SKU color sorting: acquire repeatable color

Purpose and operating boundary

Build a source-color record that survives motion and material variation before assigning a color class or semantic SKU.

This tutorial defines an engineering evaluation, not a ready-to-run product certification. The current Deyi Vision public PC example covers an EAN-13 workflow on confirmed sample images. Native OCR and detection work in the engineering code must be checked against the installed runtime, bound models and job before use. A browser template does not prove browser execution: the planned OpenCV.js backend needs separate verification. Cloud recognition requires explicit approval of the image scope, provider, model and cost. A review archive is not an executable release package. Camera, firmware and driver compatibility require their own check. All worksheet values below are hypothetical; they are not Deyi Vision measurements or customer results.

Prepare the evidence

Use actual product finishes and a neutral reference patch under the intended illuminator. Include glossy, faded and previously unseen products; declare camera color format explicitly. Before reusing the capture configuration, check that the stored pixel format and active region still match the saved recipe. Record any binning, resizing or image conversion because the object scale calculated from source pixels no longer applies unchanged after those operations. Compare the archived frame with the frame passed into the first processing node. If they differ, retain both identities and explain the conversion. During a reconnect trial, distinguish a new exposure from a repeated buffered frame. Disable automatic pass-through while acquisition identity is uncertain, and retain the event that caused the mismatch rather than silently resetting the counter. Color sorting and SKU recognition answer different questions. A color region may match a red reference while the item is the wrong size, label or package variant. Define whether the job sorts a physical color, verifies an expected SKU or requires both independent checks. Use color images with known channel order and keep exposure, white balance and light spectrum in the recipe. Test acceptable material lots, glossy surfaces, shadows and foreign products rather than only a clean reference swatch. Low saturation makes hue unreliable, and clipped highlights remove information that an algorithm cannot restore. Unknown, unobservable and mixed-color items need an explicit route. If semantic SKU recognition uses a model, validate that model separately from the color threshold and require approval before any cloud transfer. Begin with an acquisition record rather than a displayed preview. A preview may be resized, sharpened or reused after the input has stopped. Save the unmodified frame, image dimensions, pixel format, exposure, gain, lens setting, working distance, trigger identity and acquisition time. Include the smallest relevant feature and the largest allowed position variation in the same test. First capture a stationary sample to separate focus from motion blur; then repeat at the actual movement conditions without changing the optical geometry. Link every saved image to the physical sample or event. A missing or stale frame is an acquisition failure even when the last visible image looks good. Choose a Deyi Vision color area-scan evaluation route that captures actual color information; copying monochrome values into RGB channels does not supply it. Record raw pixel format, any Bayer demosaicing, and delivered RGB/BGR channel order before measuring the patch. Fix white balance, gain, exposure and illumination spectrum in the recipe; check unsaturated neutral and product references across allowed lots and motion. Verify shutter/trigger behavior and camera/driver compatibility on the installed setup. Keep color classification and SKU identity as separate acceptance tasks; selecting a color camera proves neither classification accuracy nor product identity.

Procedure

1. Capture color originals with channel order and pixel depth recorded, without silently converting the evidence to grayscale.
2. Fix exposure, gain and white balance for the comparison; record clipping and the actual illumination setting with each frame.
3. Check focus and object scale across allowed heights, then compare stationary and moving color patches at identical settings.
4. Retain the full source frame and object identity so that later background exclusion or SKU detection can be traced to its pixels.

Interpret the trial

At the hypothetical field of 120 mm across 2400 px, the scale is 0.05 mm/px. A 2 mm comparison patch spans 40 px. These dimensions describe sampling only: they do not establish colorimetric accuracy or guarantee that similarly colored SKUs are distinguishable. Color and product identity require separate acceptance.

Recording worksheet

s = 120 / 2400 = 0.05 mm/px
patch = 2 / 0.05 = 40 px

Two engineering views

Failure diagnosis

• Automatic white balance changing between products invalidates a fixed color comparison.
• A grayscale archive cannot reconstruct the lost source chromatic information; capture color again rather than guessing.

Acceptance and handover

Accept acquisition only when declared color references remain comparable without clipping across the permitted conditions. Preserve ambiguous and novel products as unknown examples for later evaluation, not as assigned known SKUs.

Keep the source images, the expected decision, the exact revision and the complete result record together. Freeze the evaluation set before selecting the final parameters. If those answers are later used to tune the workflow, mark that set as development data and obtain a new independent acceptance set. Keep Review, NoResult, execution error and missing acquisition visible in the denominator; do not turn them into passes by filtering rows. Save a reproducible failure alongside the successful sample. The handover should identify who may change the recipe, how a prior approved revision is restored and which device output remains inhibited until acceptance. A software result alone does not establish that a moving physical part reached the correct destination.