Categories: GIS Software

Terrapattern: Search Engine for Satellite Imagery

Terrapattern is a way to search visually within satellite imagery.  Developed using an open source search tool, Terrapattern is the brainchild of Golan Levin, David Newbury, and Kyle McDonald who received funding from the John S. and James L. Knight Foundation Prototype Fund.

What Terrapattern does is perform “similar-image” searches in unlabeled satellite imagery using deep learning machine vision techniques.  To train the system, a deep convolutional neural network (DCNN) was fed hundreds of thousands of categorized OpenStreepMap imagery.  After the DCNN has learned which visual features are useful for classification, descriptions for millions of satellite imagery tiles were then computed.  Since searching can take a long time and a significant amount of RAM, the developers use the CoverTree algorithm to precompute relationships between the descriptions, reducing user search time to 1-2 seconds.

This approach allows users to search for particular features across a set of satellite imagery with nonbuilding structures (structures such a bridges, bus shelters, and swimming pools) and soft infrastructure yielding the best search results.  For example, users can search for images that show yellow crosswalks, purple tennis courts, grassy fields, and more.  Terrapattern lets users search visually for geographic features that may be ephemeral or not typically mapped out.  Results from searches are then exportable as GeoJSON files.


Terrapattern is currently only available for these four major U.S. cities: New York, Detroit, Pittsburgh, and San Francisco. An alpha prototype, the developers are limited by the enormous about of active RAM (about 10GB per major city) that terrapattern takes to store the model data.  The developers are interested in hearing back from “citizen scientists, data journalists, humanitarian researchers, and other domain experts to tell us about how our app is, or could be, of use.”  Fill out the survey to provide your input.

Visit: Terrapattern and access the code on GitHub.


  • Using Visual Search to Find Geographically Similar Features on Satellite Imagery

Recent Posts

R Packages for Spatial Analysis

Looking for R spatial packages?

3 weeks ago

Case Study: How Brindisi, Italy is Using GIS to Help Preserve its Ancient Architecture

The city of Brindisi, Italy deployed a solution from Pitney Bowes® that provides 80 to 90 percent faster performance, while…

4 weeks ago

Free Ways to Learn QGIS

If you want to start learning how to use QGIS, a free and open source GIS software package, there are…

4 weeks ago

Esri Announces Student Rate for Annual International User Conference

For $100, students can attend the entire five-day Esri International User Conference which takes place from July 8 to 12,…

1 month ago

First Data from NASA’s GEDI Program to 3D Map Forests

The first set of data that have returned have demonstrated GEDI is able to measure density of forests far better…

1 month ago

Missed the Latest QGIS User Conference? Access the Videos Online

Recorded presentations from the 2019 QGIS International User Conference are now available online.

1 month ago