Cotton bale contamination is the quiet claim the spinning mill absorbs on behalf of the merchant it bought the bale from. Plastic twine, jute strapping, colored cloth, and bird-nest fiber ride into the mill inside the bale, and by the time a knitter or fabric buyer files a foreign-fiber claim, the merchant’s name is long gone from the trace. Every one of those claims lands back on the spinner’s P&L, not the merchant’s.
iFactory / Cotton bale contamination vision inspection
Catch Plastic, Jute, and Colored Cloth Before the Blowroom Opens the Lot
A camera at the bale opener that classifies contaminants by type, holds the mixing lot before carding, and writes the reason back to SAP QM — not to a WhatsApp thread.
At the opener
not at the winder
PP / jute / cloth
classified by type
SAP QM QN
on the bale batch
Mixing hold
before carding
The Problem on the Line
Plastic twine, jute strapping, colored cloth strips, and bird-nest fiber ride into the mill packed inside the bale. Once the blowroom opener has fed a bale into the mixing line, the contamination has already been dispersed across a mixing lot of thirty to forty bales — and any fragment that survives the fine cleaner is now a fragment your yarn clearer will chase during winding, or a dark speck the dyehouse will find weeks later. The CORDIS project at one European mill counted roughly 1.5 to 3.5 contamination pieces per hour on bale surfaces, with colored PP and coloured fibres dominating. What escapes the opener is what a spinner pays for later in claims.
What Escapes Into the System of Record Today
The mill knows contamination is happening. What it does not have is a record on the bale batch that a downstream inspector, a QA head, or the merchant's account manager can actually trust.
Paper
Blowroom operator marks a tally on a shift sheet — "7 PP, 3 jute today" — that never leaves the log book above the opener.
WhatsApp
Contamination photo goes to the shift supervisor's phone. The bale ID gets typed in later, or not at all, and the message scrolls out of the group by Monday.
Winder-only data
The only trace in a corporate system arrives four days later as elevated foreign-fiber cuts on the Uster Quantum yarn clearer — an outcome measurement, not a bale-batch signal a merchant can be back-charged against.
Nowhere at all
The bird-nest fiber wound into a colored cloth strip, ejected manually by an operator into a scrap bin, has no digital record at all — no photo, no bale ID, no batch link.
What the Camera or Agent Actually Sees
A vision system at the opener sees what a fine cleaner will not distinguish and what a yarn clearer only measures after the damage is spread across a mixing lot. The point is not to replace the Uster Jossi Vision Shield if you already run one; it is to add a bale-batch record before contamination is dispersed.
Polypropylene
Transparent, glossy, or optically-brightened PP from woven bale wrap and strapping — the category Uster ships VTECT UV-LED specifically for because visible-light detection alone misses it.
Class: PP-clear / PP-colored
Jute & Cordage
Jute fibers and cordage strands from packaging, coarse and distinguishable by fiber morphology from cotton at the opener chute.
Class: jute / cordage
Colored Cloth
Fragments of colored fabric, tape, or rags from harvest or transport, dyed differently from cotton and separable by color-signature at the opener camera.
Class: cloth-colored
Bird-Nest Fiber
Foreign fibers already tangled with cotton fibers into a nested clump — visible as a density and shape anomaly against clean lint on the feed table.
Class: FF-nest
What Gets Written Back
A photo and a class label are the observation. The value is in the SAP QM record that ties the contaminant to the bale batch, the merchant, and the mixing lot that has to be held before it becomes cross-contaminated with the rest of the laydown.
QMEL / QN Type F2
SAP QM
Auto-generated Quality Notification against the bale batch with contaminant class, photo attachment, and shift-operator ID. This is the record the merchant sees in the vendor scorecard.
Batch Status: RESTRICTED
SAP MM
Bale batch status flipped to restricted-use so the mixing lot cannot be released to carding via the usage decision (UD). Human override recorded with reason code.
Movement 344
SAP MM
Reversal or transfer of the mixing lot to blocked stock if contamination is confirmed on the laydown, with the original movement 101 goods-receipt visible in the audit trail.
Vendor Scorecard
SAP QM-IS
Contamination-per-bale rate rolled up to the merchant vendor code and month, so the annual purchase negotiation opens with a number, not an anecdote.
Run a bale-batch audit on your last month of shift sheets. If the count of contaminants on paper doesn't match the count of foreign-fiber cuts on your Quantum reports, the merchant is winning. Book a plant assessment — we'll show what a camera-to-QM link would have caught.
12-Week Pilot Shape on One Station
One opener, one merchant, one mixing shift. The pilot is not sized to prove the camera works — the vendors above have proven that for a decade. It is sized to prove the SoR write-back cuts a claim your mill actually files.
Weeks 1–2
Baseline & Scope
Camera installed above one bale opener. Two weeks of read-only capture with human-labeled classes to calibrate the classifier against your actual bale supply and mixing rules.
Weeks 3–4
Classifier Tuning
Class definitions locked with your QA head: what counts as PP-clear, what counts as jute, what counts as bird-nest. Confusion matrix reviewed. False-eject rate targeted below your Jossi baseline.
Weeks 5–8
Write-Back Live
Auto QN generation switched on against the bale batch in SAP QM. Batch status hold rules configured. Vendor scorecard rows begin populating for the pilot merchant only.
Weeks 9–12
Claim-Cycle Match
First 90-day claim window closes. Contamination photos from Week 5 mapped against any yarn-clearer or fabric claim that arrived. Merchant back-charge conversation held with data.
Who Owns the KPI
Vision detects. SAP records. Someone has to own the number in the operations review — and that ownership decides whether the system actually changes what the mill buys next season.
Quality Head
Contamination QN rate per bale, per merchant
Owns the classifier definitions and the threshold at which a mixing lot is held. Reviews confusion matrix monthly.
Blowroom Sup.
Photos-captured vs contaminants-ejected ratio
Owns the operator workflow at the opener — the human who confirms the class and marks the bale ID when the auto-classifier is uncertain.
Materials / GM
Contamination back-charge $ per merchant per year
Owns the vendor scorecard number and the annual negotiation. Uses the QN history to change price or terms with the merchant.
Dyehouse Liaison
Fabric claims traced to a QN on the bale batch
Owns the loop-close: every downstream fabric claim gets traced back to whether a QN existed at the opener, or whether the contaminant escaped the pilot camera.
FAQ
Do I still need my Uster Jossi Vision Shield or T-SCAN if I have this?
Yes. Those systems eject contaminants from the fiber stream after the fine cleaner and remain the reason white PP does not reach your yarn. This camera sits earlier, at the opener, and its job is not extraction — it is to create a bale-batch record in SAP QM before the mixing lot is opened and the merchant becomes untraceable. Total Contamination Control, as Uster frames it, is two-stage: preparation and winding. This adds a bale-inspection layer to the preparation stage that talks to the ERP, which the extraction systems do not.
How is this different from my yarn clearer's foreign-fiber count?
A yarn clearer measures what escaped everything upstream and reached the winder. It is the right measurement for shipping decisions and yarn-quality claims, but by the time a Quantum 3 registers an elevated foreign-fiber cut rate, the contamination has been mixed across a thirty-bale laydown and the specific bale — and specific merchant — is not recoverable from the clearer alone. Bale-batch inspection lets you attribute contamination back to the bale ID, which is the record you need for a vendor scorecard, not a shipping decision.
Will this trigger false ejects and hurt my throughput?
The first two weeks are read-only for exactly this reason. The classifier is tuned against your bale supply — different origins carry different contamination profiles, and Indian, US, and West African cotton produce different confusion matrices. False-eject rate is targeted below the eject rate you already accept on your existing Jossi or T-SCAN units. If the number does not clear that bar in the pilot, no auto-hold rule ships to production.
Stop discovering contamination four days late.
See a Bale-Batch QN Land in Your SAP QM on a Real Contaminant
Bring one merchant's last three shipments and one live opener. We'll show the classifier, the QN it generates against the bale batch, and the vendor-scorecard row it produces — before we discuss any camera install.
