Skip to main content


Crowdsourcing methodology developed

Research Achievements

Crowdsourcing methodology developed

We have developed a crowdsourcing methodology, using large-scale public participation for remote search and discovery. A first of its kind for National Geographic, Field Expedition: Mongolia encouraged online participation by the public, known as human computation, or crowdsourcing. Using high-resolution satellite images donated by GeoEye, citizen scientists analyzed real-time data, maps and other information to identify anomalies of archaeological relevance, using a novel interaction mechanism with immediate mode data flow and feedback. Combined with a live blog and reports from the field, this approach inspired thousands of online participants to create nearly 2 million human tagged data points that are being used to train supervised computers vision algorithms to identify ancient structures. A field team ground-truthed the tagged sites, using non-destructive, digital-archaeology techniques, creating an important baseline for the accuracy of crowd-based mapping and image analysis.