Identifying geomorphological changes of coastal cliffs through point cloud registration from uav images

Abstract

Cliff monitoring is essential to stakeholders for their decision-making in maintaining a healthy coastal environment. Recently, photogrammetry-based technology has shown great successes in cliff monitoring. However, many methods to date require georeferencing efforts by either measuring geographic coordinates of the ground control points (GCPs) or using global navigation satellite system (GNSS)-enabled unmanned aerial vehicles (UAVs), significantly increasing the implementation costs. In this study, we proposed an alternative cliff monitoring methodology that does not rely on any georeferencing efforts but can still yield reliable monitoring results. To this end, we treated 3D point clouds of the cliff from different periods as geometric datasets and further aligned them into the same coordinate system using a rigid registration protocol. We examined the performance of our approach through a few small-scale experiments on a rock sample as well as a full-scale field validation on a coastal cliff. The findings of this study would be particularly valuable for underserved coastal communities, where high-end GPS devices and GIS specialists may not be easily accessible resources.