The realization that all areas of life are changing is nothing new. This naturally also applies to all economic sectors, and even extends to machine vision. But what are the challenges in this comparatively young industry? In which sectors will demand rise, and where will it perhaps decrease? What major machine vision trends are expected over the next twelve months? Dr. Maximilian Lückenhaus, Director Marketing + Business Development at MVTec Software GmbH, provides a forecast.
Dr. Lückenhaus, the entire economy is facing various crises, shortages, and uncertainties. How is that affecting the machine vision sector?
Dr. Maximilian Lückenhaus: At present, the overall economic situation is naturally shaped by the challenges of the last two to three years – from the COVID-19 pandemic to the Ukraine war, which not only brought human suffering, but also had far-reaching consequences for global supply chains, the energy market, and the general price level. One of the lessons learned from these crises is that companies need to increase their resilience. And automation plays a key role here. Companies are also trying to make their supply chains more robust, which sometimes means moving production capacities back to their home countries. Increased automation is also essential in this regard. However, this trend is being hampered by the continuing skills shortage. After all, even, and especially, automated process chains require highly qualified experts to accompany the workflow with their professional expertise. Machine vision offers powerful answers to these challenges.
What exactly are those answers?
The environment we operate in has a far stronger influence on the future economic development of our sector than it used to. What’s more, the aforementioned risks and geopolitical crises are affecting investment projects in our customer sectors, also with regard to the flexible relocation of production sites for example. To this end, it is essential for machine vision components to support different user languages, but also different local industry standards, for example. That’s point one. Point two involves the developments within the machine vision sector. These are primarily driven by technology. The all-dominant topic in this regard is deep learning. For machine vision companies, this firstly means supporting their customers even more closely, for example in applications. It also means they need to take a leading role regarding technology. It will not suffice to simply master deep learning; companies also need a firm command of classic methods and to make both as easy as possible to use. After all, in addition to automation itself, making technology easy to use is an important response from machine vision to challenges such as the skills shortage.
How do you think the importance of deep learning for machine vision will develop?
Providing it is implemented in an industry-compatible manner, deep learning will become increasingly important as a means of both supplementing and developing classic machine vision methods. according to a recent study by the analysis company Gartner, deep learning is currently in a consolidation phase. In other words, the hype is dissipating, and companies are addressing the topic with greater realism.
One fitting example of this is the increasing use of AI accelerators. As deep learning places high demands on computing power, high-performance hardware components are required. Deep learning accelerators are small, energy efficient, and powerful chips designed to significantly increase the speed of deep learning processes. Thanks to their compact structure, the accelerators are now also enabling deep learning algorithms to be used in computing units with a small footprint, such as industrial PCs. This is an example of how deep learning can be used in a meaningful and appropriate way in a growing number of applications.
In which industries will machine vision gain momentum in the future?
From what we have seen, even today, machine vision is increasingly being used outside the traditional manufacturing industries, for example in the food, logistics, or pharmaceutical sectors. To take full advantage of the technology’s benefits, its use needs to be adapted to the requirements along the specific value-added chains. This can be illustrated by the following example: Customers from these sectors are increasingly asking for small, compact sensor devices. If machine vision functions are embedded into these “edge devices”, we refer to this as “embedded vision”. The key benefit here is that the results from these sensors cannot only be used locally, but can also be globally monitored and interpreted. For example, to centrally collect and evaluate quality specifications across many transported parts. For machine vision experts like us, the latter means that we also offer our customers cloud-based solutions.