AGRICULTURE
Yield Estimation
Yield Estimation. What it takes to run it.
Automated fruit/plant counting for accurate harvest forecasting.
// 01 · The question
“How much is out there, and when should the crew be booked?”
// 02 · Why it is hard
The count is the easy half. Converting visible fruit into total fruit means estimating what the canopy is hiding, and that ratio is not constant across varieties, training systems or years.
// 03 · What the labels have to be
Bounding box
Decided before the first frame
The output is a count, and a count needs instances that can be told apart, not outlines. Boxes on individual fruit give the number; the extra cost of masks buys precision the forecast never uses.
This is the decision that is expensive to reverse. The geometry has to match what the answer contains, and finding out it does not means labelling the set a second time.
// 04 · How we run it
Three parts of one loop, on this job.
LexAnnotate
Field Precision
Annotate crops, equipment, obstacles, and pathways with polygons, polylines, and semantic segmentation. Tag environmental factors like dust, shadows, and occlusions. 80k+ image annotations in production deployments.
LexInsight
Yield Intelligence
Transform plant-level detections into field-level intelligence. Map disease patterns across zones, track crop health progression over time, estimate yield from visual data.
LexAlert
Navigation Alerts
Monitor autonomous navigation models against seasonal drift. Retrain as fields change to keep navigation reliable through the seasons.
// 05 · What breaks it after launch
The whole estimate rests on a hidden-fruit ratio learned from one canopy density. A vigorous year hides more fruit behind more leaves, the visible count holds steady, and the forecast is quietly low across every block.
It is not the only one that can get this use case, it is the one that usually gets it first. All ten conditions.
// 06 · In agriculture
GPS-Denied Autonomy
The Challenge
Unreliable GPS under tree canopies and low visibility. Models failed in unstructured, obstacle-heavy terrains.
Our Execution
Annotated thousands of images with polygons and polylines. Tagged diverse environmental factors like dust, shadows, and occlusions.
The Result
Enhanced autonomy in GPS-denied environments. Faster deployment of AI-powered navigation.
80k+
Image annotations in production