Mobile road pothole classification and reporting with data quality estimates

Abstract

Harnessing the power of mobile computing platforms has opened up newer possibilities of gathering and classifying data by leveraging the use of crowd-sourcing. In the current generation that is being dominated by the mobile computing platform namely smart-phones, crowd-sourcing can be achieved in a relatively hassle-free yet effective means of collecting data from a large set of ordinary users. Tapping into this source pool, however, has a drawback of unspecified data quality and security and hence the lack of trust in the data collected. This paper proposes an approach that aims to realize the most influential factors related to destructiveness of potholes that are encountered on roadways, while supplementing the data with quality estimates derived from the completeness of the data and security and quality factors of the input device itself that is involved in the data collection process, thereby addressing the drawback of data trust. The mobile application design is described. The application use cases are presented and discussed.