Using Machine Learning to Reduce World Hunger
summary
Microsoft Research builds a low‑sensor precision agriculture system that predicts temperature moisture pH and leaf wetness from aerial imagery plus sparse ground probes, cutting reliance on $1,000 per sensor deployments by learning visual and spatial smoothness. Hardware includes waterproof sensor boards with color and infrared cameras, a drone flight planner and tethered helium balloons for imagery, and a long‑range link using TV white spaces. Pilots run in Kenya and ML models successfully infer leaf wetness from two cheap sensors and imagery.
AI-written summary · about
chapters (7)
- 0:00Precision agriculture challenge
- 1:27Sensor costs and infrastructure
- 3:12Connectivity with TV white spaces
- 4:57Drones versus balloons
- 6:22Sensor prediction with ML
- 8:08Pilot results and future
- 9:37AI for Earth and closing
related talks
28mHarnessing Machine Learning to Improve Agricultural Resilience in Africa
31mBridging IoT and Machine Learning with Python for Sustainable Agriculture
19mPlant Factory: Sensor, Data, Machine Learning
28mGame of Drones
25mFrom seed to store: AI to optimize the indoor farms of the future
21mUnderstanding World food economy with satellite
15mThe tech is there: society needs to step up
26mAI, big data, digital and future of agriculture