LAUSR

ABSTRACT The quick and accurate identification of post-earthquake rescue objects can minimize the casualties and property loss caused by earthquakes. With the rapid development of remote sensing technology, rescue objects can be identified through high-resolution remote sensing images. There are two main categories of approaches to identify rescue object through high-resolution images: automatic extraction by a computer and visual judgment by professionals. Although results can be obtained quickly by using automatic extraction, the accuracy of the results is unacceptably low. For visual judgment, the large demands for time and professionals restrict its wide practical application. In this study, we introduce crowdsourcing into the identification of post-earthquake rescue objects. First, we integrate the advantages of the computer and crowdsourcing, which means that the computer takes advantage of the speed of information processing, while crowdsourcing makes full use of human recognition capabilities. Second, we take visual judgment tasks out of the hands of professionals and entrust the tasks to workers in a crowdsourcing platform. Not only are the human resources infinite, but we can also improve the efficiency of identifying rescue objects. Third, we propose a crowdsourcing model that improves the quality of the results and saves human resources. Finally, experimental results demonstrate that our solution is feasible.