Seed NDT

Non-destructive multimodal seed quality assessment for internal integrity, moisture-related condition and automated grading.

Discuss a Site Trial

Multimodal seed quality assessment: same-seed hyperspectral, machine-vision and terahertz acquisition fused into quantitative indicators for breeding selection, quality control and inline sorting, with a six-step translation pathway

Inspection Priorities

Internal Integrity & Defect Screening

Identify internal structural abnormalities such as cracks, voids, tissue loss, insect damage, empty shells and other concealed defects.

Moisture & Physiological Condition

Assess moisture-related and dielectric differences within individual seeds to support quality indicators associated with storage condition, deterioration and viability-related status.

Identity & Quality Grading

Combine spectral, imaging and morphological information to support seed classification, admixture screening and quality grading.

Standard inline decision workflow

Inline, single-seed flow: positioning, optical / hyperspectral acquisition, terahertz internal inspection, feature fusion, AI decision, grading / rejection and data logging.

Seven-step inline decision workflow: single-seed positioning, optical / hyperspectral acquisition, terahertz internal inspection, feature fusion, AI decision, grading / rejection, data logging

Bench-top terahertz seed imaging cabinet with sample stage and workstation — Laboratory THz seed imaging platform

Inline imaging head above a conveyor or stage, linked to a computer, scanning a single seed — Inline imaging head over conveyor / stage

Sensing & System Configuration

THz + Hyperspectral / Vision · Laboratory / Inline · AI-Assisted Analysis

Complementary sensing methods and deployment platforms are selected to suit the inspection task.

Capability	Configuration & Purpose
Terahertz — THz	THz spectroscopy and imaging provide non-destructive sensitivity to internal moisture distribution, dielectric response, cracks, voids, tissue loss and concealed structural abnormalities.
Hyperspectral Imaging	Captures surface and near-surface spectral information associated with colour, texture, selected chemical or physiological correlates and visible or near-surface abnormalities.
Optical	Provides rapid assessment of seed morphology, size, shape, colour, surface damage and other externally visible features.
AI-Based Multimodal Fusion	Combines THz, hyperspectral, vision and reference-test data to generate classification, regression and anomaly-detection outputs. Converts multiple sensor signals into practical quality indicators rather than present raw spectra alone.
High-Throughput Scalable Platform	THz/hyperspectral arrays, high-speed imaging and edge inference can be developed where full-batch coverage and higher throughput are required, subject to crop- and line-specific validation.

Laboratory to inline deployment

01
Mechanism validation & calibration
Broadband THz-TDS and imaging with same-seed or matched-batch reference testing. Objective: confirm signal origin, repeatability and genuine association with target quality attributes.
02
Feature reduction
Frequency selection, feature engineering and model simplification. Objective: remove redundant information and validate robustness across seed lots and cultivars.
03
Rapid inference
Selected-frequency sensing, line scanning or array-based detection, and real-time algorithms. Objective: enable second-scale or continuous inspection.
04
Inline closed-loop operation
Conveying, positioning, detection, decision, and rejection / grading. Objective: deliver a maintainable, calibratable, production-ready inspection system.

Stages 1–2 run on the laboratory precision platform (model development); stages 3–4 on the inline production platform (industrial deployment). Modular and scalable: standardised interfaces, plug-and-play modules, remote diagnostics, continuous improvement.

THz absorbance spectra of sunflower seeds at three levels and THz images of shrivelled, insect-damaged, empty-shell and internally damaged seeds beside photographs of the same seeds — Sunflower seed — THz absorbance at 5 %, 10 % and 15 % (a, b) and THz images of shrivelled, insect-damaged, empty-shell and internally damaged seeds (c)

THz images of a wheat seed from 0 to 48 hours of germination and a predicted-versus-actual germination-stage scatter plot — Wheat seed — THz images over 0–48 h of germination and predicted vs actual germination stage

Results & Deliverables

Deliverables are defined around the selected seed type, quality objective and processing workflow.

Internal Integrity & Moisture Indicators

Decision-oriented outputs can include an Internal Integrity Index, Moisture-Related Index and Abnormality Score.

Viability-Related & Quality Scores

Support viability-related, germination-stage or broader quality scores.

Pass / Reject & Quality Grading

For inline applications, multimodal data can be translated into operational outputs such as Pass / Reject, quality grade, defect category or targeted reinspection.

Custom Seed Inspection Systems

Application-specific systems can integrate conveyor handling and sorting mechanisms around agreed crop type, seed geometry, coating condition, target defect and throughput requirements.

Systems that support this application

Seed inspection & sorting systemsTHz + hyperspectral, three throughput tiers

Customised for Your Seed & Processing Workflow

Discuss a Site Trial

We welcome the opportunity to discuss your seed types, quality priorities and existing laboratory or processing workflow.

Drawing on our THz seed-imaging experience and multimodal system architecture, we can develop a customised solution around the attributes that matter most to your application — including internal integrity, moisture-related condition, defect screening, quality grading, coated-seed inspection or storage monitoring.

The sensing combination, calibration strategy, seed presentation method, inspection speed, AI model and sorting interface can be adapted to suit your crop, seed size, coating condition, production-line layout and quality-assurance requirements.

Contact us to discuss your inspection priorities and explore a solution designed around your seed-processing or research environment.