Water · drainageStanding water
Low-lying zones holding water after rainfall. Review drainage, tile, and low-grade corners.
Know what's wrong in every field before you walk it.
Upload the drone flights you already fly. LandOS returns a per-field anomaly report — planter skips, waterway damage, standing water, nutrient stress, weed clusters. Nine classes per pass, a per-quadrant breakdown, and a priority action list ranked by urgency.
Raw flightAnomaly map
Scouting can't cover the acreage.
You fly the drones. You save the imagery. Nobody has time to interpret it. By the time a weed flush or a drainage failure is visible from the ground, the window for cheap intervention has closed.
Your agronomist walks 20 acres a day.
Your operation has 2,000. Ground scouting covers a sliver, late.
The field report is a Slack thread.
Notes from last week's walk don't feed next week's spray plan. By harvest, nobody remembers which quadrant had the waterway damage.
NDVI tells you nothing actionable.
A vegetation index flags that a spot is stressed. It doesn't say whether it's waterway damage, a planter skip, or nutrient deficiency — each with a different response.
What you get
A nine-class anomaly report from every flight.
Upload a folder of drone flights. Get back a per-field report with nine anomaly classes, a per-quadrant breakdown, and a prioritized action list keyed to the kind of problem each class represents.
01
A per-field anomaly map
Every pixel scored against nine classes. Toggle overlays in the viewer, export per-class PNGs.
02
A quadrant breakdown
NW / NE / SW / SE coverage per class. Tells your agronomist exactly where to walk first.
03
A priority action list
Each class tagged urgent / high / medium / info — replant, insurance, scouting, prescription.
How it runs
Upload a flight. Get the report the same day.
You fly the drones. You upload the orthos. We run the inference and hand back a report you can act on.
- 01
Upload the flight.
Drop a folder of RGB orthos into the viewer. Your file structure is fine.
- 02
We run inference.
A SegFormer model fine-tuned on Agriculture-Vision. Typical fields process in minutes.
- 03
Anomaly map out.
Per-pixel nine-class segmentation. Per-quadrant stats. Per-class PNG exports.
- 04
Scout what matters.
A priority action list keyed to the class — replant, drainage, scout, spray.
What ships today
A working demo.
LandOS for Ag runs a SegFormer model fine-tuned on Agriculture-Vision. The demo you sign into is the live product, pointed at five sample fields that each illustrate one of the nine classes.
Nine classes today.
Planter skip, waterway, standing water, nutrient deficiency, storm damage, double plant, drydown, endrow, weed cluster.
Same-day turnaround.
Drop orthos in the morning; read the report after lunch.
Deterministic report.
Per-quadrant coverage stats plus a priority-ordered action list. No LLM, no surprises.
Detections
What LandOS flags on every field.
Real output from five sample fields, one class per tile. The same nine-class report is what you get back on your own flights.
Water · drainage
Weed cluster · scoutWeed clusters
Concentrated weed pressure, per-quadrant. Feeds the walk-and-spray priority list for the ground rig.
ND · VRT candidateNutrient deficiency
Deficient zones for agronomy scouting. Candidate for a variable-rate prescription next pass.
Waterway · post-seasonWaterway damage
Erosion or washout on established waterways. Post-season tile and grading inspection recommended.
Nine classes today: planter skip · waterway · standing water · nutrient deficiency · storm damage · double plant · drydown · endrow · weed cluster.
Fields in the demo
Five fields, nine classes.
Each tile is the composite nine-class overlay from one of the demo fields. Sign in to toggle any single class, see per-quadrant stats, read the action list.
See LandOS on one of your fields.
Send us a recent flight — or a folder of RGB tiles. You get a nine-class anomaly map and a per-quadrant action list from your first upload.