago it was inconceivable to perform high-speed measurements on products in line with the same precision of lab measurements: now it can be done”. “Videosystems – says Alessandro Liani – is investing considerable resources in the development of AI engines for process and product control, combining vision and AI with robotic technology for the development of increasingly advanced systems”. “The 3D issue in the field of artificial intelligence promises to be already in the medium term the key element of innovations to be applied to most production processes”, reflects Lorenzo Benassi (ifm).
“Innovating means choosing automation platforms that meet the criteria of flexibility and openness to the most modern standards. This can only be achieved through flexible systems made up of machines capable of making rapid format changes and which make it easier to implement the logic of automation”, adds Mahmood Talouzi (Beckhoff).

Machine learning and deep learning
According to Giorgia Calzolari (Advanced Technologies), “the IA has reached such maturity in recent years as to see the growth and strengthening of the development of applications in several industrial fields. Within AI there is machine learning, the discipline that focuses on automatic learning and, more specifically, deep learning. One of the sectors in which deep learning techniques can be used with the greatest success is that of vision”.
This is the beginning of an in-depth analysis on deep learning which, according to Fabio Rosso (Balluff), “could give rise to new applications in areas now not covered by standard techniques yet. The basic limitation stems from the fact that we move from the field of mathematics of classical vision (2D, partly even 3D) to that of statistics of deep/machine learning. And not always, in the industrial field, a probabilistic result is accepted. But there are some application areas with a considerable amount of data where such an approach can make progress. According to Michele Giannoni (Cognex), “deep learning is making new opportunities possible in the context of automated production and inspection, generating cost savings and increasing profitability and understanding of production processes”.
“Artificial intelligence”, explains Marco Diani (iMAGE S), “has existed since the 1970s, but deep learning, i.e. the use of a multi-layer convolutional neural network (CNN), in addition to increasing computer performance, has changed the type of approach to this technology and made it more usable even at the application level. For some years now, applications in the industrial field using AI-based solutions have begun to take root in the most diverse sectors: from food & beverage to welding control, up to the automotive industry”.
Serena Monti (SICK) identifies deep learning as “one of the trends of the near future. We are developing it for image analysis. The new algorithms will allow us to solve complex production tasks to detect, verify and classify objects in a completely autonomous way. Through a large amount of images, sensors are trained to evaluate the characteristics, so as to be able to understand what are the peculiarities to be analysed and automatically set the limits that establish the goodness or otherwise of a measurement, based on the statistics of the analysis carried out on the images used during training activities”.

The target is: predictive maintenance
Behind the AI there is certainly quite a complex idea, which includes different technologies and functions. “In addition to autonomous learning, visual recognition, sensors and IoT, human-machine interaction skills, just to name a few, are hot trends”, says Nicola Lo Russo (Vision). “The functions are now the most varied: in production, marketing, sales, finance as well as daily penetration in professional areas such as healthcare, legal but also in music. Value chain management based on predictive maintenance, optimization of the production process and logistics are featured by the highest expected benefit. Thanks to additional data such as images, collected by sensors like cameras, combined with the ability of deep learning techniques to analyse a large amount of data of various kinds, we are able to detect anomalies or make predictions on the residual useful life of components”.
The current trend, according to Eugenio Meregalli (Keyence), is indeed predictive maintenance, “which results not only into the machine management and the check of proper operation, but also in the tracking and prevention of defects on the processed products”. Michele Leoni (Datalogic) provides an interesting summary. “The adoption of AI in industrial processes will, in the future, bring two main benefits. The first deals with a logic of enabling technology, i.e. the ability to finally give an answer to applications that traditional technologies are not able to solve. The second, on the other hand, is about the possibility of simplifying machinery setting, thus making it more intelligent and flexible and reducing the need for human intervention”.
“I often visit large production plants, where most of the cameras are disused ‘because they generate too much false waste’”, says Giovanni Genovese (Omron), referring to his daily work. “We are therefore trying to replace a (human) intelligence with an (artificial) intelligence, created by the human being, which should not suffer from further inefficiencies and imperfections and will contribute to reducing the time-to-market of machine vision solutions as well as entire production lines. We are all working to make our systems smarter, which in this case will use that famous concept not so much appreciated by Leonardo: experience”.