- By Axiomtek
- October 04, 2024
- Axiomtek
- Opinion
Summary
Artificial intelligence (AI) is pushing the capabilities of machine vision systems to supercharge speed, efficiency and accuracy.
Automation is progressing rapidly with Industry 4.0 initiatives aiming to improve quality, efficiency and productivity across various industries, including manufacturing, food and beverage, logistics, healthcare and more. For machine vision systems, that means enhanced data from industrial Internet of Things (IIoT) sensors and innovative high-resolution camera technology that can capture a wide range of information to help overcome issues related to real-world complexities. However, the most significant Industry 4.0 technology is artificial intelligence (AI), pushing the capabilities of machine vision systems to supercharge speed, efficiency and accuracy.
Machine vision systems empowered by edge AI can quickly and easily analyze images in real-time to recognize subtle nuances and patterns, compare patterns across an entire data set of images and retain information from each analysis to learn and improve accuracy over time continuously.
With this in mind, integrating new technologies into existing manufacturing and assembly lines comes with challenges. System integrators must choose the right hardware that can support a wide range of complex components, diverse connectivity options and the need to support current and future AI-powered systems with the processing power to analyze large quantities of data.
To meet these demands, advanced AI machine vision capabilities require industrial computers with a variety of characteristics.
High Processing power. Edge computers should have the flexibility to support a wide range of CPU and GPU requirements to meet a diverse range of future workloads. To help future-proof, systems should support the latest CPUs, such as twelfth, 12th, 13th and14th generation Intel Core processors. The latest DDR memory with high bandwidth is needed to support faster transfer of stored data for real-time analysis
Comprehensive I/O support. A wide variety of interfaces for supporting the latest devices is a key feature to look for in a machine vision edge computer. Key features include chipsets with digital interfaces for high-speed industrial cameras and I/O support for audio systems, KVM devices, serial communications (RS232/422/485), displays and external platforms.
Secure high-throughput networking. Every edge computer should feature secure, high-bandwidth connections to both internal and external networks via multi-gigabit Ethernet LAN ports. They should support 5G wireless connectivity for the increasing number of industrial sensors and devices used in remote monitoring, mobile robots, autonomous vehicles, asset tracking, AR, VR and digital twins.
Intelligent PoE device management. Edge computers should ensure safe power supply to components through DC input power while providing intelligent power management, including managing and monitoring power consumption per port for remote power distribution technologies like USB and Power over Ethernet (PoE) for connected devices, including cameras, lights and sensors.
Scalability. As AI technology evolves, edge computers should offer flexible expansion options and optional modules with additional I/O ports. This allows for easily scaling systems to accommodate current and future machine vision technologies, extending the system’s lifespan.
Industry compliance and durability. Edge computers deployed near assembly lines must be robust and designed to withstand the harsh realities of industrial and manufacturing environments, including exposure to shock, vibration, extreme temperatures, humidity, electromagnetic interference, dust and debris. For added assurance, any industrial edge hardware should comply with the latest performance and safety standards applicable to electronic equipment intended for use in industrial environments. This includes wide-operating temperature ranges, IEC/EN 61000-6-2 and 61000-6-4 EMC certifications and IEC 60068-2-27 and 60068-2-64 certifications for shock and vibration resistance.
Customization options
Since each machine vision application and environment has unique requirements, off-the-shelf systems may not always meet specific needs. System integrators should consider edge computers backed by design and integration services to meet precise project specifications. These services can customize solutions to unique requirements, allowing the flexibility for add-on peripherals such as encoder cards with real-time trigger I/O to support cameras in targeted machine vision applications.
To fulfill these demands, Axiomtek recommends an industrial-certified computing system targeted toward AI-powered machine vision applications. The ideal standardized system for machine vision applications should be scalable and modular, allowing for a wide range of system configurations and expandable to meet the requirements of various scenarios. The ability to support AI accelerators allows these systems to effectively handle computational demands for real-time decision-making in applications that require fast and accurate processing at the edge.
This piece was originally published in AUTOMATION 2024: 9th Annual Industrial Automation & Control Trends Report.
About The Author
Axiomtek, a provider of industrial PCs, is committed to advancing machine vision applications and emerging AI technologies. Our U.S.-based design engineering and integration services, alongside our DigiHub of SDKs and development resources, empower system integrators to deploy edge machine vision systems for any application.
Did you enjoy this great article?
Check out our free e-newsletters to read more great articles..
Subscribe