Developing Bare-Earth Digital Elevation Models from Structure from Motion Data on Barrier Islands

Submitted by atripp on Thu, 06/03/2021 - 09:57

Barrier islands are dynamic environments. Under calm conditions, they are gradually shaped by currents, waves, and tides; during hurricanes and other extreme storms, they can evolve within hours to days. Unmanned aircraft systems (UAS) can collect monitoring data, especially elevation information, via Structure from Motion (SfM) techniques, with a temporal resolution that is well-suited for these dynamic environments. However, SfM data represent the surface elevation, which includes the vegetation canopy. Additional processing is required for estimating bare-earth elevation, which is critical for understanding island geomorphology.

As part of the Coastal Resource Evaluation for Management Application (CREMA) project, scientists from the USGS Wetland Aquatic Research Center and St. Petersburg Coastal and Marine Science Center are developing methods to develop bare-earth digital elevation models from SfM data acquired during repeated surveys on Dauphin Island, Alabama. Specifically, the team is using vegetation and elevation surveys collected using a Real-Time Kinematic Global Positional System between September 2018 and April 2019 to develop a vegetation correction model. The model will be applied to co-located and concurrent SfM data and multispectral information collected from periodic UAS surveys over two sites on Dauphin Island to produce bare-earth digital elevation models. The team is testing several machine learning algorithms to predict and remove elevation bias caused by the presence of vegetation, including k-nearest neighbor, support vector machine, and random forest, and several gridding algorithms. The input datasets are spectral indices from UAS-based multispectral imagery and landscape position information (e.g., relative topography, elevation, and distance from shore). This effort, which is portable to other islands after any necessary calibration, could provide an efficient and repeatable framework for monitoring island geomorphology and vegetation canopy height, which are valuable data sources for tracking changes to these critical environments.

(a) Map of Dauphin Island, Alabama, (b) an example of a vegetated transect with multispectral imagery collected via an unmanned aircraft system, and (c) topographic profiles along the transect in panel b for digital elevation models (DEMs) generated using various gridding methods on the structure from motion point cloud and including a both a 15-centimeter and 50-centimeter DEM.

Author Name
Nicholas M. Enwright
Author Email