360Labs, home

48 / Hardware & IoT

AI Camera Defect Inspection for Manufactured Panels

We built a camera inspector that checks manufactured panels for defects at full line speed, scoring every panel on the device itself so production imagery stays on-site.

Built by
360 Labs (Internal)
Stack & approach
Computer VisionDefect DetectionEdge QALine SpeedOn-Prem
Interface screens from AI Camera Defect Inspection for Manufactured Panels
Interface, from the 360Labs case-study deck (page 59).

The problem

Panel QA at line speed needs inspection faster than manual review, and production imagery stays on-premise.

Solution & architecture

A vision model runs QC on the line itself, scoring every panel on-device without moving data off-site. On-device defect detection covers the full panel surface and keeps pace with production throughput, with all imagery processed locally.

Why it is built this way

  • On-device scoring

    Every panel judged locally

  • Keeps line speed

    Matches production throughput

  • Imagery stays on-site

    Processed locally, no egress