Scalable Embedded Computing for Automated Vision Workloads

SECO is introducing its end-to-end hardware and software solutions tailored for the robotics, medical, and industrial automation sectors. The integrated technology stack enables original equipment manufacturers to transition from rapid prototyping environments to production-grade deployments with unified lifecycle management.

Hardware Foundations for an Industrial AI Architecture
Demonstrated at embedded world North America 2026, taking place September 22–24 at the Anaheim Convention Center, the SOM-SMARC-Dragonwing-IQ8 illustrates a structured migration path for enterprise hardware. The platform allows engineering teams to shift initial designs tested on the Arduino VENTUNO Q rapid prototyping board directly to a production-oriented architecture utilizing Qualcomm Dragonwing IQ8 Series processors. This pipeline preserves initial software development investments while introducing industrial-grade hardware specifications, robust operating system support, embedded cybersecurity protocols, and fleet management capabilities necessary for large-scale enterprise rollouts.

Concurrent Video Processing at the Edge
Addressing high-density multimodal workloads, a multi-camera security configuration powered by the Qualcomm Dragonwing IQ9 platform demonstrates local processing capacity. The system handles up to 16 concurrent artificial intelligence video streams, merging vision-language models with real-time object detection algorithms. By utilizing the dedicated neural processing and multimedia acceleration blocks native to the IQ8 and IQ9 silicon, the architecture processes compute-intensive parallel tasks locally, reducing latency and bandwidth dependence on centralized cloud servers.

Portable Systems for Remote Field Operations
Targeting mobile enterprise applications, the Intel-based Portable 10.1-inch Edge AI Mission System functions as a rugged, battery-operated processing unit. Operating on the Intel Core Ultra Series 3 (Panther Lake) processor, the terminal supports local machine learning execution in disconnected environments. The system provides the necessary power efficiency and processing density for use cases such as field maintenance operations and isolated digital assistants, where continuous network connectivity is unavailable.

Unified Edge AI Ecosystem and Lifecycle Management
Beyond individual compute nodes, the integration of system-on-modules (SOMs), single board computers (SBCs), and human-machine interfaces (HMIs) with the Clea software framework establishes a standardized device lifecycle ecosystem. This infrastructure supports remote device management, secure over-the-air (OTA) updates, container orchestration, and artificial intelligence workflow deployment.

"Streamlining development from a working concept to something that can be manufactured, reliably deployed, and supported for years has been a major challenge," stated Rodney Feldman, VP Products, Innovation, and Marketing at SECO USA. "By bringing together compute platforms, software, artificial intelligence capabilities, and lifecycle expertise, we facilitate original equipment manufacturers to move faster without having to assemble and manage every part of the technology stack themselves."

Digital Transformation of Mechanical Vending Equipment
Legacy infrastructure applications utilize the Clea Vend software and the KarL4 compact contactless payment terminal to digitize mechanical hardware. The Tower Optical Company binocular viewers incorporate the KarL4 terminal, utilizing its low-power operation and integrated cellular connectivity to process digital-wallet and card transactions. This retrofit capability allows operators of traditional coin-operated mechanical platforms to securely authorize modern digital payments without re-engineering the primary mechanical enclosure.

Additional Context
This section details technical specifications and competitive benchmarking not included in the original news release.

The SMARC (Smart Mobility ARChitecture) form factor utilized by the Dragonwing-IQ8 provides a standardized 82mm x 50mm or 82mm x 80mm footprint, competing directly with COM Express Mini and Qseven modules. SMARC is specifically optimized for low-power ARM and hybrid silicon rather than high-wattage x86 processors, offering superior pin-out density for multimedia and camera interfaces (MIPI-CSI). Qualcomm’s IQ8 and IQ9 series processors benchmark against platforms like the NVIDIA Jetson Orin Nano and Orin NX, competing on tera-operations per second (TOPS) per watt to maximize neural processing unit (NPU) efficiency for parallel video inferencing in thermally constrained enclosures. Furthermore, the Intel Core Ultra Series 3 architecture competes with AMD Ryzen Embedded processors, focusing on hybrid core designs that incorporate dedicated NPU silicon to execute machine learning algorithms natively while maintaining legacy x86 instruction set compatibility. Software frameworks like Clea compete with platforms such as Balena and Azure IoT Edge, differentiating through native hardware-level integrations and specific containerized orchestration targeting neural network lifecycle management.

Edited by Aishwarya Mambet, Induportals Editor, with AI assistance.

www.seco.com