Being as impressive and versatile as they are, machines will continue involving themselves in the industrial space in higher and higher volume. Already we see vast saturation in this regard when you look at the use of industrial computer hubs for smart manufacturing, heavy machinery used to lift large components, additive manufacturing machines that can create parts from scratch, and much, much more. That said, robots are more and more commonly being added into this impressive array of machinery as we push further into Industry 4.0 initiatives. 

Robots are machines capable of carrying out a complex series of actions autonomously. Think of a Roomba as a barebones example. The machine is capable of detecting its surroundings with built-in sensors, and, using those detections, it can determine the best route for cleaning floors without bumping into walls. 

And while the roomba is a novel example of a robot, it actually exemplifies something unique about newer robots entering both the home and the industrial space. Unlike robots of old that simply outperformed humans in physical tasks such as lifting heavy objects, modern machines are beginning to employ more cerebral tasks due to the innovation of machine vision capabilities.

What is Machine Vision?

Machine vision is the process of a machine using digital input captured by a camera to make decisions on how to behave and what actions to take. Of course, there’s more involved than just a camera. Machine vision requires multiple pieces of hardware and software including lighting, machine sensors, software capable of processing images captured by these devices, and more.

It can help to think of machine vision as the same process involved in a human when they look at an object. Imagine a human is placed on the factory floor of a canned food supplier and it’s their job to ensure each can of product is labeled with a red label. The human looks at each can with their eyes and that image is transmitted to the brain and analyzed to determine if they, indeed, have a red label. The same thing happens with machine vision, just swap out eyes with “camera”, and brain with “processor”. 

Where machine vision sets itself apart, however, is in its lack of human error. To go back to our example, when a human has encountered hundreds, if not thousands of cans with the same red label, it’s possible that, after looking at a can with a blue label, the eyes may send a signal to the brain describing such, but the brain instead defaults to the previously, repeatedly imprinted image of the red labeled can as a sort of shorthand. This is where we begin to notice what we call “human error” on the factory line. Machines can bypass this tendency of human error by not subconsciously defaulting to imprinted images like that.

Machine Vision Use Cases 

Now that we’ve adequately answered “what is machine vision?”, let’s delve into some use cases that make it such an impressive innovation for the manufacturing sector.

Quality Control in Food and Beverage Manufacturing

Machine vision has been used rather frequently in the food and beverage industry for several purposes. Since machines lend themselves well to repetitive tasks like the one we mentioned above, many food manufacturers have specifically incorporated machine learning towards enhanced quality control.

Heineken, for example, utilizes machines tasked with scanning bottles for fill levels to ensure each bottle isn’t under or overfilled before being shipped out. The company’s bottling machine blasts through 22 bottles a second, clearly too fast for a human to catch 100% of all quality issues without stopping the machine periodically and driving down productivity. With their new machine vision solution, however, Heineken is able to inspect 80,000 bottles an hour with 100% accuracy and without making costly stops to their factory line.

Predictive Maintenance in Automotive Manufacturing  

Even momentary downtime on an auto factory floor can cost manufacturers thousands. Thankfully, machine vision can be used to diagnose breakdowns in much the same way as human vision and analysis. Analyzed in conjunction with data gathered from IoT devices and sensors, camera visuals can help computers pinpoint upcoming breakdowns in, not just automotive manufacturing lines, but in any and all supply chains that involve complex machinery.

FANUC can attest to the savings that a company stands to gain as a result of a machine vision augmented predictive maintenance system. After piloting their Zero Down Time (ZDT) platform across several automotive plants for 8 months, they observed 72 component failures that were caught and fixed before they had the chance to grow into more costly breakages.

Setting up predictive maintenance for an industrial plant can be as easy as mounting a Din Rail computer on the factory floor, using it as a kind of hub for all of the inputs drawn from your machine vision cameras and IoT sensors.

Warehouse Inventory Tracking

Many warehouses have even begun using machine vision to closely track both supplies and available product inventory. Using machine vision software that applies algorithms to feeds of inventory shelves, workers are able to receive automatic updates as soon as the machine detects a certain shelf or product has been depleted.

If workers are equipped with wearable devices or industrial tablets through which they can receive automated notifications, the right worker or team member can be notified across anywhere on the factory floor and tasked with replenishing that depleted product or ordering more stock.

Barriers to Entry

As is the case with any new piece of technology, rollout of a machine vision system comes with a few caveats manufacturers owe it to themselves to consider.  


Naturally, incorporating smart manufacturing tech such as machine vision takes time and effort. And, while the results can surely be worth it, growing pains in the form of lost productivity while teams grow accustomed to the new tech may not be affordable for smaller operations. 

Furthermore, some businesses may not need machine vision. In order to justify the cost of machine vision, there needs to be a justifiable need for remote monitoring. If you normally run into quality control issues or machine breakdowns that put you out of business, it might be worth it to implement machine vision optimization. Otherwise, it may be something to table and come back to once your operation has grown a little more. 


Unfortunately, machines and robots aren’t at the stage where they can babysit themselves. Just like machines that break down and require maintenance, machine vision software and hardware can also run into technical issues and breakdowns. Having an IT team capable of addressing these breakdowns can ensure your production doesn’t come to a halt because of a malfunctioning camera or piece of software. 

If your IT team isn’t that sophisticated, you can still employ machine vision, but ensure the customer support of the provider delivering that software and hardware is responsive. In the case of a malfunction, you’ll need a team that can address your concerns in a timely manner instead of sending you through an automated phone tree.


A common, and very founded, concern when it comes to incorporating more advanced technology such as machine vision, is the increasing dependency on interconnected devices and internet connectivity. It’s a logical fear when you consider the abundance of cyberattacks leveraged towards companies that specifically target connected devices such as wearables and ERP software.

Teams that continue to advance their smart manufacturing efforts will want to stay abreast of cybersecurity best practices for their industry. They may also find it wise to invest in other technologies geared towards bolstering cybersecurity such as edge computing or blockchain networks.

Removing Human Error From the Equation

Machine vision has long been incorporated in several industries and businesses to resounding success. Those interested in reaping some of those benefits for themselves need only keep in mind the considerations we mentioned above and proceed with a methodically tested, carefully executed rollout of machine vision technology. If you’re interested in learning more about how your factory can get started, contact an expert from Cybernet today.