Map My Crop

Map My Crop: Satellite Agronomy At Field Scale

• Ten-plus satellite constellations, with NDVI and 20-plus vegetation indices updated daily.
• Automated field boundary detection without surveys, plus crop type identification across 150-plus crops.
• Three tiers — digitise the farm, optimise operations, then command-centre analytics for enterprises.
• Built for eight buyer types, including banks, insurers, governments, cooperatives and plantation owners.
• Hardcastle Agrotech Solutions, with ISO 27001, GDPR compliance, NVIDIA Inception and ISRO authorisation displayed.

What Can Satellites Actually See In A Field?

The platform is built on 24/7 satellite health surveillance with NDVI and more than twenty vegetation indices updated daily, alongside crop type identification using multispectral analysis.

Understanding what a satellite genuinely measures explains both why this works and where it stops.

A multispectral sensor reads reflected light in bands the eye cannot see, and healthy vegetation has a distinctive signature — it absorbs red light for photosynthesis and reflects near-infrared strongly. Stressed plants reflect differently, and they do so before any visible change appears.

That gap is the whole value. A moisture deficit, a nitrogen shortage or an early disease front alters the spectral signature days or weeks before a farmer walking the field would notice, and at that point intervention still works.

What a satellite cannot do is tell you why. An index showing depressed vigour in one corner of a field is a signal that something is wrong there, not a diagnosis — and distinguishing water stress from disease from a compacted headland still requires someone to go and look.

A satellite sees stress days before a farmer walking the field would. It cannot tell you whether the cause is water, disease or compaction.

Boundary Detection Without A Survey

The first tier's headline capability is precise field maps without surveys, with automated delineation in minutes and an accuracy figure quoted across 80-plus countries.

This sounds administrative and is genuinely foundational, because almost everything downstream depends on knowing where a field is.

Field boundaries in much of the world are not digitised. They exist in paper records, in local knowledge, or in nothing at all — and without them there is no unit to attach a crop type, a yield estimate, an insurance policy or a loan to.

Manual delineation is the conventional alternative and it is slow and expensive at scale. Automated detection from imagery turns a survey project into a processing job, which is what makes national-scale monitoring feasible at all.

The consequences reach well beyond agronomy. Insurance claim verification, subsidy administration and lending against land all rest on an accurate boundary, and that is why the same platform ends up selling to banks and governments rather than only to farmers.

Indices, Alerts And Acting On Them

Beyond monitoring, the platform offers disease identification in under five seconds, yield forecasts up to three months ahead, irrigation modelling with zone-specific water stress maps, and risk scoring for lenders.

Those four have very different evidential standing, and it is worth separating them.

Index monitoring and alerting is the solid ground. Detecting that a zone has diverged from the field average is a measurement, and acting on it is straightforward — go and look, then irrigate, treat or leave it.

Yield forecasting is harder and more valuable. It depends on modelling how the rest of the season will go, and the further ahead the forecast, the more it rests on weather that has not happened. A three-month-ahead forecast is a projection with a confidence interval, and the interval is the part that matters.

Risk scoring for lenders is the most consequential and the least visible. A score deciding whether a smallholder gets credit is a decision about someone's livelihood made partly by a model, and the right questions are what it uses, how it was validated, and what recourse a farmer has when it is wrong — the same governance questions any data platform selling into lending should expect, as with Intelligence Node in retail pricing.

• Treat an index alert as a reason to visit, not a diagnosis.
• Ask for the confidence interval on any yield forecast.
• Check the revisit rate and cloud handling for your region.
• Validate boundary detection against fields you know precisely.
• For lending scores, ask what recourse a farmer has if one is wrong.

Coverage, Cloud And Revisit Rates

The platform states ten-plus satellite constellations and daily index updates across 900 million-plus hectares monthly in 80-plus countries.

Using many constellations rather than one is the correct architectural choice, and the reason is weather.

Optical satellites cannot see through cloud. A single constellation with a five-day revisit gives you an image every five days in theory and, during a wet growing season in a cloudy region, potentially nothing useful for a month — precisely when crop stress is developing fastest.

Combining constellations multiplies the chances of a clear view, because they pass at different times and orbits. It is the difference between a monitoring service and an occasional picture.

Radar-based sensing is the other answer to cloud, since radar penetrates it, and the site does not say whether any of its constellations are radar. That is the specific question to ask for any wet-season or tropical deployment, along with what the effective observation frequency is in your own region rather than the stated update cadence.

Reading A Page Dense With Numbers

The homepage carries an unusual density of quantified claims: yield increases, input cost reductions, forecast accuracy, water savings, default prediction improvement, customer retention and payback period, alongside two different figures for AI accuracy in different sections.

None carries a source, a study, a sample or a definition, and that is the thing to hold in mind while reading the rest.

A stated average yield increase, for instance, cannot be evaluated without knowing against what baseline, on which crops, in which regions and over how many seasons. Agriculture varies so enormously by geography and season that an average across 80 countries is close to uninterpretable.

One displayed testimonial refers to a yield improvement figure so large that it is very unlikely to describe yield in the agronomic sense, which is the kind of thing a careful editor would query before publishing.

None of that impugns the technology, which is real. It means the numbers are marketing rather than evidence, and the credible signals on the page are elsewhere: ISO 27001, a named ISRO authorisation, NVIDIA's programme, and the specificity of the product tiers. Exported field data still has to be analysed somewhere, which is Quadratic's job, and the source-linking discipline Daloopa applies to filings would suit these numbers well.

Monitoring A Farm From Orbit

Satellite agronomy fails when alerts arrive and nobody acts, which happens as soon as the alerts outnumber the people who can respond.

This sequence establishes trust in the data before scaling the coverage.

1. Validate boundary detection on fields whose acreage you know exactly.
2. Check the effective clear-image frequency in your region, not the stated cadence.
3. Ground-truth the first month of alerts by visiting the flagged zones.
4. Learn which index divergences matter on your crops and which do not.
5. Compare one yield forecast against your actual harvest before relying on it.
6. Decide who responds to an alert before switching alerts on.

Map My Crop: Data And Coverage

Is Satellite Monitoring Useful On Your Farm?

Watch from orbit when

• Your holding is too large to inspect thoroughly every week.
• Field boundaries are not yet digitised anywhere reliable.
• You need evidence for an insurance claim or a subsidy return.
• Irrigation is a major cost and zones vary across a field.
• You are lending, insuring or aggregating across many farms.

Walk the field when

• You farm a few fields you walk daily anyway.
• Persistent cloud makes optical imagery unreliable in your region.
• Nobody would act on an alert once it arrived.
• You need a diagnosis rather than a signal that something is wrong.
• Your crop or region falls outside what the platform covers.