In-Sight® 3900 sets a new standard for high-speed, high-precision inspection at the edge – no external PC required
“The In-Sight 3900 represents a major leap forward in what embedded AI vision can do on the shop floor,” said Matt Moschner, president and CEO of Cognex. “Manufacturers no longer have to choose between inspection depth and line speed. We’ve developed a system that offers both – with the reliability and simplicity Cognex is known for.”
High-speed AI vision without compromise
The In-Sight 3900 eliminates the traditional trade-off between depth of inspection and line speed that traditional vision systems force. By combining Cognex-developed edge AI, advanced AI and rule-based vision tools with a powerful embedded computing engine, the system enables deterministic, real-time inspections at full line speed, supporting next-generation applications in packaging, automotive, electronics and consumer goods.
“Our packaging lines run at extremely high speeds, which previously limited us to using traditional OCR tools,” said Andrea Sabbadini, Engineering Manager at Fuji Seal. “With the In-Sight 3900, we can now use Cognex’s Edge AI read tools at full production speed without compromising throughput. The result is a more robust inspection process that is faster to set up and easier to maintain across our lines.”
Key features
- 4 times faster processing at scale: Performs inspections up to four times faster than previous generation Cognex vision systems – maximizing line speeds and eliminating bottlenecks in mass production.
- High-resolution imaging up to 25 MP: Supports resolutions up to 25 megapixels for wider fields of view, finer measurements, and improved defect detection – all in a single shot.
- High-throughput, real-time execution at the edge: Built-in AI acceleration and optimized processing pipelines enable deterministic, real-time inspections that are synchronized with high-speed production lines.
- Industrial-grade connectivity: Dual Ethernet architecture ensures reliable communication with PLCs, robots and enterprise systems.
Designed for advanced AI at the edge
The In-Sight 3900 is based on the latest generation of Cognex embedded AI architecture and features a dedicated high-performance AI processor for fast, deterministic decision making right at the edge. The platform supports:
- Edge AI tools for rapid deployment and inspection stability.
- Advanced AI tools for complex applications with high variability.
- PC-independent operation to simplify validation, deployment and lifecycle management.
“Cognex has pushed the boundaries of what embedded AI image processing can do at the edge,” said Shyam Krishnamurthy, senior vice president at Qualcomm. “We are proud to drive this breakthrough with Qualcomm technology and look forward to what this collaboration will bring to manufacturers performing demanding inspections at full production speed.”
Part of a cloud-to-edge AI vision ecosystem
In combination with Cognex OneVision™, the In-Sight 3900 becomes part of a scalable cloud-to-edge AI vision ecosystem. OneVision streamlines model development, cross-site collaboration, and deployment to multiple devices, allowing manufacturers to train AI models centrally while performing inspections locally at production speed.
About Cognex Corporation
For over 40 years, Cognex has made advanced machine vision simple, paving the way for manufacturing and distribution companies to become faster, smarter, and more efficient through automation. Innovative technology in our vision sensors and systems solves critical manufacturing and distribution challenges, delivering unparalleled performance for industries ranging from automotive to consumer electronics to consumer goods. With a long-standing focus on AI, Cognex makes these tools more powerful and easier to deploy, helping factories and warehouses improve quality and maximize efficiency without the need for highly specialized technical expertise. Headquartered near Boston, USA, we have locations in over 30 countries and more than 30,000 customers worldwide
– – – – – –
Further links
👉 www.cognex.com
Photos: Cognex
