Electronics manufacturer USI has launched an AI Smart Camera designed to run computer vision workloads directly on the factory floor. The system combines high-resolution imaging, a low-power edge computing platform, and AI vision software, allowing image analysis and AI inference to take place locally.

The camera processes visual data at the edge rather than relying solely on centralised computing infrastructure. USI said the architecture is intended to support real-time decisions in manufacturing inspection and automation applications.

Keeping vision processing local

Supported applications for USI’s Smart Camera include object detection, defect identification, optical character recognition, component positioning and tracking, assembly verification, and product classification. USI also lists operation behaviour detection and analysis among the workloads supported by the system.

The hardware includes a high-resolution camera module designed to operate in low-light conditions alongside the edge computing system. Connectivity includes Ethernet, Power over Ethernet, HDMI, and MicroSD, while USI describes the design as ruggedised for integration with industrial equipment, machine vision systems, and automation platforms.

USI has combined the camera hardware with embedded software and tools covering data collection, dataset generation, model training, and deployment. The company said its no-code/low-code AI model training platform is intended to reduce dependence on specialised AI expertise when developing and deploying vision applications.

“Today’s manufacturers are seeking more than just camera hardware,” Justin Chang, director of the Vertical Mobility Solution Center at USI, said.

Machine-vision workloads can place different demands on factory networks than conventional automation traffic. A joint reference architecture from Cisco and Cognex said many production networks were built for deterministic control traffic and basic automation rather than high-bandwidth image data, with higher-resolution cameras and AI inference adding requirements around latency, synchronisation, and bandwidth.

Cisco and Cognex recommend running AI inference close to the image source to reduce latency and limit bandwidth consumption. Their architecture also calls for time-sensitive control and synchronisation traffic to be prioritised separately from high-bandwidth image streams.

USI said the Smart Camera can provide data for identifying process deviations, as well as real-time visual guidance and feedback to robotic systems. The launch announcement does not provide response-time figures or specify whether the system can independently reject defective products or alter production equipment.

Edge inference does not mean all computing moves onto the factory floor. Current AWS manufacturing guidance describes machine-learning models running at the edge for real-time defect detection while factory data can still be sent to cloud services, and Cisco and Cognex note that computing resources for model training can reside in the cloud or separate on-premises infrastructure.

Cisco and Cognex also recommend local inference as a way to keep sensitive image data under local control. They note that industrial vision systems can capture proprietary products, manufacturing processes, or customer information, and recommend selective extraction, filtering, and forwarding rather than indiscriminately transporting raw images.

USI’s launch announcement does not identify the processor used by the edge computing platform, its AI performance, or supported model frameworks. It also does not provide pricing, a general availability date, or benchmarks comparing local inference with a centralised or cloud-based deployment.

USI said the AI Smart Camera has already been deployed within its own manufacturing operations. The announcement does not identify the production sites using the new camera or provide performance figures from those deployments.

AI inspection already in use

USI has previously deployed AI-assisted visual inspection within its manufacturing operations. The company said an automated optical inspection and AI defect-detection system was introduced at its Jinqiao facility in 2023 and at its Zhangjiang facility in 2024.

According to USI, the system uses deep-learning models to identify more than 85% of defect types and improved inspection efficiency by more than 60% compared with processes that relied on human visual rechecks. Those figures relate to the earlier AOI system, not the newly launched Smart Camera.

USI also reports using AI in an automated functional circuit testing station, where it said individual test times fell by about 50%. USI’s Smart Camera announcement does not link the new system’s models or inspection architecture to these earlier deployments.

(Photo by Carl Tronders)

See also: Netmore adds Kinéis satellite connectivity to LPWAN network

Want to learn more about the IoT from industry leaders? Check out IoT Tech Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including AI & Big Data Expo and the Cyber Security Expo. Click here for more information.

IoT News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.