
Terrapattern is an open-source visual search tool for satellite imagery, designed to help journalists, citizen scientists, and researchers quickly identify and track specific visual features across large geographical areas. Utilizing machine learning techniques, it allows users to perform visual queries by clicking on a feature of interest, enabling the discovery of patterns that are often overlooked in traditional mapping. The tool aims to democratize access to geospatial intelligence, making it easier for non-experts to explore and analyze satellite imagery for sociological, humanitarian, scientific, and cultural insights. Currently, it supports searches in metropolitan regions of New York City, San Francisco, and Pittsburgh, covering over 2,200 square miles.

Terrapattern is an open-source visual search tool for satellite imagery, designed to help journalists, citizen scientists, and researchers quickly identify and track specific visual features across large geographical areas. Utilizing machine learning techniques, it allows users to perform visual queries by clicking on a feature of interest, enabling the discovery of patterns that are often overlooked in traditional mapping. The tool aims to democratize access to geospatial intelligence, making it easier for non-experts to explore and analyze satellite imagery for sociological, humanitarian, scientific, and cultural insights. Currently, it supports searches in metropolitan regions of New York City, San Francisco, and Pittsburgh, covering over 2,200 square miles.
Product: Reverse image/visual search engine for satellite and aerial imagery
Founders: Golan Levin; David Newbury; Kyle McDonald
Tech: Convolutional neural network for indexing and matching map tiles
Funding signal: Supported by the Knight Foundation Prototype Fund
Coverage: Searchable in metropolitan areas including New York City, San Francisco, and Pittsburgh
Geospatial search and analysis of satellite/aerial imagery for sociological, humanitarian, scientific, and cultural insights.
Geospatial / Mapping / Computer Vision
30000.00
“Supported by the Knight Foundation Prototype Fund”