TerraSkyAI — ML & Computer Vision for Precision Agriculture
Machine learning and computer vision on drone imagery—off-type and weed detection, plant stand counts, and yield estimation for seed and crop companies
Overview
TerraSkyAI is an ML and computer vision platform for precision agriculture. Seed and crop companies fly drones over fields; models detect off-types and weeds, run plant stand counts, and estimate yield—so teams don’t walk every row by hand. Aerial images become map pins, QC-reviewed findings, and agronomy-ready reports. Live at terraskyai.com.
Project Details
Project Gallery
Key Results
The Challenge
Manual field scouting doesn’t scale. The hard problem was computer vision and ML on real drone imagery: reliable plant-level detection across growth stages and crops, counts and yield signals from photos, geospatial placement of every finding, and a human QC loop before anything goes back to the field.
Our Solution
We delivered TerraSkyAI as an ML / computer vision product with a full field-ops portal: • Core CV/ML pipeline on uploaded drone photos—object detection and analytics per crop and growth stage • Canola vision models: off-type / volunteer (VC) detection, early prescout, weed detection, plant stand count, yield estimation, male vs female bay areas • Potato: off-type detection, weeds (where allowed), plant count + yield together • Corn: plant count + yield; early prescout checks • Farms & fields on a map with flight history; flights record altitude/drone, photo upload, and analysis jobs • Results on the map—CV pins with crop photos; “clean” GPS images for coverage; reports and field stats • QC workflow—confirm, reject, or correct model detections before export/share for field follow-up • Pilots, alerts, cloud storage, roles/SSO, and day-to-day ops tools around the vision stack
The Results
Production ML and computer vision for AgTech at scale: • Drone imagery → detections, counts, and yield estimates without walking every row • Crop-specific CV models for canola, potato, and corn use cases • Map-grounded findings agronomists can trust after QC • Faster scouting and clearer follow-up points for seed/crop operations • Live platform: https://www.terraskyai.com/
Technologies Used
“TerraSkyAI’s computer vision and ML turn our drone flights into plant-level detections, stand counts, and yield signals—practical AI we can QC and take to the field.”
Agronomy Operations
TerraSkyAI