---
title: "How to check plastic pellets and flake with computer vision"
description: "A camera with even light over a thin layer of pellets or flake on a belt, chute or sample tray takes photos as material passes. A model trained on the plant own images, labeled against its own specs, marks likely problems such as black specks, burnt or off-color pieces, gels, fines and dust, odd shapes and sizes, and foreign material such as metal, paper, labels or other plastics, and it logs a count by lot and time. Flagged lots can be held for a closer look, and the trend helps the team find a filter, a screen or an incoming bale that is causing trouble. Aim the camera at the material, not at people. The quality lead reviews the counts and samples and decides what ships, is reworked or is blended. The model does not replace melt tests, lab checks or customer specs, which stay in place, and people make every quality and release call."
url: https://lidvizion.ai/answers/how-to-check-plastic-pellets-and-flake-with-computer-vision/
site: Lid Vizion
index: https://lidvizion.ai/llms.txt
---

[Direct answers](https://lidvizion.ai/answers/)

# How can a compounder or plastics recycler use computer vision to check pellets and flake for contamination?

A camera with even light over a thin layer of pellets or flake on a belt, chute or sample tray takes photos as material passes. A model trained on the plant own images, labeled against its own specs, marks likely problems such as black specks, burnt or off-color pieces, gels, fines and dust, odd shapes and sizes, and foreign material such as metal, paper, labels or other plastics, and it logs a count by lot and time. Flagged lots can be held for a closer look, and the trend helps the team find a filter, a screen or an incoming bale that is causing trouble. Aim the camera at the material, not at people. The quality lead reviews the counts and samples and decides what ships, is reworked or is blended. The model does not replace melt tests, lab checks or customer specs, which stay in place, and people make every quality and release call.

## What this means in practice

- Start with one line or one sample station.
- Label photos from your own material against your own specs.
- Spread material in a thin even layer under steady light.
- Log counts by lot and time.
- Keep melt tests and lab checks in place.

## How Lid Vizion approaches it

We start with the business decision and operating context, define ownership and measurable milestones, then use configuration, integrations, automation, or custom software only where each earns its place.

## More on Automation & custom AI

- [How can a quarry, mine or aggregate plant use computer vision to keep an eye on rock size between lab tests?](https://lidvizion.ai/answers/how-to-estimate-rock-and-aggregate-size-with-computer-vision/)
- [How can a garage owner or operator use computer vision to keep track of wear and damage between engineer visits?](https://lidvizion.ai/answers/how-to-inspect-parking-garages-with-computer-vision/)
- [How can a property manager or building owner use computer vision to check walls, windows, gutters and other exterior parts?](https://lidvizion.ai/answers/how-to-inspect-building-exteriors-with-computer-vision/)

[Compare services](https://lidvizion.ai/services/) · [Use a planning tool](https://lidvizion.ai/tools/) · [Explore workflow use cases](https://lidvizion.ai/use-cases/) · [More direct answers](https://lidvizion.ai/answers/)

Talk it through

## Have a version of this question inside your company?

Bring the decision, the constraints, and what a good outcome looks like.

[Talk through the decision](https://lidvizion.ai/contact/?need=ai-workflow&from=%2Fanswers%2Fhow-to-check-plastic-pellets-and-flake-with-computer-vision%2F#write) · [Book a call](https://calendar.app.google/bB26LdA9e2gmGFp69)