Challenge

There is a data gap in agriculture that prevents the agribusiness industry from making informed decisions, forcing them to farm reactively instead of proactively. Current methods of taking physical samples to a lab or using satellite imagery are expensive and inefficient, resulting in higher greenhouse gas emissions that contribute to climate change. Growers need a verifiable system to analyze crop and soil nutrients, predict yields, and monitor carbon sequestration in soil at scale.

Solutions

Cloud Agronomics scans crops and soil using custom hyperspectral imaging apparatuses on manned aircraft, collecting 300 times more data per pixel than satellites. The data is sent to Azure, where georeferencing, calibration, and analysis algorithms convert the raw data into insights. Cloud Agronomics is building one of the largest tagged datasets for agriculture, enabling growers to manage crops proactively and efficiently. In addition, real-time carbon monitoring will give carbon credit marketplaces the first standardized carbon index to provide a financial incentive to transition to carbon farming, which is the practice of removing excess carbon from the atmosphere and storing it in soil to aid plant growth.

How Cloud Agronomics uses Azure

  • Blob Storage to store calibrated hyperspectral and satellite data for geospatial querying
  • Databricks for pre-processing and orchestrating machine learning
  • Container Registry to provide the storage and registration for containerized applications
  • Container Instances to run containerized applications, including ML models, visualization, and user-facing apps
  • Machine Learning Service for real-time analytics for data science R&D
  • Kubernetes Service for orchestration and cluster management for processing calibrated hyperspectral imagery
Fog on a field bordered by trees.

Predictive insights for agriculture powered by geospatial imaging

Stalk of wheat at sunset.