Cognex Company’s In-Sight 3900 imaginative and prescient system. | Supply: Cognex
Cognex Corp. this week launched the In-Sight 3900 Imaginative and prescient System. The absolutely built-in imaginative and prescient system is designed to ship pace, accuracy, and backbone on the edge. It allows producers to run demanding inspections with out sacrificing throughput or simplicity, mentioned the corporate.
“The In-Sight 3900 represents a significant leap ahead in what embedded AI imaginative and prescient can ship on the manufacturing facility ground,” acknowledged Matt Moschner, the president and CEO of Cognex. “Producers not have to decide on between inspection depth and line pace. We’ve constructed a system that delivers each, with the reliability and ease Cognex is understood for.”
Cognex constructed its newest system on a brand new technology of its embedded AI imaginative and prescient know-how, powered by Qualcomm Dragonwing platforms. The corporate mentioned the In-Sight 3900 eliminates the standard tradeoff between inspection depth and line pace that legacy imaginative and prescient methods impose.
The system combines Cognex-developed edge AI, superior AI, and rule-based imaginative and prescient instruments with a high-performance embedded compute. With these, it allows deterministic, actual‑time inspections at full line pace, supporting subsequent‑technology packaging, automotive, electronics, and shopper items purposes.
“Our packaging strains run at extraordinarily excessive speeds, which beforehand restricted us to utilizing conventional OCR instruments,” mentioned Andrea Sabbadini, the engineering supervisor at Fuji Seal. “The In‑Sight 3900 now permits us to deploy Cognex’s Edge AI Learn instruments at full manufacturing pace with out compromising throughput. The result’s a extra strong inspection course of that’s sooner to arrange and simpler to take care of throughout our strains.”
Cognex supplies AI on the edge
Key options of the system embrace:
- 4X sooner processing at scale: In-Sight 3900 runs inspections as much as 4 occasions sooner than earlier‑technology Cognex imaginative and prescient methods. It maximizes line charges and eliminates bottlenecks in excessive‑quantity manufacturing.
- Excessive-resolution imaging as much as 25 MP: The system helps resolutions as much as 25 megapixels for wider fields of view, finer measurements, and improved defect detection. It does this all inside a single acquisition.
- Actual-time, high-throughput edge execution: Embedded AI acceleration and optimized processing pipelines allow deterministic, real-time inspection synchronized with high-speed manufacturing strains.
- Industrial-grade connectivity: Twin Ethernet structure ensures dependable communication with PLCs, robots, and enterprise methods.
Cognex mentioned it constructed the In-Sight 3900 on the most recent technology of its embedded AI structure. The platform incorporates a devoted high-performance AI processor for quick, deterministic choice making instantly on the edge.
The In-Sight 3900 helps edge AI instruments for quick deployment and inspection stability, superior AI instruments for complicated, excessive‑variability purposes, and PC‑free operation to simplify validation, deployment, and lifecycle administration.
“Cognex has pushed the boundaries of what embedded AI imaginative and prescient can ship on the edge,” mentioned Shyam Krishnamurthy, senior vice chairman at Qualcomm. “We’re proud to energy this breakthrough with Qualcomm know-how, and we look ahead to what this collaboration will proceed to ship for producers operating demanding inspections at full manufacturing pace.”
When paired with Cognex OneVision, the In-Sight 3900 turns into a part of a scalable, cloud-to-edge AI imaginative and prescient ecosystem. OneVision streamlines mannequin improvement, cross-site collaboration, and deployment throughout units, permitting producers to coach AI fashions centrally whereas executing inspections domestically at manufacturing pace.
The submit Cognex releases absolutely built-in AI-powered imaginative and prescient system for robotics appeared first on The Robotic Report.
