Remote sensing from unmanned aerial vehicles for 3D urban modelling: case study of Loulé, Portugal

Abstract

Data acquisition using Unmanned Aerial Vehicles (UAVs) is currently a fast and low-cost Remote Sensing approach. UAVs have become standard platforms for applications in photogrammetric data capture. This capture is still only feasible for small areas, around dozens of hectares. In view of our description, we can say that the data obtained using UAVs are useful for 3D urban modelling, i.e., buildings and urban areas, with a view to support large scale urbanism and urban planning studies. The data were gathered from RGB UAV imagery at a height of 100 meters covering the city centre of Loulé, Portugal. To conduct the UAV survey the following steps were followed from data acquisition through processing 3D urban models: a) establishing flight plan parameters (height, overlapping, time, focal distance, camera data, spatial resolution) and collecting data survey (regarding preparation aspects such as safety, interferences, and calibration); b) loading photos; inspecting loaded images and removing unnecessary images; c) aligning photos; d) building dense point clouds; e) building a mesh; building a 3D polygonal model; f) texturing the 3D model; g) calculating DEM and DTM; h) creating an orthoimage; and i) exporting results. After processing the data, we arrived at two highly accurate 3D urban models. Taking these results into consideration, we have integrated them into urban plans in order to show their usability in the planning process of a city.