What the Human Eye Can’t See: Inline Thermal Inspection with Eigen Innovations and Advantech

Partnership enables manufacturers to leverage AI-based, Eigen-powered thermal vision inspection utilizing Advantech high-performance box IPCs

Temperature can play a vital role in many of the processes used to produce of goods across diverse industries like automotive, appliance, aerospace, construction, and medical device manufacturing. Whether it’s steel grate or plastic welding, injection molding, drywall production, or adhesive and coating applications, precise temperature monitoring on the factory floor is essential to guarantee products meet quality standards.

Temperature fluctuations and other process variations on production lines can have far-reaching consequences. More than just impacting product quality, they can trigger a cascade of operational issues and negatively affect a company’s bottom line. Without real-time monitoring and defect detection solutions, manufacturers expose themselves to product recalls, returns, equipment damage, downtime, and increased waste—all of which can erode customer trust and profitability.

Historically, manufacturers relied on end-of-line human visual inspections to monitor and test the quality of parts and products; however, this approach has significant limitations. It’s not only time-consuming and costly, but also prone to errors. Furthermore, companies that depend solely on manual checks lack a standardized, digital record of inspections and data.

Time to go thermal

Eigen Innovations and Advantech have partnered to provide a streamlined solution for AI-driven thermal quality inspection, making it more attainable for manufacturers. The end-to-end system supplies critical data and tools needed for comprehensive inline inspection of parts and products—all with compatibility for integration into existing manufacturing infrastructure. Powerful AI and machine learning capabilities not only facilitate process intelligence beyond conventional visual inspection solutions, but also allow users to develop advanced inspection algorithms that continuously learn and adapt.

Eigen empowers manufacturers to build, manage, and integrate machine vision inline inspection solutions. As the sole provider combining non-contact temperature measurement, imaging, and AI for inline quality inspection, Eigen sets a new standard and distinguishes itself from other machine vision inspection providers by enabling teams to swiftly pinpoint root cause of quality issues.

With a dynamic AI approach, Eigen’s solutions help improve accuracy over time, reducing false positives and enhancing defect detection capabilities. The result is an inline thermal inspection solution that identifies issues and predicts and prevents future defects, leading to significant cost savings and improved product quality.

Utilizing AI model training capabilities, Eigen OneView users can leverage images enhanced by algorithmically-generated labels (refined with human validation) to develop advanced AI models. Factory teams then have the flexibility to either maintain a hybrid strategy or fully adopt supervised AI models. This heightened transparency and adaptability bring immense benefits to manufacturers and end customers.

Benefits of Thermal Imaging for AI-powered Datasets

Industrial edge hardware

The solution relies on Advantech’s industrial edge hardware to link factory floor assets and gather data with precision. The Intel-based MIC-770 series is a powerful and dependable embedded IPC (OneView Edge) that facilitates connections to cameras, PLCs, and cloud services.

With the provision of multiple I/O for device connectivity, MIC-770 systems serve as a data gateway and industrial controller. OneView Edge serves as a bridge between the network edge and the cloud, enabling the transmission of image and process data from production line equipment. The MIC-770 processes the image and machine data on the factory floor or from connected cameras, and sends the standardized, consolidated data to the Eigen cloud platform. The AI inspection models are trained in the cloud and deployed to the MIC-770, ensuring ongoing inspection even if network connectivity is lost.

It is ideal for AI-based edge computing and inference, designed for use in harsh factory environments, and available for global deployment in enterprise, cloud, and edge applications with limited space. By combining edge devices like the MIC-770 with Eigen vision solutions and OneView software, AI and machine learning can streamline and automate factory processes, leading to more efficient and dependable operations. This powerful combination also unlocks data-driven decision-making, revealing new insights and business intelligence.

With access to insights like production trends, defect reports, and root cause analyses, manufacturers can view factory processes in a new light. Beyond just preventing defects, the detailed data can reveal new opportunities to improve overall operations. Production and quality teams can have new confidence in the quality of every end product, backed by the assurance of standardized, thorough digital inspections.

How it works

Inspection Solution Architecture

Solution Use Cases

Ensuring plastic weld integrity to prevent leaks in plastic gas tanks

Thermal images and process data from the welder are fed to an OneView Edge embedded IPC running Eigen’s OneView software. The software consolidates all captured data and generates standardized weld-over-weld data required to generate an AI model to inspect each weld inline.

Detecting weld integrity issues in steel production

Eigen, in collaboration with cloud partner Amazon Web Services (AWS), developed a vision solution to address Nucor’s inspection challenges. The system captures images and data to facilitate automated inline weld inspection. OneView software, running on an industrial PC, combines thermal images of the grate welds with process data from the welding unit. The Eigen solution creates a standardized visual record for every weld, enhancing quality control and traceability.

Leveraging the standardized data and tools within OneView, the quality team is able to oversee a machine-learning algorithm that actively monitors for heat signatures and identifies quality defects in real time.

Avoiding quality-related complications in specialty paper manufacturing

Sappi was facing rising quality-related costs due to downtime and equipment damage on its specialty papers lines. The swift accumulation of coatings in specialty paper production was causing quality problems, including streaks. The company’s existing optical machine vision system struggled to keep up, failing to detect moisture changes on the production line that led to these defects.

To address these issues, the Eigen team helped to implement a thermal-based vision solution utilizing thermal cameras to detect temperature variations indicative of streaks and coating build-up. The vision solution helps monitor production 24/7 and alerts operators to potential issues.

For real-time monitoring, production and quality teams integrated the camera with Eigen’s OneView software. The system is interlocked to automatically unload calendars when significant build-up events occur. Beyond reducing downtime, this closed-loop automation solution also enabled standardized record and data access. Additionally, the company saw more than $1M in savings across three sites by implementing this proactive approach.

Bringing thermal imaging and AI into manufacturing is about more than implementing new technology—it is revolutionizing inline quality inspections—especially in thermal manufacturing processes. By reducing or eliminating manual checks, manufacturers can streamline processes and let team members focus on what truly counts. It’s about building better products, smarter.