3D UAS-SfM Point Cloud Classification in Urban Terrestrial Environment

EasyChair Preprint 10655

Abstract

The use of lidar technologies for point cloud acquisition is financially costly. However, photogrammetry using unmanned aerial systems (UAS) imagery and structure-from-motion (SfM) techniques have proven a practical and costeffective way to collect point cloud data (Liu and Boehm, 2015). SfM uses two-dimensional (2D) images to produce high-quality three-dimensional (3D) point clouds. Classified point cloud data is useful in environmental modelling, cultural heritage preservation and navigation applications (Grilli et al., 2017; Roynard et al., 2018; Croce et al., 2021). This study focuses on classifying a 3D UAS-SfM point cloud of a heterogeneous urban environment into three land cover categories: ground, high vegetation and buildings.

Keyphrases: Classification, Mapping, Structure from Motion, Unmanned Aerial Systems, point cloud

@booklet{EasyChair:10655,
  author    = {Simiso Ntuli and Angus Forbes and Mayshree Singh and Mulemwa Akombelwa},
  title     = {3D UAS-SfM Point Cloud Classification in Urban Terrestrial Environment},
  howpublished = {EasyChair Preprint 10655},
  year      = {EasyChair, 2023}}