Page 13 - EE Times Europe November 2021 final

EE|Times EUROPE 13

OPINION | SMART FACTORY limited the range of applications that could be
automated within a factory.
This challenge resided in capturing scenes
How Advancements in in motion and the seemingly “inherent”
tradeoff between quality and speed. Imagine
Machine Vision Propel products or product components that are
placed on a moving conveyor belt. When
they reach a robot equipped with a 3D-vision
Factory Revolution system, the vision system scans the parts, one
by one. The output of this scan is a 3D point
cloud with precise X, Y, and Z coordinates.
This 3D data is used to navigate the robot
By Andrea Pufflerova, Photoneo to approach each part, pick it, and place it
at another location or perform some further
action with it. Or the 3D data can be used for
The Fourth Industrial Revolution is ramping up. But what is it inspection and quality control. These robotic
exactly? What technological advancements enable this new wave of tasks may seem rather simple, but they are
transition to more advanced production means and processes? And not. In fact, they represent the most challeng-
what is the role of machine vision in this huge gear train? Let’s take ing applications for machine vision.
a quick spin through the history of manufacturing to understand Here is why: Traditional 3D sensing
the context of transformation that currently occurs within factories, technologies have not been able to provide
marked by a number of milestones introducing new means to mecha- a high-quality point cloud of objects moving
nize production and push it to the next level. at a fast speed. Time-of-flight systems, for
The first breakthrough (the First Industrial Revolution) was the instance, can provide a fast scanning speed
invention of the steam engine in the mid-18th century. At the end of and nearly real-time processing, but they fail
the 19th century, steam power started to be displaced by electric power. This Second Industrial to deliver a high level of detail at moderate
Revolution enabled mass production through the invention of assembly lines. noise levels. The result is a low-resolution
The late 1900s saw the emergence of computers, electronics, and digital technology, which output. On the other hand, structured light
ignited the spread of automation. This Third Industrial Revolution, or “Digital Revolution,” systems offer submillimeter resolution and
enabled the automation of entire production processes through the deployment of comput- high accuracy, but those come at the cost of
ers, machines, and robots. And then human and machine capabilities started to merge. These speed. In other words, structured light sys-
“cyberphysical systems” marked the Fourth Industrial Revolution, or Industry 4.0, transforming tems can deliver high-quality 3D data if the
traditional production facilities into smart factories, where everything is fully connected through scanned object or the camera does not move.
a communication network for data exchange — between machines, humans, and facilities. The tradeoff between quality and speed
Within this interplay of technologies, machine vision plays a pivotal part. Let’s take a look at its limits vision-guided robotics and machine-
impact on the manufacturing automation of the future and how it propels factory transformation. vision applications to tasks that involve static
scenes and fixed-vision systems. However,
WHY A SMART FACTORY CANNOT BE SMART WITHOUT MACHINE VISION Parallel Structured Light, which enables 3D
A smart factory is a highly digitized, fully automated, connected, and flexible manufacturing area scanning in motion while delivering high
environment that relies on data and communication. It makes use of the most advanced technol- resolution and accuracy, can overcome this
ogies that enable the collection, communication, and analysis of data, including machine vision, limitation. The technology was developed by
artificial intelligence, and the industrial internet of things. Photoneo and enables the capture of moving
Machine vision plays a core part in data generation and collection: It captures the physical scenes without motion artifacts.
world and transforms it into digital data in the form of point clouds so that the data can be fur- The possibility of scanning dynamic scenes
ther evaluated and translated by AI algorithms into valuable information. opens up countless applications that could
It also extends robotic capabilities to unprecedented levels. Robots equipped with 3D machine not be automated before. Among these are
vision and intelligence can perform the most complex and sophisticated tasks within a factory. tasks that require hand-eye coordination —
3D vision helps robots navigate spaces and accomplish operations that require dexterity. It is that is, mounting a 3D-vision system directly
critical for tasks such as real-time process control, product inspection and quality control, object onto the robotic arm. Traditionally, the robot
handling and sorting, robot guidance, and predictive maintenance of machines. 3D data helps to needed to stop moving to make a high-
detect issues such as defective machines and to facilitate fast intervention. quality 3D scan. This is not necessary any-
To enable these robotic tasks, machine vision needs to provide large amounts of high-quality more, which significantly shortens cycle times
real 3D data. This is necessary so AI algorithms can work with this data and process it into useful and increases productivity and efficiency.
information, which is then further communicated throughout and outside of the factory to other The resistance to the effects of move-
technologies. These can analyze it and learn from it so decisions can be made accordingly. ment or vibrations is a new machine-vision
Facilities that adopt machine vision to optimize manufacturing operations can see an expo- capability that starts a new era of manu-
nential increase in productivity and efficiency. This leads to lower costs, better product quality, facturing automation. Together with other
less waste, and the prevention of crises related to the shrinking labor force. advancements in the field, it helps transform
The market offers a number of different machine-vision technologies. So what criteria should traditional production facilities into the smart
be applied when selecting the right machine vision for a smart factory? factories of tomorrow. ■

MACHINE-VISION CHALLENGES AND ADVANCEMENTS Andrea Pufflerova is a public relations
The development of machine-vision technologies has not reached the final stage. Developers of specialist at Photoneo and writer of tech
3D-vision systems constantly improve their solutions to take vision-guided robotics one level articles on smart automation powered by
further. Yet one challenge could not be solved with standard technologies, which tremendously robotic vision and intelligence.

www.eetimes.eu | NOVEMBER 2021