How NorrStudio by NorrSpect detects machine oil marks, lubricant deposits, and hydrocarbon contamination on fabric rolls at production speed locating the machine source, preventing downstream processing of contaminated cloth, and eliminating oil-related buyer charge-backs.

Overview

Colour variation is the single most commercially damaging quality issue in dyed and printed textile manufacturing. A shade difference of just 0.5 ΔE imperceptible in isolation but clearly visible when two panels are sewn together can render an entire cut order unsellable if the colour drift has occurred within a roll or between rolls in the same dye lot. Retail buyers operating just-in-time replenishment programmes have zero tolerance for within-lot shade variation: garments that cannot be mixed across size ratios due to shade difference are a direct inventory liability.

NorrStudio, developed by NorrSpect, uses calibrated multi-channel colour imaging and continuous ΔE mapping against approved golden sample baselines to detect colour variation including side-centre-side shade difference, end-to-end shade drift, and between-roll dye lot inconsistency inline at production speed, replacing intermittent spectrophotometer spot-sampling with continuous spatial colour profiling across every roll.

About NorrSpect

NorrSpect is a Swedish AI company headquartered in Umeå, Sweden, specialising in industrial visual inspection for precision manufacturing. Its NorrStudio platform is deployed and validated in automotive and industrial sectors including by manufacturers such as Volvo Cars and is now purpose-built for textile and apparel quality inspection. Colour deviation thresholds and dye-specific calibration models are defined and validated during the pilot phase using the client's approved colour standards and spectrophotometer reference data.

Industry challenge: why colour variation escapes conventional quality control

Standard dye house colour QC relies on spectrophotometer spot-sampling cutting a small swatch from a defined position on the roll and measuring its ΔE against the approved standard. This approach has two fundamental limitations. First, it provides point-in-time, point-in-space coverage: a sample taken from the roll centre at the roll midpoint tells nothing about colour at the selvedge, at the roll ends, or in the metre before and after the sample point. Second, it is performed after dyeing is complete by which point any colour deviation has already propagated through the entire dye lot.

Within-roll colour variation in particular where shade drifts gradually from one end of the roll to the other, or from selvedge to centre is almost entirely invisible to spot-sampling QC and only discovered when garments cut from different roll positions are assembled and the shade mismatch becomes visible under retail lighting.

Solution: NorrStudio AI colour variation inspection

NorrStudio deploys calibrated multi-channel colour imaging across the full fabric width, measuring colour continuously at every point of the roll and expressing deviation as a ΔE value against the approved golden sample baseline stored in the system. Continuous spatial ΔE mapping replaces point-sample spectrophotometry with full-roll colour profiling generating a colour heat map for every roll that shows exactly where, how much, and in which direction colour has deviated from specification.

• Measures colour deviation continuously across the full fabric width at production speed, detecting SCS variation, end-to-end drift, and localised unlevel dyeing in a single pass

• Expresses colour deviation as ΔE against the approved golden sample baseline, with buyer-specific tolerance thresholds encoded as pass / conditional / reject grading criteria

• Detects shade bars and listing events in real time, enabling immediate process correction during continuous dyeing before the fault propagates through the full batch

• Compares between-roll colour profiles within a dye lot to identify rolls that exceed shade pairing tolerance preventing incompatible rolls from entering the same cut order

• Identifies metamerism-prone dye lots via multi-illuminant colour assessment flagging lots that match under assessment light but deviate under D65 daylight or retail illuminant A

• Generates full-roll colour spatial maps shareable with buyers as digital colour QA certificates replacing physical swatch submission

• Integrates colour deviation data with dye house process parameters temperature, pH, dye concentration for closed-loop dye process correction

Use case: knit fabric dye house side-centre-side shade elimination for fast-fashion buyers

The problem: A large knit fabric dye house processing cotton jersey and interlock for fast-fashion buyers was experiencing chronic SCS shade variation problems on its pad-batch dyeing range. Approximately 8–12% of rolls per dye lot were being rejected at buyer incoming inspection for SCS shade difference exceeding the buyer's ΔE 1.0 tolerance a threshold the dye house's spot-sample QC process was structurally incapable of detecting because samples were always taken from the roll centre, never from the selvedge zones where SCS deviation is most severe.

The NorrStudio solution: NorrStudio was installed at the pad-batch dyeing range exit and at the stenter exit. Full-width continuous ΔE mapping was configured against the dye house's approved colour standards for its 120 active shades. The system immediately revealed that SCS deviation on the pad-batch range was consistently greatest in the first 20 metres of each roll corresponding to the roll start before the pad liquor reached thermal equilibrium and in a 15cm zone on each selvedge edge. Pad recipe and mangle pressure adjustments were made within the first two weeks, reducing SCS deviation across the full roll width.

Results: