Improvements on road centerline extraction by combining voronoi diagram and intensity feature from 3D UAV-based point cloud
MetadataShow full item record
CitationBiçici, S., & Zeybek, M. (2021, November). Improvements on Road Centerline Extraction by Combining Voronoi Diagram and Intensity Feature from 3D UAV-Based Point Cloud. In The Proceedings of the International Conference on Smart City Applications (pp. 935-944). Springer, Cham.
This study presents an application for road data users to make it easier to identify the centerline of roads. Images obtained from high-resolution unmanned aerial vehicles (UAV) provide a reliable database for fundamental applications such as road safety, road maintenance, traffic network, city planning, and vehicle navigation systems, thanks to accurate road extraction and centerline. Road extraction methods are based on algorithms that usually classify roads from 2D images. However, such data are difficult to provide high accuracy spatial information. Moreover, there are limitations for spatial information extraction problems for the road centerline. To overcome these limitations, we present a method to extract road centerline with 3D data based on point clouds that provide reliable information from UAV images. Commonly used three measures, namely Completeness, Correctness and Quality, for the road centerline extraction are 0.905, 0.999 and 0.905, respectively.