• Subscribe

iSGTW Feature - A picture is worth a thousand words

Feature - A picture is worth a thousand words . . . or a single image query


Image courtesy EarthPerspectives.com

What if you could enter an image, or a tiny piece of one, instead of text to search a database for matching images?

Search engines allow you to enter text to query their databases. The engine finds and returns to you documents that match your input text. For image searches, it tries to match text that describes the image, its origins, its format or some other attribute; it can't match the image itself.

But the Geospatial Information Retrieval and Indexing System, or GeoIRIS, could change this. Created by EarthPerspectives, it will let users easily query satellite imagery databases using images or image fragments rather than generic textual metadata. Currently in the prototype stage, the GeoIRIS image content search engine was developed at the Center for Geospatial Intelligence (CGI) at the University of Missouri at Columbia. The U.S. government could begin using GeoIRIS for intelligence purposes as early as the next few years.

Satellite image analysts try to locate certain objects or features in a large volume of imagery. These are often small areas of densely-packed photographic images - the proverbial needle in a haystack. For instance, if a homeland security officer wanted to know the location of all large storage tanks in a specific geographic area that might hold oil, gas or petroleum, she could submit an image chip - a small area of an image - containing a storage tank as query input to GeoIRIS. The search engine would then search the database and return objects that match the image content.

GeoIRIS v4 supports Query-By-Example using either image regions (tiles) or objects - such as planes, storage tanks, swimming pools, etc. This screenshot shows one query result of many; results are ranked in order of "best fit," and metadata about each result can be displayed. Image courtesy earthperspectives.com.

"Rather than having to look at all of the imagery chip by chip or frame by frame, analysts can use GeoIRIS to quickly and easily locate imagery of interest through an online interface," said Curt Davis, director of CGI.

For the search engine to work, however, all of the imagery in the database has to be processed offline beforehand and indexed. CGI researchers have completed this step for their first data set. They started by dividing the large satellite images into tens of thousands smaller image chips. Then they used Appistry Enterprise Application Fabric* (EAF) to manage the distribution of the chips across the computing nodes, where features and objects were extracted from them. Appistry EAF allows researchers to easily manage feature extraction tasks on computer clusters without any specialized knowledge of computing. Once these computations were completed, the Appistry software collected and reassembled the results.

"We didn't have to write any process orchestration functions," Davis said. "It really made the distribution and collection processing for GeoIRIS a lot easier."

-Amelia Williamson, for iSGTW

*Appistry EAF was renamed Appistry CloudIQ Platform 4.0 on March 9

Join the conversation

Do you have story ideas or something to contribute? Let us know!

Copyright © 2019 Science Node ™  |  Privacy Notice  |  Sitemap

Disclaimer: While Science Node ™ does its best to provide complete and up-to-date information, it does not warrant that the information is error-free and disclaims all liability with respect to results from the use of the information.

Republish

We encourage you to republish this article online and in print, it’s free under our creative commons attribution license, but please follow some simple guidelines:
  1. You have to credit our authors.
  2. You have to credit ScienceNode.org — where possible include our logo with a link back to the original article.
  3. You can simply run the first few lines of the article and then add: “Read the full article on ScienceNode.org” containing a link back to the original article.
  4. The easiest way to get the article on your site is to embed the code below.