Applications and cases
Five material streams, one engineering method.
Each industry asks something different from the imaging path. These overviews show the material, the separation problem, the recommended camera / compute / valve configuration and the metric the line is judged on, from Purity Ratio to foreign matter removal.

Industry overview
Start from your material
Configuration follows the material, not the other way round. The combination of camera path, compute tier and valve board changes with each stream.
01
Grain
Focus: Purity Ratio and good:bad ratio
Typical material
Milled rice, wheat and seeds
Chalky kernels, discoloured grain and light impurities arrive in the same stream at full machine width, so the valve has to reject fast without carrying good grain.
Recommended configuration
02
Tea
Focus: Foreign matter removal rate
Typical material
Green and black tea leaf, stems, fibre
Stems, fibre and foreign matter sit close to the leaf in visible colour, while leaf shape changes with every harvest.
03
Chinese Herbal Medicine
Focus: Grade consistency across batches
Typical material
Sliced and whole herbal material
Slices differ in shape and colour, and the acceptable appearance shifts with origin and season, so a fixed threshold does not hold across batches.
Recommended configuration
04
Ore & Mineral
Focus: Ejection accuracy at high throughput
Typical material
Crushed ore and mineral fragments
Density and internal structure carry the value, surface colour does not, and the ejection stage has to keep up with a wide, heavy stream.
05
Metal
Focus: Alloy purity
Typical material
Mixed alloy scrap and fragments
Alloy families look alike under visible light, and mixed fragments arrive in unpredictable shapes and sizes.
Recommended configuration
Reference cases
Configurations already on machines
Client names stay with the customer. Each case describes the machine builder, the material and the configuration that went into the field.
How delivery runs
From a sample to a supplied module
Four stages, each with a decision attached, so a project does not drift between the sample test and the production order.
Scope
Material, target Purity Ratio, good:bad ratio and cabinet constraints reviewed with your engineering team.
Sample imaging
Samples run against visible, NIR, hyperspectral or X-ray paths to confirm which channel can separate the defect.
Prototype
Camera, edge box and valve board configured together, then ejection timing validated on your own machine.
Supply
Batch supply with recipes stored on the panel PC and models upgraded remotely.
Bring us the material you cannot separate.
A sample description is usually enough to tell whether the problem is imaging, ejection or both. We answer with the path we would take.