Modular Vision MX8M-Plus: using an NPU for machine learning in an HMI. Suggested use case: HMI on an industrial automation machine. a camera performing AOI (automated optical inspection) brings images of a product just assembled into the HMI via Ethernet or USB. The NPU includes an algorithm that analyzes the incoming image for critical parameters (color, shape, relative positions and lengths for example). The HMI therefore provides a user interface, edge AI, and IoT to deal with the results of AOI.
Smarter Quality Control: Merging AOI with HMI
Automated optical inspection (AOI) has long been a cornerstone of industrial automation, particularly in electronics manufacturing. But AOI systems can present major integration headaches due to their use of highly-specialized hardware needed to keep up with the massive computation demands of computer vision.
One promising solution is to migrate AOI algorithms from specialized hardware to the general-purpose computers running the human machine interface (HMI). HMI devices, also called Panel PCs, incorporate a touch display with an embedded computing platform and peripheral interfaces, into a single enclosure suitable for deployment in industrial environments. The processors in modern HMI systems have extraordinary performance, including built-in artificial intelligence (AI) accelerators that can help take AOI to the next level. This is a convenient option because HMI and AOI systems usually work in concert, with human operators using an HMI to train the AOI system, monitor its operation, and intervene as needed.
Benefits of Integrating AOI and HMI
Today’s manufacturing requires seamless cooperation between multiple machines, control systems, and software solutions to maintain efficient production flow. Using the HMI as the basis for this integration considerably simplifies matters.
In this setup, AOI cameras bring images of assembled products while still on the assembly line to the HMI, addressing several challenges:
- By coupling the AOI and HMI, technicians can train the AOI directly on the manufacturing line. This leads to better results, as it accounts for factors like the lighting conditions on the factory floor.
- Systems must adapt quickly to newly introduced products or production changes without causing long changeover times. Having the HMI and AOI integrated makes this easier.
- AOI systems can produce false negatives (missing faulty components) or false positives (marking good components as defective), necessitating rechecks and manual intervention—a task that requires an HMI.
An intuitive interface facilitates all of these tasks. Thus, it is critical to look for a modern HMI with a large, high-quality display.
In addition to technical benefits, combination of the AOI and HMI computing platforms reduces cost of equipment purchase and ongoing maintenance.
Manufacturing Execution Systems and IIoT Integration
Modern HMI hardware can also supply the connectivity needed to integrate AOI with manufacturing execution systems (MES) and Industrial Internet of Things (IIoT) platforms. This integration enables manufacturers to:
- Track quality metrics across the entire production line.
- Identify trends that could indicate emerging issues.
- Share inspection data with other manufacturing systems.
- Deploy updates to inspection algorithms across multiple systems.
- Monitor and optimize quality control processes in real-time.
These capabilities take a combined HMI + AOI system beyond a mere checkpoint on the production line. They turn an inspection point into a critical element of the overall smart factory.
SECO Solution with SMARC Module and AI Processing
To help companies enhance their production environments with smarter AOI, SECO has created the Modular Vision 10.1 MX8M-Plus. This all-in-one HMI combines features including:
- Ruggedized 10.1” touchscreen with 400 cd/m² brightness for easy in-factory readability.
- An NXP i.MX8M Plus microprocessor with a neural processing unit (NPU) that delivers 2.3 TOPS of AI performance.
- A high-performance GPU for smooth user interfaces.
- Gigabit Ethernet, USB-C, and USB 2.0 ports (among other interfaces).
- RS-232, RS-485, GPIO, and other interfaces for connection to collocated equipment and sensors.
- Four mounting options (panel, flush, rear, and VESA mounts) for maximum flexibility to incorporate within industrial machinery.
With this hardware, the Modular Vision 10.1 MX8M-Plus can connect to a camera via USB or Ethernet, process the camera data with its NPU (which facilitates image operations such as feature extraction and identification), and display the results on a high-quality screen.
The platform is powered by Clea OS, a Yocto Linux-based operating system that supports cross-platform consistency, enhanced security, and seamless integration into IoT networks. Clea OS includes the libraries needed to facilitate edge IoT data orchestration (collection and processing) for minimized latency and subsequent communication to the cloud for enterprise-wide use. This includes the deployment of AI-based algorithms. The operating system also includes a full-featured device manager and platform security functions. Clea OS is part of the Clea ecosystem that incorporates a number of tools for device management, data orchestration, and cloud connectivity.
Conclusion
Modern production environments require intelligent manufacturing systems, particularly for tasks like AOI. Success depends on high integration density and innovative AI algorithms. SECO addresses these needs by providing complete systems that combine embedded processing, touch display, camera and industrial equipment connectivity, an IoT-ready edge Linux operating system, and an enterprise-grade IoT software suite – for rapid deployment into industrial environments.
Ready to enhance your production line with AI-powered inspection? Contact us today.