Authors: Sara Durgan*, Florida Atlantic University, Caiyun Zhang, Florida Atlantic University
Topics: UAS / UAV, Remote Sensing, Environmental Science
Keywords: wetland, drone, UAV, terrain correction, DEM, photogrammetry, structure from motion
Session Type: Paper
Start / End Time: 3:05 PM / 4:45 PM
Room: Taylor, Marriott, Mezzanine Level
Presentation File: No File Uploaded
Accurate three-dimensional models, including bare-earth digital elevation models, are essential for understanding the biophysical settings of wetland ecosystems and forecasting future geomorphic change. Light Detection and Ranging (LiDAR) is the current data standard for generating three-dimensional models of terrain and surface features. Despite its advantages, LiDAR data acquisition remains laborious and expensive; therefore data availability is limited in the Everglades. Advances in unmanned aerial vehicle (UAV) technology and Structure from Motion photogrammetry present the ability to generate high-resolution data products frequently and at a low cost. However, a major challenge for generating terrain models using Structure from Motion photogrammetry is the inability to penetrate dense vegetation canopies that exist in these environments. The objective of this research is to develop a species-based terrain correction procedure for a UAV-derived digital elevation model for a restored wetland in the coastal Everglades. Contemporary remote sensing techniques including object-based image analysis (OBIA) and machine learning algorithms are tested and compared for the correction procedure. The corrected UAV-derived digital elevation model is compared with a terrain model generated using the best-available LiDAR in terms of precision and accuracy. The results of this research provide a method to increase UAV data integrity in wetland environments to guide the utilization of these data products in coastal wetland research and management.