Image in modal.

Adapt or die — the distillation of a quote famously misattributed to Charles Darwin — is a concept that rings true in a myriad of ways in the business world. For instance, despite inventing the digital camera in 1975, Kodak failed to embrace digital photography. In 2000, Blockbuster made the decision to not purchase Netflix for $50 million. As time passed, those decisions proved to be monumental missed opportunities.

Today, especially in areas like manufacturing and logistics, the same “adapt or die” theory applies. Customer demands continue to expand, and the ability to compete and drive revenue is increasingly contingent upon finding ways to make processes faster and more efficient to maintain or increase throughput. Industrial automation technologies like machine vision help companies meet those goals. And, fortunately for businesses around the globe, machine vision technology providers today continue to evolve and deliver innovative improvements that help keep goods and products flowing from the factory floor or warehouse to the consumers’ front steps.

Warehouse Automation, Food Demand on the Rise

Despite some contraction in 2023, the machine vision market overall remains healthy. In fact, market intelligence company Interact Analysis predicts a compound annual growth rate (CAGR) of 6.4% for machine vision between 2022 and 2028, representing increased revenues from $6.5 billion to $9.3 billion over that period. In the report, Interact Analysis shows inspection representing 40% of use cases in 2022. Around the world, businesses of all types are relying on machine vision systems to perform various types of inspections, including classic tasks such as presence/absence checking and defect detection.

While machine vision systems are deployed across a range of different industries and markets, two areas that could help drive growth in the coming years are in food and warehouse automation. Data and business intelligence company Statista, for example, predicts 6.53% growth from 2024 to 2028 for the global consumer food market, with confectionery and snacks representing the market’s largest segment at $1.77 trillion in 2024. Data in the report is modeled, among other factors, on industrial production statistics. If food production capacities increase to meet growing demands, expect machine vision and automation technologies to play a critical role.

In warehouse automation, meanwhile, Interact Analysis is predicting moderate growth in 2024 following a slowdown in 2023. When COVID-19 first hit, e-commerce sales saw significant growth, which led to increased warehouse construction, which in turn drove the demand for automation technologies in warehousing and logistics centers upward. In 2023, sales cooled off somewhat and interest rates went up, which led to a decline in warehouse construction and warehouse automation sales. In 2025 and beyond, however, warehouse automation will see double-digit growth.

Growth in the overall warehouse automation space will inherently require machine vision systems, but in order for this market or the food production market to keep pace with the predicted increase in global demand over the next few years, machine vision systems will also have to adapt. Speeding up a production line while maintaining accuracy, for example, requires faster cameras, but it also requires innovations in machine vision lighting, which is fundamental to acquiring high-quality images.

Conquering Common High-Speed Imaging Challenges

Machine vision technologies must advance in tandem with evolving automation needs and challenges. Here, examples abound. In warehousing environments, for instance, primary and secondary packaging must be inspected to ensure product quality and labels must be verified for accuracy and compliance to prevent escapes and maintain customer reputation. As processes speed up, though, new challenges emerge, such as verifying labels or alphanumeric characters through shrink-wrap, which becomes a difficult task for cameras or barcode readers because of the glare. While polarizers can effectively eliminate glare, using them reduces overall light output, which decreases image clarity and overall read effectiveness.

Meeting increased speed requirements in food production environments also poses a unique set of obstacles from a machine vision system design standpoint. Systems integrators must find ways to boost speeds and overall throughput on a production line, for example, without relying on longer exposure times, and while also avoiding pixel blur and maintaining the system accuracy. To overcome these sorts of high-speed imaging challenges in both warehouse and food production environments, implementing a new lighting design is often the answer.

Many machine vision lights today offer overdrive mode, which runs the LEDs at approximately 2X to 4X the current of the light in continuous operation mode. This new design combines two overdrive engines: a standard overdrive and a “deca” overdrive, which provides a 1 ms light pulse that is 10X the LED drive current than continuous mode. In operation, the dual overdrive mode starts with a 1 ms burst of deca mode and if the exposure goes beyond that, the light drops to standard overdrive — which is still 5X the current of continuous mode — to extend the output. While the lights have a lower duty cycle than a continuous mode light, the impact is negligible. For example, if an application has a 100 µs pulse width at a 3% duty cycle, deca overdrive can repeat every 3.3 ms, which is the equivalent of 300 fps.

Brighter, Safer Light Pulses

While dual overdrive lights can help integrators overcome some common machine vision obstacles, deploying these lights means bright strobes in open areas. These lights can distract or cause headaches and even seizures for employees working around them. To avoid the cost, space, and complexity of light shielding, system designers could use the lights in a continuous, non-strobed mode of operation. However, doing so loses the advantage of the extra light provided by overdrive mode and requires the use of 5 to 10 times the number of lights to achieve the same intensity. To effectively remove the negative impacts of bright strobing lights while allowing companies to reap the benefits of dual overdrive lighting for their machine vision systems, a new hardware feature was developed that allows the LEDs to internally trigger thousands of times per second in overdrive mode. The short-triggered pulses occur continuously over the light’s duty cycle, delivering the benefits of the extreme output of deca overdrive, while maintaining the safety and ease of use of a continuous light source.

Since people perceive light that flashes faster than 60–100 times per second as continuous, a dual overdrive light equipped with the hidden strobe feature will appear to be a light in continuous mode. For example, when a light triggers in sync with a camera for image capture, the pulses stop, and the programmed pulse for the camera is executed. After image capture, the light returns to its continuously generated pulses, showing what appears to be a “continuously on” appearance, with only a slight flicker noticeable to anyone working nearby between the switches.

Deploying dual overdrive lights equipped with this new feature not only reduces the impact of bright strobe lights for nearby employees, but also delivers the benefits of brighter light pulses for challenging applications such as high-speed food inspection.

Lights Boost Can Inspection Rates

For one system integrator, deploying these dual overdrive lights meant a can inspection rate increase of 2,000 ppm without sacrificing exposure time or accuracy. In a previous design, machine vision systems integrator EAMVision’s 360° can inspection system could inspect approximately 4,000 ppm. A Banner through-beam fiber optic sensor detected the presence of a can or bottle, which triggered four 5 MPixel cameras to capture images of the entire circumference of the can or bottle. Images were then transmitted to an industrial PC running a custom inspection application for image processing. Four images were taken from each camera and the software dewarps and stitches them together to create a flat projection of an entire can or bottle label for inspection.

Figure 2: Dual overdrive lights in an existing can inspection system.
Figure 2: Deploying dual overdrive lights in an existing can inspection system allowed EAMVision to increase its number of cans inspected per minute from 4,000 to approximately 6,000.

Keeping up with the speed requirements of a 360° inspection system poses challenges when it comes to capturing images of a continuously moving target like a can without motion blur. As the speed of the system increases, it becomes more difficult. In the previous setup, EAMVision had four bar lights running in overdrive mode, which allowed the system to capture images with 40 to 50 µs exposure time on the cans, with variation occurring due to pixel blur. Though shorter exposure time was desirable to further limit motion blur and increase speeds, the camera required 12 dB of gain to properly expose the image based on the available light.

The noise caused by the high gain limited further improvements and affected the image quality. To cut down exposure time while maintaining the benefits of overdrive lighting, EAMVision installed four dual overdrive lights. In doing so, the system’s exposure time went down to 12 µs with no additional gain and zero motion blur, effectively increasing the overall speed from 4,000 ppm to approximately 6,000 ppm. According to EAMVision, these lights have allowed them to build inspection systems that keep up with some of the fastest automation applications on the market today.

Guide Automation Forward

Sales figures aside, machine vision and automation technologies will only become increasingly important to disparate businesses worldwide in the coming years. Whether it’s for an application in a warehouse, food production line, electric vehicle battery manufacturing plant, or in the pharmaceutical supply chain, machine vision will continue to adapt in order to provide solutions as these processes expand. And remember, a machine vision system is greater than the sum of its parts — all of these technologies must progress in tandem to guide the technology into the future.