Earlier this year, market research firm Gartner forecasted that by 2027 more than half of warehouse operations will use artificial intelligence (AI)-enabled vision systems. These systems will replace traditional scanning-based cycle-counting processes and lead to improvements in efficiency and performance.
This ramp up in machine vision will begin in 2025 as it moves from just a specialized tool used for specific operations to the pillar of intelligent systems. Advancements in hardware, software and integration technologies like AI, real-time processing and internet of things (IoT) integration promises to bring machine vision to new heights this year.
“Companies are looking beyond traditional solutions as pressures mount to continuously improve operational process performance,” said Carly West, senior director analyst in the Gartner Supply Chain Practice, in a statement. “AI-enabled vision systems will propagate quickly in warehouse operations as the value proposition is so evident; not only for inventory management, but also monitoring that can identify safety issues and ergonomic problems for workers in real time.”
According to market research firm Interact Analysis, the machine vision market will grow from $6.5 billion in 2022 to $9.3 billion in 2028. The market will have a compound annual growth rate of 6.4%.
AI dominance
Two of the main benefits to AI-enabled machine vision are deep learning for complex tasks and transfer learning.
AI-enabled vision systems combine industrial 3D cameras, computer vision software and AI pattern recognition technologies and machine learning to radically change how many industrial processes are performed, according to data from system vendor Cognex. Instead of humans doing the manual inputs, AI will be able to dominate machine vision tasks such as:
- Anomaly detection
- Object recognition
- Predictive quality control
Transfer learning pertains to pre-trained AI models that enable rapid customizations for use cases that might be more specialized than general warehouse operations. This will lead to a reduction in development times.
How AI will be used
In these AI-enabled machine vision systems will be used in three ways:
- Explainable AI
- Deep learning
- Generalized AI models
This rise in explainable machine vision systems will allow users to understand and, maybe more importantly, trust decisions made by AI. This will be particularly essential for industries like healthcare and automotive where safety is the top priority.
Deep learning for tasks like pattern recognition, defect detection and complex decision-making will be a pinnacle use for machine visions systems in industrial imaging.
Finally, generalized AI models will be developed to be capable of learning new tasks with minimal retraining. These pre-trained models will enable rapid customization for specific use cases as well as reduce development times.
Inspection dominates
Not surprisingly, the largest application for machine vision in 2025 will remain inspection. By 2028, it is expected that inspection will generate around $3.9 billion, or 42% of the total machine vision market, according to Interact Analysis.
Inspection particularly needs machine vision to ensure product quality, operational efficiency and cost-effectiveness. Specifically, the top inspection areas include:
- Detection of imperfections
- Measurement
- Validation and confirmation of operations
- Positional accuracy and robotics with vision
- Product tracking and verification
In detection of imperfections, AI-enabled algorithms in these systems analyze data on-site and determine if defects are present. This can take just minutes with the results being transmitted to operations center to review the data. Critical defects are flagged by AI for an immediate response.
Additionally, AI-enabled machine vision will boost efficiency by allowing companies to free up humans to do more complex tasks while AI does more of the mundane jobs.
Quality control
Another significant use case for AI-enabled machine vision is in quality control where manufacturers use the technology to help with increased regulatory requirements, new manufacturing techniques and labor shortages, according to market research firm ABI Research.
“AI is accelerating and improving the efficiency of the MV market,” said James Prestwood, indsturial and manufacturing industry analyst at ABI Research. “It increases inspection speeds and enables the movement of quality upstream, and AI systems are more adaptable than traditional software. However, although many AI solutions can easily integrate with existing MV hardware and software, making it a low-hanging fruit for manufacturers to leverage, its lack of explainability can be challenging.”
Without this functionality, AI could struggle to gain traction in some of these markets, Prestwood said.
Challenges
And that challenge is just one of many that face AI-based machine vision.
There will likely be issues with finding enough skilled professionals in AI to meet demand. This may lead to a slowdown in machine vision adoption.
Additionally, the complexity of combining vision systems with existing workflows and IoT networks will remain a challenge for engineers and designers.
There will also be regulatory and data privacy issues in the realm of AI-enabled machine vision as companies voice concerns about data being funneled through a deep learning algorithm.
Future use cases
While inspection will remain the No. 1 use case for machine vision, looking ahead, the technology is likely to play an increasingly significant role in:
- Waste reduction
- Personalized focused-vision applications
- Vision-guided cobots
- Multispectral imaging
- Real-time vision
- Embedded vision in edge computing
- 3D vision and sensing
In-depth articles by Electronics360 on these subjects will likely be covered throughout 2025 and beyond.
