Non-destructive quality inspection for produce — ripeness, internal defects, size and weight judged in seconds on the packing line, so every crate that leaves the shed is graded the same way by the same standard.
Premium produce is still graded by hand — tapping a durian, smelling the stem, judging a colour by eye. A good grader is genuinely skilled, but they get tired, they disagree with the next shift, and they cannot be hired fast enough in season.
The cost lands in two places: a bad fruit that reaches an export customer, and a good fruit that gets downgraded for nothing. Both are margin, and neither shows up in a report.
Our flagship for this line. Built first for durian — the hardest case, because the part that matters is sealed inside a spiky shell — and retrained from there for other high-value crops.
Judged from the combined signature of the fruit rather than a single cue, so the call matches what the buyer will find when they open it.
Hollow heart, wet core, uneven flesh and pest damage flagged without cutting — the defects that cause returns.
Colour, bruising, cracks, mould and blemish coverage scored consistently under controlled lighting.
Weighed and dimensioned in the same pass, so grade and class come out together on one label.
Trained on your own varieties — Musang King and Black Thorn do not grade the same, and the model should know that.
Every unit logged with its images, its scores and its grade, exportable for the buyer or the certification body.
Start with one station to prove the grading against your own crop, then scale to the throughput the season demands.
Single-unit, operator-fed
Conveyor station with diverter
Multi-lane, full traceability
Indicative figures for the reference build. Throughput, sensing set and grade rules are all configurable to the crop and the packhouse.
| Parameter | Specification |
|---|---|
| Inspection method | Non-destructive: multi-angle vision under controlled lighting, combined with weight and dimensional measurement; additional sensing modalities selected per crop |
| Crops | Built first for durian; retrainable for mango, jackfruit, avocado, melon and other high-value produce |
| Grading output | Configurable grade classes with a confidence score and a reject reason per unit |
| Throughput | Sized to the packhouse target; multi-lane configurations for peak season volume |
| AI processing | On-device inference on an edge controller — the line keeps grading if the internet drops |
| Training | Models trained on your own crop and your own grade rules, retrainable on-site as the season changes |
| Traceability | Per-unit record with images and scores; CSV / API export to packhouse ERP |
| Environment | Washdown-tolerant enclosure for packhouse conditions; humidity and temperature rated for the tropics |
| Interface | Local touchscreen plus a browser dashboard for batch and shift reporting |
Specifications are indicative and subject to change. Contact us for the current datasheet of a specific configuration.
Export grading & sorting
Pre-shipment verification
Grading at intake
Lot-level quality checks
Raw material screening
Condition on arrival
Auditable grade records
Varietal & agronomy trials
Tell us the produce, the grades you sell into and your peak-season volume. We will run a feasibility trial on your own fruit before any hardware is committed.