Skip to content
Stage 19 of 42 · Manufacturing

Fabric Inspection

4-point / 10-point, AI camera inspection and shade grouping.

The lesson for this stage
Chapter 15 · Fabric Inspection & Classification

What really happens at this stage

Fabric inspection determines whether a received lot is fit for cutting by scoring visible defects such as holes, slubs, oil marks, weaving faults, and shade variation against an accepted grading system, commonly a 4-point or 10-point scale. Under the 4-point system, defects are scored by length and severity, capped at 4 points per linear yard, and the lot's total defect points per hundred square metres are compared against an agreed acceptance threshold set with the buyer or brand. Inspection is typically done on a percentage sample of rolls per lot rather than every roll, using a defined sampling plan, with escalation to full inspection if the sample fails.

Modern fabric inspection increasingly uses AI-assisted camera systems mounted on inspection frames, which scan fabric at running speed and flag defects automatically by type and location, reducing dependence on inspector fatigue for repetitive fault detection. Alongside defect scoring, shade grouping is carried out by comparing rolls against a physical or digital shade standard under standardised lighting, sorting rolls into shade bands (A, B, C) so that a single garment or bundle is cut from one shade band only. Shade banding is critical because even fabric that passes defect scoring can produce visibly mismatched garments if rolls from different dye lots are mixed within the same style or bundle.

How it is done

  1. 1
    Select the sample rolls

    Draw a defined percentage of rolls per lot per the agreed sampling plan, ensuring rolls from different production batches are represented.

  2. 2
    Mount and run the inspection frame

    Feed each sampled roll through the inspection frame at a controlled speed with adequate lighting, or through an AI camera system if available.

  3. 3
    Score defects by the agreed system

    Record each defect's type, length, and point value under the 4-point or 10-point scale as it is identified.

  4. 4
    Calculate points per hundred square metres

    Total the defect points for the roll and convert to points per 100 square metres to compare against the acceptance limit.

  5. 5
    Group rolls by shade

    Compare each roll to the shade standard under standard lighting and assign it to a shade band before it is released for cutting.

  6. 6
    Accept, hold, or reject the lot

    Pass rolls within the acceptance limit to the cutting store, hold borderline rolls for re-check, and reject or negotiate on rolls exceeding the limit.

Key metrics (indicative targets)

MetricWorking targetWhy it matters
Defect points per hundred square metreswithin the agreed acceptance limit for the orderdirectly predicts the rate of fabric-related garment rejects
Inspection coverageper the agreed sampling percentage of rollsensures the sample is representative of the full lot
Shade banding accuracy100% of released rolls correctly bandedprevents shade mismatch within a single garment or bundle
Inspection turnaround per lotwithin 24-48 hours of receiptdelays here push back cutting and threaten the T&A calendar
False accept/reject rate (AI systems)low single-digit percentagemeasures reliability of automated detection versus manual check

Targets are indicative working ranges, not standard or legal limits.

Control points to check and sign off

  • Sampling plan and acceptance limit are agreed with the buyer before inspection
  • Inspection frame lighting and speed are set to standard conditions
  • Defects are scored consistently using the same grading system throughout the lot
  • Every roll is shade-banded before release to the cutting store
  • Rejected or held rolls are physically segregated and documented

Common pitfalls and their consequences

  • Inspecting at inconsistent lighting or frame speed, producing unreliable defect scores between shifts
  • Mixing rolls from different shade bands into the same cutting bundle, producing visibly mismatched garments
  • Relying solely on AI detection without periodic manual verification, missing defect types the system was not trained on
  • Using a sampling percentage too small for a highly variable lot, letting defective rolls through undetected
  • Releasing borderline rolls to cutting without a documented hold-and-recheck decision, creating unclear accountability for later claims

Main activities

  • Production planning and line balancing
  • Cutting, bundling and sewing operations
  • Printing, embroidery and value-add processes

Quality risks

  • Missed defects → cutting waste

Sustainability risks

  • Fabric write-off

AI opportunities

  • Defect map to cutting

Official sources

Learn the skills used at this stage

Industry chapters that cover this stage

The GARMENT TECH Ed industry chapters whose lessons are applied here — each chapter has a reference page and a taught lesson you can read straight through.

Track your course progress

Saved on this device only

Loading

Study anonymously on this device, or sign in to save across devices. See your certificate.

Questions about this stage

Ask about stage 19, Fabric Inspection. No account needed — questions are published once reviewed, with the answer alongside them.

No questions yet — be the first to ask.

Stay in touch

New chapters, delivered quietly.

A short note when a new story, reflection or milestone is added. No noise, no spam — unsubscribe with a single click.