Terrapattern is the first visual search engine for satellite imagery. It is created by Golan Levin, David Newbury, and Kyle McDonald and with the funding support from John S. and James L. Knight Foundation. Terrapattern is an open-source search engine. Terrapattern is very fast and accurate in providing the similar visual search results. The important part is that you can download the list of locations in the GeoJSON format.
Terrapattern returns the relevant satellite visual images when clicked on any featured image of your interest or location searched by you. For example, you searched the schools bus-station in PITTSBURG, than it will return the similar visual search results from the PITTSBURG along with additional information like Geographical plot, Similarity Plot. These plots give you the insight that how many such similar plots are there. The results are based on the visual features, color and most important it mathes the search results based on the description which sometimes does not give accurate results. Terrapattern is based on Deep Convolutional Neural Network (DCNN) which is feeded by thousands of Google satellite images and Cover Tree algorithm to help increase the search accuracy.
Currently Terrapattern is only limited to few cities like PITTSBURG, SAN FRANCISCO, NEW YORK CITY, DETROIT. Terrapattern says more cities would come soon. The visual of terrapattern are based on Google satellite images. Are you gonna use it if your city is added? Share your review. Know the comprehensive list of search engine of 2016
Keep sharing. Stay tuned to Tech Tunes.