AI for Earth technical resources

APIs

AI for Earth APIs allow application developers to accelerate their conservation applications with machine learning.

Land Cover Mapping

The Land Cover Mapping API leverages machine learning to provide high-resolution land cover information.

Species Classification

The Species Classification API enables developers to leverage deep learning models for recognizing plants and animals.

Camera Trap Image Processing

The Camera Trap Image Processing API accelerates biodiversity surveys by detecting and localizing animals in camera trap images.

Data

AI for Earth hosts key geospatial and conservation datasets on Azure, so that our grantees – and anyone else applying technology to conservation – can use the power of the Azure cloud to operate on global-scale environmental data. Do you have feedback about these datasets, or about other environmental datasets that would accelerate your work on Azure? Contact us at aiforearthdatasets@microsoft.com.

Infrastructure

Our grants program provides cloud-compute resources to hundreds of organizations. We provide open-source infrastructure and hands-on development support to help grantees go from a machine learning project to a web API running on Azure, so their research can be immediately consumed by other environmental scientists. This is the same infrastructure we use to host our own Web applications.

Open-source code

We are committed to making our work available for others to build on, by releasing code and trained models on GitHub.

Research

In addition to the research we support through our grants program, we work with collaborators in Microsoft Research to drive innovation in environmental science. Explore these AI for Earth projects at Microsoft Research.