Sensor & Payload

Hyperspectral Imaging

Use it when objects look similar in RGB but differ in spectral response.

Technology overview

How it works

Captures a spectrum at every image pixel, turning subtle wavelength-dependent differences into spatial classification or quality maps.

Where it adds value

Seed and crop grading; coating / surface-condition assessment; material classification; inline sorting where surface or near-surface chemistry matters.

Combine it with

Pair with RGB for morphology and localisation; THz when internal structure or moisture must also be assessed; AI for classification and grading.

Practical note

Best deployed as benchtop, inline or UAV-based sensing, depending on target size, speed and required coverage.

Systems that carry it

Seed inspection & sorting systems

THz + hyperspectral, three throughput tiers

Laboratory, online and high-throughput configurations. Hyperspectral reads the coat; terahertz reads the kernel; both fuse at the edge and drive pneumatic or mechanical sorters.

Used in:Seed NDT

Laboratory	Benchtop THz-TDS + hyperspectral
Online	Line-scan THz + push-broom, ~5–10 seeds / s
High-throughput	Area-array THz + snapshot, ≥ 20 seeds / s

Where we use it

Seed NDT

Colour sorting sees the coat. We see the kernel.

Terahertz + hyperspectral high-throughput sorting

Agriculture · ≥ 20 SEEDS / S

Engineering principle: start with the measurement question, then select the minimum sensing stack required.

Partner with us →