A Case Study of Crowdsourcing Imagery Coding in Natural Disasters

Abstract

Crowdsourcing and open licensing allow more people to participate in research and humanitarian activities. Open data, such as geographic information shared through OpenStreetMap and image datasets from disasters, can be useful for disaster response and recovery work. This chapter shares a real-world case study of humanitarian-driven imagery analysis, using open-source crowdsourcing technology. Shared philosophies in open technologies and digital humanities, including remixing and the wisdom of the crowd, are reflected in this case study.

Sustaining Citizen Science beyond an Emergency

“…Methods for module design in this open source citizen science platform are open source on GitHub [16,17], and an earlier publication [18] included examples of code for how different types of questions were structured. The underlying open source code had previously been adapted for use in light pollution [19] and forest [20] crowdsourced monitoring projects.…”

“…The researchers' conclusion referred to the kind of physics data for which CERN is best known; however, it applies as well to citizen science projects aiming for sustainability. This has been a focus of an earlier publication about GeoTag-X [18] and processes in which data shared in publications and reports were also open data shared on GitHub. Because this project was based on the open source Pybossa platform, with bespoke code shared on GitHub, knowledge about analysis workflows and processes can still be accessed despite the end of the research project.…”

“…Though GeoTag-X was retired some years after its European Commission FP7 project funding finished, a related H2020 project was funded, called E2mC (Evolution of Emergency Copernicus services). A theme overlapping both projects has been whether digital volunteers can help structure datasets of disaster information in ways useful to emergency responders [18,42,51,52]. This thread of research and knowledge is one of several in this field [53], not all of which have explicitly humanitarian goals.…”

“…Citizen science in disaster response can be approached with sustainable development and digital humanitarian mindsets [18] with open code; it can also be considered from security or defense perspectives [55] and using closed platforms [56]. If citizen science in disaster contexts aims to improve state responses to citizens' needs, then authorities' involvement in citizen science is implied in organizing or interpreting data in ways used by emergency response authorities.…”

Processing Collections of Geo-Referenced Images for Natural Disasters

“…There are some platforms such as Tomnod (http://www.tomnod.com), GeoTag-X [12], and some research works that address the problem of using volunteers for the processing of georeferenced images [9][13] [14][15] [16]. In particular, Tomnod of the company DigitalGlobe is using Artificial Intelligence (AI) driven by crowdsourcing to automatically identify the characteristics of interest in satellite and aerial images.…”