Advances in components, software, and integration techniques continue to power growth in machine vision and automated inspection for industrial automation. Key to success is to keep up with changing trends, broadening component offerings, and new imaging and analysis techniques. In this article, we will review the “state of the market” and discuss some established technologies that are maturing to provide value to more end users, as well as some “cutting-edge” technologies that may bear watching. Our discussion also will focus on practical application of certain technologies and evaluate where capabilities have not yet lived up to the market and user expectations.

Market Challenges Are Easing

The past year in industrial automation could be fairly characterized as “tumultuous.” With a return in the early part of 2022 to some normalcy from effects of the pandemic, machine vision and related tech industries faced the usual as well as new challenges. The shortage of skilled workers has been both a hardship and advantage for manufacturers in this marketplace for some time and the trend likely will continue. Customers seeking to mitigate loss of productivity driven by staffing problems have turned more than ever to automation as a solution with machine vision and robotics at the center of the discussion. Conversely, producers of these technologies face the same staffing issues, and the ability to deliver components and systems has markedly impacted implementation in some areas. The trend is compounded by the newly trending supply chain crisis that eased only slightly as 2022 unfolded. Shortages of sensors, integrated circuits (ICs), FPGAs (field programmable gate arrays), memory, and other core technologies continued in many cases to result in long lead times in machine vision components like cameras, smart cameras, PCs, embedded computers, and even illumination systems. Also impacted were enabling technologies like robotics and process control. Although some relief is in sight now, with lead times coming back to two to three months (from a worst case at some points of five to eight months) for some components, conventional wisdom in the market suggests that it may be into the second quarter of the year before normal delivery returns.

Trends In The State Of The Machine Vision Market

Despite challenges, the machine vision marketplace (and the closely related robotic marketplace) continued to experience strong growth in the first half of the 2022 as reported by A3 (Association for Advancing Automation) although final numbers are not yet available. However, in a member survey from Q2 taken by A3, sentiment about growth in the market for machine vision components and systems was decidedly leaning towards a downturn. This sentiment might be only a reflection of the market challenges noted earlier since other market pundits and analysts both inside and outside of the industry seem to indicate in recent published studies that machine vision and robotics overall will continue to grow at a good rate over the next four to five years. Driving home the impact of the current market environment, the A3 survey also showed that supply chain disruption, shortage of skilled labor, economic uncertainty, and inflation were (in that order) the biggest immediate challenges to their respective companies. Ultimately, the key trend commercially though seems to be that demand remains high as manufacturing seeks to continue its drive towards greater productivity and mitigation of workforce shortages through implementation of automation, with this initiative strongly supported by ongoing advances in machine vision and related technologies.

Technology Trends – Imaging Components

One of the more visible trends in technology involves both a technology and an application base. The technology is 3D imaging, and the application base is 3D vision guided robotics. While machine vision has long been a key enabler in robotic automation for a variety of use cases, the continued evolution of the capability of 3D imaging systems is paving the way for growth in vision guided robotics solutions over widely varying industrial markets. Most notable has been the rise of standard solutions that incorporate the 3D imaging technology, advanced software, and a robot and related components in integrated systems that hold the promise of broader and easier implementation and integration in applications requiring 3D location and part handling and even random object pick and place. Some use cases that can immediately realize the benefits of these systems would be logistics and distribution, and machine tending. Similar application-targeted robotic solutions also are using either 3D imaging or 2D imaging for a broad range of flexible inspection applications with at least one system notably incorporating analysis of a CAD model to automatically drive an imaging system to inspection points on an object without robot programming.

The trend in 3D imaging is buoyed by ongoing technological advances as well. Many systems now natively provide either grayscale or RGB “texture” images correlated with the 3D data. This type of imaging enables new strategies for object identification and segmentation as part of 3D location or measurement. A long-time challenge for some types of 3D cameras, high-resolution, wide-area imaging of parts in motion at relatively high rates is now available in the market.

A maturing technology, non-visible imaging continues to be a trend as it gains momentum in the marketplace. The broad value of non-visible imaging, particularly in shortwave infrared (SWIR) for inspection and mid- to longwave infrared (MWIR, LWIR) for thermal imaging has resulted in growing adoption of these technologies in the industrial automation space. Shortwave infrared is particularly of interest due to its unique interactions with some materials, like the ability to render some opaque plastics transparent while leaving water non-transparent to certain SWIR wavelengths. SWIR is used in many implementations of hyperspectral and multispectral imaging to identify materials and chemicals by spectral signature. Illumination components that deliver SWIR wavelengths also continue to be on the rise in availability, helping to advance and facilitate more potential use cases. Also, with the reference above to hyperspectral imaging it is important to note that increased understanding of this technology and awareness of critical use cases in industrial automation make this a “trending” imaging technique as well.

Ongoing advances in camera technology always make the “trending” list. Imaging sensors continue to achieve higher resolutions and imaging rates. These added capabilities need to be supported by interconnect protocols that can achieve higher bandwidth, and the market has responded as we see GigEVision implemented at up to 100Gbps, and interface standards like CameraLink HS and CoaXPress 2.0 also providing increased transfer rates. Other sensor advances introduced in prior years are trending as camera manufacturers take advantage of developments like the Sony SenSwir™ for visible and SWIR imaging and the Depthsense™ time of flight 3D imager. In SWIR imaging also, advances in CQD (colloidal quantum dot) imaging sensors promise to lower costs while increasing resolutions and performance. A recent unique sensor development, the “neuromorphic” or “event” imager has been noticed by the market and is worth mentioning though in industrial automation still in search of broad use cases.

Technology Trends – Software

“AI” (in the form of deep learning) still seems to receive a high amount of attention in the marketplace. Many products promote “AI” in their name or as part of their software offering though it is difficult at times to perceive exactly what the “AI” refers to. General purpose machine vision software and components that use deep learning for differentiation and classification, as in defect detection, continue to see adoption, and in many cases the deep learning is offered as a tool in a broader library or software implementation to further enhance the overall capabilities.

In parallel, embedded computing and embedded vision, particularly when combined with compute platforms that support deep learning, are trending as well. Use cases for embedded vision extend well beyond industrial automation though in particular new smart camera offerings make good use of this technology. As noted earlier also, a distinct trending direction for software is the targeting of specific application use cases with easy-to-configure systems and packaged solutions and we can expect this movement to continue.

Trending Markets And Application Use Cases

One of the most interesting trends in industrial automation has been revealed in recent statistics about robotic adoption by use case and market. Over the past few years, a shift, small but noticeable, has taken place in who is implementing the most robots. Traditionally automotive use cases have dominated non-automotive. However, in the previous year, and potentially in 2022 when final numbers are in, non-automotive use exceeds automotive applications. Overall, the general trend is that non-automotive applications are growing and automotive applications, while strong, are mostly level. What does this have to do with machine vision trends for markets and applications? Historically, machine vision as a technology market has tracked very closely with the robotic marketplace for several reasons. In short, when we want to predict what might be the next direction for machine vision, one place to look is where the robot sales are going.

That said, there are two important markets that likely will be key growth areas in the near future.

  • Precision agriculture and vertical/indoor farming: One of the keys to productivity in agriculture lies in automation of the everyday tasks of farming and reduction of raw materials needed. One use case where the latter is already seen is in automated weeding using vision guidance and laser destruction of the weeds. Precision application of fertilizer in selective spray technology uses these technologies also to reduce costs and chemical usage. In vertical and/or indoor farming robotics and machine vision strive to make the process as near to fully automated as possible.
  • Logistics/warehousing: While perhaps slowing somewhat, the use of machine vision as a standalone technology in logistics is still an important use case. Applications like “track and trace”, package/object identification, and volumetric package analysis remain strong. VGR also continues to be a strong target with AMR (autonomous mobile robots) used prolifically in moving product for order fulfillment.

The Most Important Trend Of All

For anyone who uses, specifies, designs, or develops machine vision systems, solutions, or technologies in industrial automation, the most important trend to remember is “don’t follow trends.” Jumping on some bandwagon and specifying a technology because it is the most recent hype or because you’ve heard that you’ll be left behind if it’s not in your automation plan is a sure recipe for disaster. Evaluate the capabilities and applicability of technology based on the needs, both functional and commercial, of each application and specify the components and solutions that provide a known deliverable and result.