SnapDurian

Snap. Know. Act. · Offline durian disease & pest scouting

Spot durian disease in the orchard, even with no signal.

SnapDurian is a handheld device. Photograph a leaf or trunk during your orchard walk, and it suggests which disease or pest it is likely to be in about a third of a second. Everything runs on the device, so it works deep in the orchard where phone apps stop working.

Join the grower trial Read the research
EXAMPLE · leaf lesion

Illustration, not a product photo

~350 msper photo, on the device
12common diseases and pests
0 barsof signal needed
11 sitesof field photos, Peninsular Malaysia and Sabah

Why it matters

A missed symptom costs more than a scouting walk.

RM 10,000+

Every tree counts

A mature durian tree can be worth RM 10,000 to 20,000. Catching a problem early protects years of growth.

RM 2,700

Spraying is expensive

One spray of an 8 ha orchard can cost around RM 2,700. Knowing what you are treating helps avoid wasted sprays.

No data

Orchards are remote

Many orchards have weak or no mobile signal. Disease apps that need the internet stop working right where you need them.

Figures are from a durian grower we interviewed in Sabah. Costs vary by orchard.

How it works

Three steps during your normal orchard walk.

Point and shoot

Photograph a leaf or trunk that looks wrong. One button, no app, no login.

Check on the spot

The device runs the model itself and shows its best suggestion in about a third of a second.

Decide what to do

Use the suggestion to decide what to check next or who to call. It is a decision aid, not a replacement for your agronomist.

Built for the field

Tested in tropical heat, on battery, far from Wi-Fi.

Every design choice came from problems we hit in real orchards.

Power
Runs on batteryBuilt on a Raspberry Pi 5, so it is affordable to make and easy to repair.
Connectivity
Fully offlineNo internet, no account and no app to install. It works the same at the far end of the orchard.
Heat
Keeps running in the heatWe traced field shutdowns to the board's 82 °C safety cut-off and added short rest cycles so it stays on.
Small pests
Does not lose the small onesWe chose a compression method that keeps the smallest pests detectable instead of the fastest option that dropped them.
Speed
About 350 ms per photoMerging two separate models into one halved the waiting time without losing accuracy.
Hardware choice
Simple to maintainWe tested a faster, pricier board but chose the one a grower can use without technical help.

Tested where it will be used

Field photos from real orchards, checked on farms the model had never seen.

  • 1,372 field photos from 11 orchard sites in Peninsular Malaysia and Sabah.
  • Labelled to a standard agreed with an experienced crop-protection practitioner.
  • Every model is tested on farms it was not trained on, because your farm is new to it too.

What we tell growers up front

Accuracy on a brand-new farm is lower than in testing on familiar farms, and some pests are harder than others. That is why every trial farm helps: photos from your orchard make the next version better for you.

Research

Peer-reviewed and published.

Thermal cut-off, not frequency throttling: duty-cycle scheduling and a dataset-integrity audit for on-farm edge disease detection in durian

Lin, D. S. (2026). Smart Agricultural Technology, 15, 102526. Elsevier. Open access.

doi.org/10.1016/j.atech.2026.102526

Grower trial · late 2026

Try SnapDurian in your orchard.

We are looking for durian growers in Malaysia to use the device during their normal scouting walks and tell us what works and what doesn't. Trial devices are provided free of charge.

lindingshan@snapdurian.com

Tell us where your orchard is, roughly how many trees you have, and the best way to reach you.