uAgri Studio
Back home
Agricultural Intelligence

Agricultural intelligence connected to the Digital Twin

Terrain, soil, climate, satellite, water and planning are organised around each parcel. FIWARE preserves operational reality; uAgri adds scenarios, provenance and plan-versus-actual comparison.

Map-first workspace

The parcel is the entry point to layers, analysis and planning.

Geometry, inspector, legend, provenance and availability states share one surface.

Decoupled providers

CNIG/Copernicus, SoilGrids, ERA5/Open-Meteo and CDSE use common contracts.

Regional routing, fallback, cache, budgets and freshness stay outside the agronomic core.

Explainable engines

Suitability, water, irrigation and economics use visible rules and parameters.

Insufficient results fail closed and are never filled with generative values.

Versioned plan

UAgriAgronomicPlan separates planning, parcel reality and economic campaign.

Each approved version is immutable and relates to parcel, campaign, operations and evidence.

Earth Observation

Sentinel-2 catalogue, COG/tiles and temporal indices document vigour, water and change.

NDVI, NDWI, NDMI, EVI and SAVI are shown as derivatives with date, resolution and algorithm.

Plan versus actual

The approved plan is compared with irrigation, costs, harvest, operations and observed imagery.

FIWARE provides facts; Intelligence produces traceable comparisons and reports.

Agricultural Intelligence

From territorial data to explainable agricultural plans

A modular provider and deterministic-engine layer analyses every parcel, preserves provenance and prepares decisions later checked against FIWARE.

01Understand

Understand the parcel

The workspace combines geometry and external sources without confusing estimates with measurements.

  • Elevation, slope and aspect
  • Texture, pH, carbon and water capacity
  • Climate, weather and evapotranspiration
  • Sentinel-2, NDVI, NDWI, NDMI, EVI and SAVI
02Plan

Plan with scenarios

Engines produce explainable ranges and never replace certified engineering or advice.

  • Suitability with rules and blockers
  • Water demand and irrigation pre-design
  • CAPEX/OPEX scenarios
  • Versioned and approvable UAgriAgronomicPlan
03Learn

Compare plan and reality

The Digital Twin provides actual observations to learn from each campaign.

  • Consumption and operations versus forecast
  • Harvest versus expected yield
  • Vegetation development versus EO baseline
  • Provenance report and ERC-7943 v2 readiness

Progressive architecture

Open, modular and FIWARE-native

  1. 01Regional and global providers
  2. 02GIS/physical/rules analysis
  3. 03Versioned approved plan
  4. 04FIWARE actual and evidence

Readiness

Open parcel intelligence

Enter the demo, select parcela-norte-jerte and explore data, plan and comparison with real observations.

Estimates, laboratory results and sensors are displayed as distinct data classes.

Heavy rasters and artifacts remain outside Orion; FIWARE receives summaries and relationships.

ML yield forecasting remains in research until validated cohorts are available.