Novel deep learning and GIS-based approach for road inventory survey

Abstract

The existing road inventory preparation methods are time-consuming, labor-intensive, inefficient, and there is no acceptable method for 3D urban visualisation. Accordingly, the study proposed a new cost-effective application to prepare road inventors & 3D urban visualiaation utilising deep learning technologies. The study comprised three main stages. In the first stage, the study conducted literature reviews. In the second stage, the study develops the application. Finally, the study validated the developed application by using Ranna as a case study. Further, the application recorded an accepted level of kappa accuracy. i.e., 92% & 90% for two models in the case study. transport planners and urban planners can employ the proposed application to prepare road inventors and 3D urban visualisation as the main contribution of this study.