QualityNew

AI fabric inspection on knitting machines that stops the machine

Added 2026-10-14

Role

quality manager, knitting supervisor

The problem

In circular knitting, fabric is inspected only after a batch comes off the machine, so defects are found after long lengths have been produced. Yarn, dye and energy are wasted on unusable fabric, recurring errors repeat, and tired operators miss subtle defects or grade inconsistently.

Step by step

1. Measure defects per roll and waste by defect type for your top fabrics.
2. Pilot sensors on a few machines running a high-volume fabric.
3. Define which defects are critical and the recurrence threshold that stops a machine.
4. Require supervisor sign-off with root cause before restart.
5. Use defect data by machine, yarn lot and setting to fix causes.
6. Compare defects and waste on equipped versus non-equipped machines.

Tools

• Smartex.AI: used at Yeşim Group
• Other fabric-inspection systems: Cognex/Keyence vision, EVS, Shelton

As reported by the source; most figures are vendor or self-reported and not independently verified.

Human review: cautions

• Figure from an awards profile based on company input, for one fabric type.
• Auto-stop rules that are too tight will hurt output; tune thresholds.
• Retrain the model for new fabrics and yarns.

You are a textile quality engineer. I will paste defect records (date, machine, fabric, yarn lot, defect type, length affected).
1. Rank defect types by waste.
2. Find patterns by machine, yarn lot or shift.
3. Propose stop rules (which defect, how many in what length) and the checks a supervisor should do.
4. Suggest root causes to investigate.

Sources

• https://www.just-style.com/featured-company/yesim-group-ai-inspection-technology-innovation/