A recent customer survey conducted by Rockwell Automation reveals that enhancing product quality is perceived as the foremost catalyst for the digital transition in manufacturing. Additionally, respondents identified that the primary area where artificial intelligence (AI) exerts the most significant influence is in closed-loop quality control systems.
At the highly anticipated Automation Fair 2024, Amanda Thompson, the product manager for FactoryTalk Analytics VisionAI at Rockwell Automation, delivered an insightful session dedicated to the latest version of VisionAI, which debuted in September 2024. This fresh iteration of VisionAI specifically targets the quality and AI capabilities that Rockwell Automation’s customers prioritize.
Thompson elaborated that, although VisionAI incorporates advanced machine vision technology, it transcends traditional machine vision systems. “This is a quality inspection platform designed to help you understand the quality of the product you’re producing,” she emphasized. “It not only tells you if a part or product is good or bad, it also elucidates the reasons behind it.”
The system is engineered to facilitate vision inspection at impressive line speeds, accommodating between 500 to 600 parts per minute, subject to the specific application. It adeptly performs functions such as reading barcodes, classifying parts for identification and sorting, detecting surface defects, conducting presence/absence assessments, and verifying text accuracy.
Beyond strict inspection functions, VisionAI offers robust traceability features, allowing users to access images flagged by the system as defective. This drives essential comparisons against acceptable images and provides irrefutable proof of specific defects. “VisionAI encompasses root cause analysis capabilities that reveal why an item failed, in addition to providing timely notifications so that detected issues can be remedied promptly,” Thompson noted.
In alignment with Rockwell Automation’s commitment to developing AI-powered technologies that adapt based on user domain expertise, Thompson stated that VisionAI is designed as a no-code platform tailored for operations and quality control personnel. It “relies on your expertise to provide the information needed for the system to make good quality decisions,” she added.
VisionAI architecture and operation
The architecture of VisionAI extends from the cloud to edge devices, seamlessly integrating embedded analytics, data storage solutions, and application programming interfaces for synchronization with manufacturing execution systems and enterprise resource planning systems. The cloud is tasked with hosting the AI engine and offering remote access capabilities for monitoring, management, and deployment of VisionAI’s edge functionalities. At the edge, users engage with local hardware components such as edge computers, human-machine interfaces, PLC integration, alongside the vision system cameras, lenses, and lighting apparatus.
Thompson noted that VisionAI currently supports Basler cameras, with plans to announce compatibility with additional third-party vision systems in 2025.
Detailing the operational mechanism of VisionAI, Thompson explained that the initial step requires the user to “define the type of inspection they want VisionAI to run.” Following this, they can begin capturing images and classifying them as good or bad. The quantity of images necessary for system training relies on the complexity of the inspection task.
VisionAI generates a chart for image training, indicating to the user when a sufficient number of images have been provided for effective model training. A corresponding training report is produced by VisionAI to highlight model accuracy during its development phase, pinpointing any mislabeled images, thus allowing users to rectify them and enhance overall accuracy prior to deployment.
Once the cloud-based model has been successfully trained, it is deployed to the edge where the system can manage up to eight cameras simultaneously, with live camera feed capabilities available both onsite and remotely.
“The defect carousel feature showcases failed images, enabling users to double-click on these images to gain deeper insights into the reasons behind the failures,” she further explained. “VisionAI also consolidates data across different systems and time periods, allowing its dashboard to support comparisons of batches, stations, production days, and additional variables to assist in precise troubleshooting endeavors.”
VisionAI autonomously generates quality reports, ensuring that every stakeholder receives consistent and accurate information.
Industry applications
Thompson identified several industry applications where VisionAI exhibits substantial effectiveness, including:
Packaging. VisionAI can conduct a variety of packaging inspections, including evaluating bottles for defects, examining expiration dates, and verifying label time stamps. It also assesses caps for proper twisting and potential leaks. Thompson highlighted that all these inspections can occur concurrently, yielding a single comprehensive quality result for the product. VisionAI operates on a recipe concept. “If you are running different shampoo brands on a line, and you need to implement various inspection criteria, you can tailor an inspection recipe for each product or SKU,” she elaborated.
Food and Beverage. VisionAI’s capacity to identify issues extends even over complex backgrounds. For instance, it can be employed to inspect cereal as it moves along a conveyor, ensuring correct shapes and colors while flagging imperfections or foreign contaminants. Moreover, it evaluates the placement of toppings on products, identifying discrepancies such as inadequate topping application on bagels or doughnuts.
Automotive. In automotive applications, VisionAI can verify the presence and proper placement of required labels, check for any wrinkles or bunching on car seat covers, and pinpoint specific locations on the seat where these imperfections occur for easier physical assessment.
What role does root cause analysis play in VisionAI’s ability to improve manufacturing processes?
**Interview with Amanda Thompson, Product Manager for FactoryTalk Analytics VisionAI at Rockwell Automation**
**Interviewer:** Thank you for joining us today, Amanda. Let’s dive into the recent customer survey conducted by Rockwell Automation that indicated enhancing product quality is the primary catalyst for the digital transition in manufacturing. How do you see this trend influencing the development of technologies like VisionAI?
**Amanda Thompson:** Thank you for having me! The insights from the survey reflect a significant shift in the manufacturing landscape. As companies increasingly prioritize product quality, they are turning to AI-driven solutions like VisionAI to facilitate this transition. Our platform is specifically designed to automate quality inspections and provide detailed analyses that help manufacturers understand the root causes of defects. This capability is essential for achieving a robust quality control system that goes beyond mere pass/fail assessments.
**Interviewer:** You mentioned that the latest version of VisionAI debuted in September 2024 and focuses on quality and AI capabilities. What are some key features that set this system apart from traditional machine vision technology?
**Amanda Thompson:** VisionAI incorporates advanced machine vision but is much more than just an inspection tool. It not only identifies whether a product is acceptable or defective but also elucidates the reasons behind each assessment. This makes it a powerful platform for closed-loop quality control. For example, VisionAI operates at impressive line speeds—between 500 and 600 parts per minute—and performs a variety of functions, from reading barcodes to detecting surface defects and verifying text accuracy. Its traceability features allow users to access flagged images, enhancing accountability and comparison against standards.
**Interviewer:** That sounds incredibly efficient! Can you elaborate on how VisionAI’s root cause analysis capabilities work and their importance in the manufacturing process?
**Amanda Thompson:** Absolutely. One of the standout features of VisionAI is its ability to conduct root cause analysis. When a defect is detected, the system provides timely notifications and insights into why that failure occurred. This timely feedback is crucial for manufacturers, as it enables them to take corrective action before issues escalate further down the production line. Being able to trace a defect back to its source allows for continuous improvement and the adjustment of processes and quality standards.
**Interviewer:** You also mentioned that VisionAI is designed as a no-code solution. How does this impact usability for operations and quality control personnel?
**Amanda Thompson:** The no-code aspect of VisionAI is a game-changer for many organizations. It empowers operations and quality control personnel—who may not have extensive technical backgrounds—to define their inspection criteria and manage the system effectively. Users are able to leverage their domain expertise without needing to write complex code, thus making the technology more accessible and scalable across teams.
**Interviewer:** It’s fascinating to see how technology is evolving in manufacturing. Can you provide some insights into how VisionAI integrates with existing systems?
**Amanda Thompson:** Sure! VisionAI adopts a cloud-to-edge architecture, meaning it integrates seamlessly with existing manufacturing execution systems and enterprise resource planning systems. The cloud hosts the AI engine, while edge devices manage local operations. Users can connect various hardware components like cameras and interfaces easily. Our current compatibility with Basler cameras, with plans to support more third-party systems, demonstrates our commitment to offering versatile solutions.
**Interviewer:** what does the future hold for VisionAI and its capabilities in the manufacturing sector?
**Amanda Thompson:** Looking ahead, we’re excited about expanding VisionAI’s capabilities even further. As we announce support for additional vision systems and continuously improve our AI models, we anticipate seeing even greater adoption of our technology across different manufacturing sectors. Our goal is to help manufacturers not only enhance product quality but also drive efficiencies and insights that can transform entire operations.
**Interviewer:** Thank you for your insights, Amanda! It’s clear that VisionAI is at the forefront of the digital transition in manufacturing.
**Amanda Thompson:** Thank you for having me! I’m excited to see how these innovations will shape the future of manufacturing.
