Multispectral: where it differs
A vegetation index reliably shows that a patch is unlike its neighbour, and how differently the zones have developed. That's the base, but it isn't yet the cause.
Science & practice
We verify, we don't promise
Fine words about the future are worth nothing without evidence. Not “trust our hyperspectral magic”, but a structured base: science, our cases and operational metrics. We don't say “it works perfectly”. We say “verified, with a confidence level and honest limits”.
nitrogen moisture normal confidence 81% Layer 1 · Research base
Why NDVI alone isn't enough
A vegetation index sees greenness and biomass. Narrow biochemical and physiological signals (chlorophyll, the red-edge shift, water absorption, pigments) are read from a denser spectrum and from all-weather radar.
Hyperspectral: which signal changed
Where multispectral sees a few colours, hyperspectral sees hundreds of bands: chlorophyll, the red edge, pigments, water absorption. The red edge is closely tied to leaf chlorophyll and nitrogen status.
Radar: structure and moisture under clouds
Radar is all-weather, day and night. It keeps observation continuous where cloud jams the optical for weeks, and it senses moisture and canopy structure.
Ground and drone: confirmation
The spectrum gives a hypothesis; a soil and tissue sample, scouting and a drone confirm it. The model is calibrated on the farm, it doesn't pass off someone else's percentages as yours.
Layer 2 · Our cases
What hyperspectral saw before NDVI
Every case is built on the same scheme: signal, diagnosis, confirmation, action, result. With a confidence level and explicit limits. Pick a case.
- Problem
- NDVI showed a weak zone; regular monitoring suggested “add nitrogen”.
- Signal
- The red edge is weak, but SWIR water bands and radar point to a moisture limit.
- Hypothesis
- Weakness from lack of water, not from nitrogen hunger.
- Confirmation
- The signal agrees across optical, SWIR moisture bands and all-weather radar; a field check of moisture and compaction.
- Action
- Don't increase nitrogen until moisture is confirmed; check compaction.
- Result
- Nitrogen didn't go into a zone where it wouldn't have worked. A wrong application was prevented.
- Limits
- The exact cause (drying out or compaction) is confirmed on site.
Layer 3 · Operational metrics
You pay not for pixels, but for confirmed decisions
On every pilot we measure operational results, not “pretty maps”. Including an honest false-positive rate.
- 01
Confirmed anomalies
Prove that something real was found, not noise.
- 02
Anomalies missed by the current system
The value of replacing a weak diagnostic layer.
- 03
Hectares where the action changed
The agronomic value of the decision.
- 04
Hectares excluded from application
A direct resource saving.
- 05
Prioritized scouting trips
Value for the process and the agronomist's time.
- 06
Task maps created and applied
Value for the machinery.
- 07
Drone and samples ordered
Service value and confirmation.
- 08
False-positive rate
We show it: this is about trust, not gloss.
Benchmark
Already using something? Put us alongside
Keep your current system. We run AstraField over the same fields, the same dates and with the same agronomists and compare confirmed decisions: nitrogen, moisture, unseeded hectares, scouting priorities, task maps and the false-positive rate.
Let AstraField run alongside for 6-8 weeks. If we don't find confirmed decisions your system missed or confused, then don't roll us out.