San Diego-based Tomnod combines crowdsourcing and machine intelligence to analyze massive datasets, particularly geospatial imagery. Some of their recent applications for crowdsourced imagery analysis include mapping building damage after the Christchurch earthquake and the search for the tomb of Genghis Khan. Most recently, the company partnered with Amnesty International and DigitalGlobe to monitor human rights violations in Syria.
The Genghis Khan work was done through a partnership with the National Geographic Society, gathering 10,000 contributors who made 1.8 million human analytical contributions on high-resolution Geo-Eye satellite imagery. The scale and the scope of this work won them a USGIF Academic Research Award.
Key to Tomnod’s technology are algorithms that rank the reliability and accuracy of individual contributors. With their CrowdRank technology, they can better train contributors as well as improve crowd-based analysis. In addition to their cloud-based toolsets for desktop work, they also have a Rapid On-the-Ground Response System (ROTGRS) that integrates unstructured details with analytics for mobile personnel in the field.
This winning combination of online repository, machine-guided analytics, and on-the-ground tools points to the future of geospatial intelligence gathering. It will be interesting to learn of their next project.
{ 2 comments… read them below or add one }
Thanks for the shout-out Matt!
Tomnod is very excited about the potential for distributed human computation — aka crowdsourcing — as a new tool for analyzing the billions of images, videos, photos etc. that are being collected around the world daily.
Since you asked, Tomnod’s current project is a bit more whimsical: we’re monitoring, analyzing and mapping every Tweet to find the spookiest Halloween costumes in San Diego!
http://www.signonsandiego.com/halloween-map/
This TweetMap technology will have applications for GeoInt too though: now we are crowdsourcing combine real-time, on-the-ground reports as well as overhead imagery analysis.
Thanks for the update, and good to see the new TweetMap application.