Autonomous agricultural drone surveying
Define a field once. The drones keep watch.
LandRithm plans survey missions, flies them, and finds crop stress and irrigation problems in the imagery — then plans the close-up inspection flight. Every step is validated in simulation before it flies.
- MAVLink 2
- PX4
- ArduPilot
- STAC
- COG
- GeoJSON
- PostGIS
- QGIS
How it works
From field boundary to follow-up flight.
Most drone software stops at the map. LandRithm closes the loop: what the analysis finds becomes the next mission.
- 01
Plan
Repeatable coverage missions from a field boundary, a target ground resolution, and an overlap — with coverage and image-quality checks on every flight.
- 02
Fly
The same mission runs against an emulated drone or real PX4 and ArduPilot firmware in simulation, over standard MAVLink — and later a real airframe, through the same interface.
- 03
Analyze
Vegetation-index heatmaps, per-zone health scores, and anomaly polygons — standing water, dry stress, weed clusters, planting gaps — each with a confidence score.
- 04
Inspect
Each flagged anomaly comes with a planned low-altitude inspection mission: the step that closes the loop from detection back to the air.
Use cases
One survey pipeline, many questions about the field.
The same plan-fly-analyze pipeline answers agronomy, irrigation, operations, and terrain questions — with wildfire risk as the newest extension.
Crop health
Agronomists
Vegetation-index heatmaps (ExG, VARI and GLI from RGB; NDVI and NDRE from satellite or multispectral), per-zone health scores, and trends across surveys.
Irrigation anomalies
Growers & irrigation managers
Standing-water and dry-stress polygons with severity scores, compared against zone baselines, with alerts when new anomalies appear.
Pest, disease & stress
Agronomists & scouts
Flagged anomaly clusters — weed patches, planting gaps, suspect stress — with confidence, imagery chips, and a recommended inspection mission.
Field surveying
Farm operations
Repeatable, parameterized coverage missions with coverage and image-quality QA on every flight.
Terrain mapping
Farm ops & drainage planning
Orthomosaics and surface and terrain models from photogrammetry, with slope and drainage derivatives.
Change detection
Agronomists
Survey-over-survey index changes registered to the field grid, with Sentinel-2 time series as the long-horizon baseline.
Wildfire & fuel risk
Land managers
Dryness and fuel-load indicators, burn-severity indices from Sentinel-2, and terrain-aware risk overlays for wildfire-adjacent land.
Open by design
No closed hardware. No locked-up data.
Built on the open standards the drone and geospatial communities already use, so it works with the airframes and tools you have.
Simulation-first
The whole stack — planning, autonomy, telemetry, perception, analytics, and the operator console — is built and evaluated without drone hardware. The same code runs against an emulator, real autopilot firmware in simulation, and photoreal rendering.
Open protocols
MAVLink, MAVSDK, PX4 and ArduPilot from day one, behind a single drone interface. Moving from simulation to a real airframe is an adapter swap, not a rewrite.
Open data
Fields, imagery, and results live in open formats — STAC, Cloud-Optimized GeoTIFF, GeoJSON, PostGIS — and load directly in QGIS. The farm owns its data.
Status
Built in simulation, headed for the field.
LandRithm is in active development. The software is built and validated end to end in simulation — emulated flights, real autopilot firmware, photoreal rendered imagery — before any of it flies on a real drone.
- Mission planning, emulated flight, perception, maps and reportsDone
- Real autopilot firmware in simulation, orthomosaics, satellite baselinesDone
- Photoreal simulation and synthetic training dataIn progress
- Closed-loop re-surveyPlanned
- Bench, indoor, and outdoor validation on real aircraftPlanned
Contact
Growers, agronomists, researchers, drone operators — let's talk.
Interested in piloting LandRithm, collaborating on research, or just following along? We'd like to hear from you.
contact@landrithm.com