The designing of road networks in civil engineering requires significant labor and is often time consuming. This report proposes the utilisation of deep learning techniques, leveraging high-resolution aerial imagery, to classify road surface areas from non-road areas. The objective of this report is to accurately estimate road surface area by calculating distance from a road centerline to the road edge, which will assist municipalities in efficiently assessing and managing their road assets.
Ofentse Menwe
Honours Student
Dr Clement Nyirenda
Project Supervisor
Kevin Johnson
Project Co-Supervisor