Making Green Transport a Reality: A Classification Based Data Analysis Method to Identify Properties Suitable for Electric Vehicle Charging Point Installation

Abstract

With Electric Vehicles (EVs) emerging as the dominant mode of green transportation in the UK, it is critical that local authorities and urban planners can accurately map the existing EV infrastructures in place. In this paper, we demonstrate a novel data processing pipeline to analyse remotely sensed image data to highlight areas of a city most suitable for EV infrastructure. By applying deep transfer learning to multiple datasets, we are able to identify individual addresses suitable for the installation of home EV charging points. Using this same methodology, we also highlight areas where community charging points would be most effectively installed. We improve on previous methods by integrating topographical data, Census data, and remotely sensed image data to achieve a fully automated system capable of large-scale surveying of external building characteristics.