3D Coastal Monitoring from very dense UAV-Based Photogrammetric Point Clouds

Abstract

In the present study, the potential use of unmanned aerial vehicles (UAVs) as a platform to flexibly obtain sequence of images along coastal areas from which producing high quality SfM-MVS based geospatial data is tested. A flight campaign was conducted over a coastal test site covering an area of around 4 has near Malaga (Spain). Images were taken on 1st December 2015 at a height above the ground ranging from 113.5 to 118 meters by using a Sony α6000® consumer camera mounted on a UFOCAM XXL v2® octocopter. 40 RTK-GPS surveyed ground points were evenly distributed over the whole working area. Furthermore, a very dense and accurate point cloud was collected by using a FARO Focus 3D X-130 terrestrial laser scanner (TLS). The photogrammetric block was computed by using two widely known SfM-MVS commercial software implementations such as Inpho UASMaster® and PhotoScan Professional®. PhotoScan provided a highly accurate bundle adjustment with errors of 1.5 cm, 1.5 cm and 6.1 cm along X, Y and Z axis respectively. The triangulation errors computed from UASMaster turned out to be slightly poorer along Z axis. In this sense, the very high resolution Surface Model built up from the corresponding photogrammetric point cloud depicted higher Z-differences with respect to the reference TLS derived surface model in the case of the UASMaster workflow. Summing up, the high degree of automation and efficient data acquisition provided by UAV-based digital photogrammetry makes this approach competitive and useful to be applied in high resolution 3D coastal mapping.