The startup, Earth Observing Systems (EOS), uses NVIDIA GTX 1080 GPUs, and NVIDIA Tesla V100 GPUs on the Amazon Cloud, with the cuDNN-accelerated TensorFlow deep learning framework to train their algorithm on both historical and current observations, including satellite imagery and historical data. Once trained, EOS implements the deep learning system to calculate crop conditions in a particular area of interest.
The ability to measure crop production rates provides valuable insight for supply chain customers, the startup mentioned on their site.
“[We] developed a multi-level deep learning architecture that targets land cover and classifies crop types from multi-temporal, multi-source satellite imagery,” according to the company’s press release. “The key element of the architecture is an unsupervised neural network that is used for optical imagery segmentation and missing data restoration.”
The company says their analytical tool allows clients and partners to get a real-time view of the field’s performance.
Farmers, suppliers, commodity traders, crop insurance companies, and all members of the agriculture supply-chain can now use the EOS tools to support their businesses.
Read more >
AI Helps Farmers Predict Crop Production
Apr 02, 2018
Discuss (0)
Related resources
- GTC session: Transforming Agriculture with AI and Computer Vision
- GTC session: Democratizing AI for Agriculture: Bridging the Digital Divide
- GTC session: Predicting El Nino: Machine Learning’s Role in Advancing Complex System Understanding
- NGC Containers: MATLAB
- Webinar: Move Enterprise AI Use Cases From Development to Production With Full-Stack AI Inferencing
- Webinar: Accelerate AI Model Inference at Scale for Financial Services
About the Authors
