AntMapper-based workflow for the crowdsourced Mapillary data preprocessing

Mapillary imagery is a novel crowdsourced data offering street-level imagery and GPS trajectory data simultaneously. Image inconsistency and measurement error of GPS trajectory data are two main issues obstacle Mapillary data applications in urban studies. In this paper, proposed workflow of crowdsourced Mapillary data cleaned random images and matched the GPS trajectory to correct road segment in Inner London by using AntMapper algorithm. This preprocessing work ensures the data quality and consistency of the crowdsourced Mapillary dataset for future spatio-temporal urban analytics in various scenarios.