Blog · AI visual inspection
Notes from the inspection line.
Anomaly detection, few-shot training, edge deployment and the metrics that matter: practical writing for solution partners, quality managers and automation engineers.
- 01
The rare-defect problem: why you can never collect enough defect images
Defects are rare by design, so you never collect enough defect images. Good-parts-only anomaly detection trains on about 20 good parts and sidesteps class imbalance.
12 October 2026 · 7 min read
- 02
Total cost of ownership of an AI inspection system (without a price tag)
A framework to reason about the lifetime cost of AI inspection, engineering hours, escapes avoided, maintenance and egress, without quoting a price.
8 October 2026 · 6 min read
- 03
Unsupervised defect detection: flagging a defect it has never seen
How anomaly detection catches a defect it was never trained on: it learns a good part and scores deviation, so first-seen flaws are flagged, not matched.
5 October 2026 · 6 min read
- 04
Adding a new part variant to a running inspection line
When a customer adds a part, how fast can the line inspect it and who does the work? Capture 20 good images, train in under 48 hours, deploy by USB.
1 October 2026 · 6 min read
- 05
Plastic injection quality control with AI: short-shot, flash and colour
How anomaly detection trained on good shots only catches short-shots, flash, sink and colour drift on moulded parts, with millimetre gauging at cycle speed.
28 September 2026 · 7 min read
- 06
Metal stamping and machining inspection: burrs, scratches and gauging
How AI surface inspection plus millimetre dimensional gauging catches burrs, deformation and scratches on stamped and machined metal parts.
24 September 2026 · 6 min read
- 07
Chemical and petrochemical inspection: AI vision for hazardous (ATEX) areas
What a hazardous-area (ATEX) inspection unit needs, and how the AV-X100 (IP66) is positioned for chemical and petrochemical lines. Certification to verify.
21 September 2026 · 6 min read
- 08
Escapes vs false rejects: the two errors that set inspection cost
One threshold trades escapes against false rejects, and the escape usually costs more. The confusion matrix in shop-floor terms, and 0.69% FNR at 99.65% F1 on a live cap line.
17 September 2026 · 7 min read
- 09
Automotive quality inspection with AI: headlight and bumper recognition
How edge-AI vision recognises left vs right headlights and bumper variants on a mixed automotive line, so the robot runs the right program.
14 September 2026 · 7 min read
- 10
Glass inspection and positioning with AI: transparent and coated panels
How an R&D project located near-invisible transparent and coated glass with laser-line and Raman-shift techniques, feeding a handling robot the position in millimetres.
10 September 2026 · 8 min read