Crowdsourcing Based Event Reporting System Using Smartphones with Accurate Localization and Photo Tamper Detection

Abstract

Crowdsourcing provides a novel and efficient paradigm to leverage numerous smartphone users to report timely information (time, location, photo, etc.) about events, enabling various urban sensing applications such as environment monitoring and public space management. Localization and photographing are two of the most important function modules in crowdsourcing based event reporting systems, which are used to tell people where and what events happen. However, the existing systems tend to localize the user instead of the event, resulting in the poor accuracy. Meanwhile, the existing event localization approaches need either complex user operation or complex computation process. For this reason, we propose a simple, effective, and light-weight approach for event localization. On the other hand, the existing systems cannot guarantee the photo authenticity, as the emerging various image editing and processing softwares have made the (malicious) photo tampering easier and easier. Accordingly, we design a novel smartphone-based photo tamper detection approach with high security. Detailed system implementation and performance evaluation are provided to verify the effectiveness of our proposed approaches.