Verdantis
Verdantis

Sensors & Services

A live field-sensor reading from the Devinska Nova Ves pilot plot, shown here as a public demo of the capability - more services for this section are planned.

Two datasources, one picture

Verdantis combines two fundamental data sources: (A) satellite intelligence and (B) sensor ground-truth data. The two work in symbiosis, because of how differently each one scales. Satellite intelligence identifies the general pattern within a plot and helps you understand roughly where a problem sits - but it can't pin down an exact position, or an exact value.

This is where sensors come in. Sensors are purpose-built, metrologically calibrated devices capable of measuring the variables that actually drive canopy health on your plot - soil moisture, salinity, temperature, air conditions, and more. True values, in real time, with no delay.

Sensors become even more powerful combined with Sentinel-2 imagery. Sentinel-2 uncovers the general shape of a problem - drought, or one side of a plot losing moisture earlier than the rest. Once that pattern is identified, sensors can be deployed exactly where it's unfolding, pointing to the concrete crops or trees sitting on dry soil - no need to blanket-deploy across hectares of fields, just the specific locations that matter. That saves money, computing time, and work.

Which sensors

Soil moisture

Understand and navigate new drought periods with ease.

Soil components

Is your soil healthy, and can fertilization patterns be improved?

Temperature

Fight frost effectively, and understand exactly when it happens.

Location sensors

For machinery on your plot - see the patterns of routine operations and use those insights to improve harvest.

Try it: a 7-sensor network simulation

The real field-sensor API is temporarily offline, so this is built against a fake API (sensor_fake_api/ in the repo) matching its exact shape - 6 fixed sensors plus 1 moving sensor simulating a tractor crossing the plot, a map, a 10-day soil-moisture chart, and a crop/season-aware "critically low" status readout. Shown here as a periodically-regenerated snapshot (this site has no backend to poll); run sensor_fake_api/app.py yourself locally for a genuinely live, polling version.

Open the simulation →

Where this sensor sits

The live sensor map isn't available right now.

Live dashboard demonstration

A preview of what a live-updating dashboard looks like, built with a lightweight (~50KB), canvas-rendered charting library rather than a heavy framework - chosen so the chart stays fast on this plain, no-build-step site. The data below is real: every point is an actual historical reading from the Devinska pilot's temperature/soil-moisture sensor, not simulated.

Demo · historical readings, not a live feed yet
Temperature (°C) Soil moisture (raw units)

What it would take to make this genuinely live (updating on its own, with no redeploy of this site): one small addition on the sensor service's side - a public, unauthenticated, read-only endpoint that this page can poll directly every minute or so. The service already has the two pieces that make this a config change rather than a rebuild: an "only what's new since I last checked" query parameter, and a planned switch from a 30-day-delayed view to a live one. The homepage's weather cards already work exactly this way (polling a public API straight from the browser) - the same pattern just needs a sensor-side endpoint to point at.