Communications of the ACM
Volunteers and Algorithms Need Training to Find Mh370
Crowdsourcing has emerged as a way of recruiting the public to help review satellite imagery to try to locate the missing Malaysia Airlines Flight 370.
Credit: EPA
As the search continues for Malaysia Airlines Flight 370, crowdsourcing has emerged as a solution that could enable the public to help scan satellite imagery to find the missing jetliner.
U.S.-based satellite imagery firm DigitalGlobe has created a website that enables the public to study satellite images square by square, and more than 2 million people worldwide are participating.
Disaster responders are increasingly relying on images from satellites, planes, and volunteers to map out important features of an area. Volunteers can then access these images through Web-based platforms and collaboratively identify specific elements in the pictures. During the Haiti earthquake response effort, for example, satellite maps helped identify collapsed buildings and navigable routes.
Although mapping projects been successful, search efforts are far more difficult, requiring good eyesight, an understanding of debris features, and contextual knowledge such as the angle from which the picture was taken.
Large crowdsourcing efforts require a combination of human computation techniques and machine-learning algorithms, with an initial training phase before the search begins, writes University of Southampton lecturer Sarvapali Ramchurn. In addition, Ramchurn says volunteers become less accurate over time with repetitive tasks, and should take breaks and be tested periodically for accuracy.
From The Conversation
View Full Article
Abstracts Copyright © 2014 Information Inc., Bethesda, Maryland, USA
No entries found