V
Loading...
Loading...
Loading...
Saving Harvests from Space
Computer vision models analyzing satellite imagery detect crop diseases weeks before visible symptoms, saving harvests for over 100,000 farmers across East Africa.
40%
Yield Increase
500K
Hectares Monitored
100K+
Farmers Served
3 weeks
Early Warning
Crop diseases cause average losses of 30-40% annually for African farmers, with some outbreaks destroying entire harvests. Early detection is key to prevention, but traditional scouting methods can't cover the vast areas farmed by smallholders.
Using satellite imagery and computer vision, we've built an early warning system that detects crop disease indicators weeks before symptoms are visible to the human eye. Farmers receive alerts on their phones, allowing them to take preventive action and save their harvests.
The system monitors over 500,000 hectares across Kenya, Tanzania, and Rwanda, serving more than 100,000 farmers with timely disease alerts.
Understanding the unique obstacles we're working to overcome.
By the time disease symptoms are visible, it's often too late for effective treatment.
Smallholder farms are scattered across vast areas impossible to scout manually.
Frequent cloud cover in tropical regions limits satellite imagery availability.
Limited labeled data for training models on African crop diseases.
The methods and techniques we've developed to address these challenges.
Using spectral bands beyond visible light to detect plant stress before visible symptoms.
Tracking changes over time to identify anomalous vegetation patterns.
Using radar imagery to see through cloud cover and monitor crop conditions.
Network of agronomists and farmers providing validation data for model training.
Measurable outcomes from our research and deployments.
40%
Yield Increase
Farmers receiving alerts see 40% higher yields compared to control groups.
500K
Hectares Monitored
System monitors over 500,000 hectares of farmland across East Africa.
100K+
Farmers Served
Over 100,000 smallholder farmers receive disease alerts via mobile.
3 weeks
Early Warning
Alerts provided an average of 3 weeks before visible disease symptoms.
Okafor, Z., Banda, M., et al.
Scaling the disease detection system across East Africa with FAO partnership.
Partners:
FAO, CGIAR, Kenya Agricultural Research Institute
Zara Okafor
Product Lead
Michael Banda
Remote Sensing Lead
We're always looking for collaborators, partners, and talented researchers to advance this work.