Rainfed or irrigated olive groves: water stress and alternate bearing under control.
−0 %water
Indicators recorded on partner farms — full detail in the case studies
Olive trees tolerate drought but pay for it the following year. Olive Care tracks the vigour of every tree through satellite and drone imagery, flags water-deficit zones before fruit set, and secures the deficit irrigation schedule.
Tree-by-tree vigour index and detection of missing trees
Water stress monitoring ahead of flowering and fruit set
Early detection of peacock spot disease
Harvest forecasting by grove zone
What it changes: Reduced alternate bearing · Deficit irrigation under control · Targeted rather than blanket treatments
Cereal Care
Durum wheat and barley: nitrogen management and yield forecasting.
−0 %nitrogen
Indicators recorded on partner farms — full detail in the case studies
In cereals, everything is decided between tillering and heading. Cereal Care turns the NDRE index into a variable-rate nitrogen map, monitors emergence, and delivers a zone-level yield estimate weeks before harvest.
Variable-rate nitrogen map derived from NDRE
Emergence and plant density checks after sowing
Detection of lodging and water-stress zones
Yield estimation by homogeneous zone
What it changes: Nitrogen applied where it pays · Re-sowing decided in time · Harvest logistics planned ahead
Citrus Care
Citrus: salinity, drip irrigation and fruit size.
−0 %water
Indicators recorded on partner farms — full detail in the case studies
Citrus reacts quickly to excess salt and uneven irrigation. Citrus Care combines soil sensors, conductivity readings and imagery to fine-tune watering cycles, sector by sector, all the way to the target fruit size at harvest.
Soil salinity and conductivity monitoring
Drip irrigation uniformity checks
Stress alerts ahead of physiological fruit drop
Vigour tracking by irrigation sector
What it changes: More consistent fruit size · Fewer fruits lost · Irrigation network under control
Greenhouse Care
Greenhouses: climate, fertigation and continuous pest pressure monitoring.
−0 %treatments
Indicators recorded on partner farms — full detail in the case studies
Under cover, a temperature or humidity drift costs a week of production. Greenhouse Care aggregates the sensors of every bay, monitors VPD and raises the alert before the crop falls behind.
Temperature, humidity and VPD per bay
Fertigation and conductivity monitoring
Pest and disease pressure alerts
Climate history compared across cycles
What it changes: More consistent cycles · Fewer curative treatments · Energy consumption under control
What each solution measures first
Quatre cultures, quatre priorités de mesure. Le chiffre en tête de colonne indique l'évolution mesurée par la solution correspondante.
Olive Care
−28% water
Tree-by-tree vigour index and detection of missing trees
Water stress monitoring ahead of flowering and fruit set
Early detection of peacock spot disease
Harvest forecasting by grove zone
Cereal Care
−19% nitrogen
Variable-rate nitrogen map derived from NDRE
Emergence and plant density checks after sowing
Detection of lodging and water-stress zones
Yield estimation by homogeneous zone
Citrus Care
−21% water
Soil salinity and conductivity monitoring
Drip irrigation uniformity checks
Stress alerts ahead of physiological fruit drop
Vigour tracking by irrigation sector
Greenhouse Care
−24% treatments
Temperature, humidity and VPD per bay
Fertigation and conductivity monitoring
Pest and disease pressure alerts
Climate history compared across cycles
Indicators recorded on partner farms — full detail in the case studies
The fundamentals
What happens between the satellite and your field
Before choosing a solution, it helps to understand what it rests on. Here are the six agronomic and technical building blocks behind every RoboCare recommendation.
01
Satellite imagery and remote sensing
The Copernicus programme’s Sentinel-2 satellites pass over Tunisia every five days and measure the light reflected by the canopy across thirteen spectral bands, from blue to shortwave infrared. A healthy plant absorbs red for photosynthesis and reflects near infrared heavily; a stressed plant does the opposite, progressively.
Five-day revisit, archives available since 2015Thirteen spectral bands, three of them in the red edgeAutomated atmospheric correction and cloud masking
That imbalance becomes measurable long before it turns into visible yellowing. This is the whole value of remote sensing: it gives access to a physiological signal, not merely a photograph of the field.
02
Multispectral drones
When the question drops to the scale of a tree or a row, ten-metre resolution is no longer enough. The drone carries a multispectral and a thermal camera and flies a few dozen metres up: each pixel then covers two to five centimetres of ground.
Centimetre resolution, tree-by-tree healthThermal imaging to check irrigation uniformityPlanned missions, automatic orthomosaic and report
At that resolution you count trees one by one, spot a blocked emitter by its thermal signature and map a disease outbreak to the metre. Flights are planned in advance and the report is generated automatically after processing.
03
Ground IoT sensors
Satellites observe the canopy; they cannot see what happens forty centimetres down. IoT stations fill that blind spot: staged moisture probes, soil temperature, electrical conductivity and, depending on the site, a rain gauge and anemometer, all sampled every fifteen minutes.
Moisture measured at several root depthsLoRaWAN or 4G depending on site coverageSolar powered, annual maintenance
Transmitted over LoRaWAN — several kilometres of range at a power draw that allows years of autonomy on battery and solar panel — these readings calibrate the satellite model’s thresholds against the farm’s actual soil conditions.
04
Applied artificial intelligence
A raw vegetation index is noisy: sun angle, surface moisture, phenological stage and soil type all shift it without any agronomic problem being involved. The models’ job is precisely to separate that noise from the useful signal.
Image segmentation to isolate stress zonesBreak detection on the field’s time seriesField feedback loop and seasonal retraining
We combine segmentation networks to outline abnormal zones, time-series models to detect a break in trend, and classifiers trained on field photographs to propose a disease hypothesis. Every prediction is checked against the agronomist’s observations, and models are retrained at the end of each season.
05
Precision agriculture
Precision agriculture starts from a simple observation: a field is never uniform. Soil depth changes, available water capacity varies by tens of millimetres from one end to the other, and a single rate applied everywhere is necessarily excessive somewhere and insufficient elsewhere.
Within-field zoning derived from spectral responseVariable rates instead of uniform onesSeason-to-season comparison on the same grid
Managing with precision means splitting the field into coherent zones and treating each according to its real need. Remote sensing makes that possible without probing the soil every twenty metres: the canopy’s spectral response already integrates the combined effect of soil, water and nutrition.
06
Smart irrigation
Managing irrigation is not about watering more, but about knowing which sector has actually depleted its reserve. The water balance combines reference evapotranspiration, the crop coefficient for the current stage, effective rainfall and probe moisture readings.
Water balance per sector, revised at every passDetection of drip-network unevennessManaged deficit irrigation on perennial crops
The output is a volume per sector and per watering round, revised at every satellite pass. On olive, it supports controlled deficit irrigation: deliberately reducing supply during less sensitive phases without compromising fruit set.
NDVI layer applied to the field, with stress hotspots flagged in red
Index curve across the season and anomalies detected automatically
Change detection in six classes, from severe decline to strong growth
Comparison of two satellite passes and scan history with cloud cover
Dashboard
Health status of every field, open alerts and today's tasks on a single screen.
Field mapping
Boundaries, areas, crops and per-field history, with KML/Shapefile import.
NDVI / NDRE satellite monitoring
Vegetation index time series and side-by-side comparison between two dates.
Weather intelligence
7-day forecasts, rainfall totals and evapotranspiration at field scale.
Smart reports
Weekly PDF report per field, exportable for the cooperative or the bank.
Multichannel alerts
Email, SMS and WhatsApp — with escalation if the alert goes unread.
Going further
What you actually do inside the platform
Beyond the six modules, four uses shape the daily routine of monitored farms.
Field mapping and management
Each field is a geographic object: outline, computed area, crop, variety, sowing or planting date, irrigation sector. Outlines are drawn by hand over satellite imagery or imported from a KML, GeoJSON or Shapefile the cooperative already uses.
KML, GeoJSON and Shapefile import with automatic reprojection
Split into irrigation sectors and management zones
Stackable layers: indices, soil, elevation, history
Analysis, statistics and history
Every measurement is kept. You follow an index curve across the season, overlay it on last year’s, and compare two satellite passes side by side to see what changed between two dates.
Time series by field, by zone and by crop
Comparison between two dates and two seasons
Statistics aggregated at farm level
Smart alerts and escalation
An alert only fires when a deviation exceeds the threshold calibrated for that crop and growth stage — not on every index fluctuation. It carries a zone, a probable cause, a severity level and a recommended action.
Thresholds per crop and per phenological stage
Email, SMS and WhatsApp, with escalation if unread
Alert history and resolution tracking
Farm management
Several users share one account with distinct rights: the head grower views and comments, the technical manager validates interventions, the cooperative gets a consolidated view of its members.
Multi-user accounts and differentiated roles
Consolidated view for cooperatives and grower groups
Log of recorded interventions and followed recommendations
From raw data to the agronomic decision
No hardware to buy to get started: the satellite is enough.
01
Create your account
Sign up online, with no hardware and no installation.
02
Add your fields
Draw the boundary on the map or import a KML/Shapefile.
03
Monitor your crops
Every satellite pass updates the vegetation status of the field.
04
Receive your decisions
Alerts and recommendations by email, SMS or WhatsApp, in French or Arabic.
Under the hood
How the platform is built
Five layers, from downloading a raw image to the notification landing on a phone.
01
Ingestion
Scheduled retrieval of Sentinel-2 and Landsat scenes, collection of sensor frames through the LoRaWAN gateway and the 4G network, import of drone flights and local weather data.
02
Processing
Atmospheric correction, masking of clouds and their shadows, clipping per field, index computation and inference model execution across parallelised jobs.
03
Storage
Rasters kept in a cloud-optimised format so only the requested window is read, geometries and time series in a spatial database, cold archives for multi-year history.
04
API layer
A documented REST API exposes fields, indices, alerts and reports. It serves the web interface, the mobile app and third-party integrations through the same authentication.
05
Delivery
Vector maps rendered in the browser, PDF reports generated on demand, notifications routed to email, SMS or WhatsApp according to each user’s preference.
Trust
Cloud, security and synchronisation
Your field boundaries and yield history are sensitive data. Here is how they are handled.
Cloud hosting
Infrastructure hosted in ISO 27001 certified European data centres, with daily encrypted backups and periodically tested restoration.
Encryption
Traffic encrypted end to end with TLS 1.3, data encrypted at rest, token-based authentication with expiry and two-factor authentication available on administrator accounts.
Data ownership
Your data stays yours. It is never sold or shared with third parties, and both a full export and permanent deletion are available on request.
Availability and synchronisation
99 % availability measured over twelve months. Data already loaded stays readable offline, and synchronisation resumes automatically once the network is back.
Frequently asked questions
Arboriculture and large-scale crops
What RoboCare changes in practice, orchard by orchard and field by field.
Can RoboCare monitor olive orchards?
Yes — olive monitoring is one of RoboCare’s core solutions, Olive Care. It tracks tree-by-tree vigor using satellite and drone imagery, flags missing or declining trees, monitors water stress ahead of flowering and fruit set, and screens for early signs of peacock spot disease. Farms using it have reduced irrigation water use by around 28% while managing the tree’s natural alternate-bearing cycle more predictably.
How can precision agriculture improve olive production?
Olive trees tolerate drought but pay for it the following year through reduced yield — the alternate-bearing effect. By tracking vigor and water stress tree-by-tree before flowering, RoboCare lets growers time deficit irrigation precisely, keeping trees under just enough stress to manage vegetative growth without triggering a yield crash the next season. Early peacock spot detection also means treatment goes only where disease pressure is confirmed.
Can RoboCare detect stress in fruit trees?
Yes. Beyond olives, RoboCare’s Citrus Care solution applies the same tree-level monitoring to citrus orchards — combining satellite indices with drone imagery when finer detail is needed to spot water or nutrient stress in individual trees before it’s visible on the ground. The same approach extends to other fruit tree crops as farms bring new orchards onto the platform.
Is RoboCare suitable for large farms?
Yes — RoboCare already monitors over 100,000 hectares, and pricing scales with a degressive tier beyond 100 hectares plus a specific rate for cooperatives grouping several members. Cereal Care, for example, turns NDRE readings into nitrogen modulation maps across large blocks, tracks emergence and lodging, and estimates yield by zone weeks before harvest — built for scale, not just single small plots.
How can large-scale farms optimize irrigation and fertilization with RoboCare?
Across a large farm, RoboCare’s platform zones every field automatically from satellite indices, so irrigation and nitrogen decisions are made per homogeneous zone instead of per whole field. NDWI flags where water stress starts first, NDRE drives variable-rate nitrogen maps, and the same maps export directly to the formats most spreading consoles already accept.
Can RoboCare monitor multiple agricultural plots?
Yes — the platform is built around managing a full portfolio of fields from one dashboard, not a single plot. Each field keeps its own history, indices, and alerts, while the dashboard surfaces which fields need attention across the whole operation at a glance. This is exactly the use case cooperatives rely on, since it lets one account track many members’ plots under a single, consistent view.
Let's talk about your fields
Which crop do you want to monitor first?
Describe your fields and a RoboCare agronomist will reply within 48 hours with the solution calibrated for your crop.
