L3.10: Assignment

PrintPrint

Crowdsourcing: Turning to the Crowd for Help

Tomnod is an example of a crowdsourcing platform that is used to extract GEOINT data from satellite imagery. DigitalGlobe’s constellation of satellites captures high-resolution images of an area the size of India every day. Tomnod expedites the analysis of this huge amount of satellite imagery by engaging a large public crowd. For example, millions of people took part in a Tomnod campaign to search for signs of the missing Malaysian Airlines flight MH370.

Tomnod responds to global events like natural disasters, security incidents, search and rescue missions, or mapping remote populations. Tomnod loads current satellite images onto their website where they can be explored by a global community of volunteer image taggers. Each volunteer scans a subset of the imagery, pixel by pixel, and places a “tag” over objects of interest. Within hours, the crowd can cover thousands of square kilometers many times over. By gathering multiple, independent views of every location, crowdsourced consensus begins to emerge. Even though most individuals are novice imagery interpreters, the “wisdom of the crowd” converges on the locations that are most important. Using a statistical geospatial algorithm, this consensus can be extracted and used to focus the efforts of expert analysts (known as “tipping and cueing”) or provided to responders on the ground.

see text version of image
Figure 3.16: The process of turning raw crowd tags into actionable information
Source: Tomnod

Try Tomnod for yourself!

See how Tomnod helped to track the damage from the Boles Fire near Weed, California. The Boles Fire started on September 15, 2014, southeast of Weed, California. The wind-driven fire quickly moved into town destroying 100 homes and forcing 1500 people to evacuate. DigitalGlobe’s satellites collect imagery in the infrared spectrum in addition to the red/green/blue colors visible to the human eye. Using such infrared imagery, in this Tomnod campaign the images are “false-colored” which highlight the burned areas in black, while areas of healthy vegetation appear red.

Use the following link to access the Tomnod Boles Fire near Weed, California site.