Earth Index uses our large earth observation model to find places that are similar to each other and identify locations that have changed over time in similar ways. Using image embeddings, a "DNA sequence" which uniquely identifies features in a stack of imagery, the technology allows for efficient searching and classification workflows.
Leverage the power of multispectral data
From crop mapping to carbon biomass estimation, Earth Index can draw on descriptive multispectral data to find what you’re looking for.
Search over time
The world is dynamic. Earth Index is designed with this principal at its core, and easily allows for a user to monitor locations that have changed over time in similar ways.
Embeddings designed for structural searches
Built using the latest geospatial machine learning and computer vision models, Earth Index is designed to recognize spatial patterns in imagery.
Efficient image compression
Intelligent image compression allows for rapid searching, classification, and retrieval.
Flexible deployment on any cloud platform
Built with the intention to operate flexibly at scale using cloud-based platforms such as Google Earth Engine, GCP, and AWS.
Why we built it
Earth Genome generates environmental intelligence for decision makers. Earth Index was born from the need to more efficiently detect and monitor change on the planet. Our projects over the years have required finding things that have never been accounted for or things that change over time – discovering new mines in the Amazon, creating an inventory of all high concentration cattle operations in the United States, finding undocumented waste sites in Indonesia or distinguishing wheat fields from other types of agriculture in Ukraine. We purpose built the tools to search within an ocean of pixels – across space and time. Now, Earth Index allows us – or anyone – to build these kinds of datasets in a way that is better, faster, and cheaper than ever before.