In the field of Remote Sensing, Unoccupied Aerial Vehicles (UAVs) are emerging as powerful tools in the management of natural resources and landscapes. Imagery collected using UAVs, and digital elevation models that are derived from UAV imagery facilitate landscape mapping at temporal and spatial scales that are difficult to achieve using traditional satellite-based Remote Sensing techniques. The quality of UAV-derived imagery and associated datasets varies based on data collection methods and data processing approaches. In this research the accuracy of elevation data derived from UAV imagery processed using three different methods was assessed at River Ridge Ranch in Springville, CA. UAV imagery was collected in this remote setting using a senseFly eBee Plus with S.O.D.A. photogrammetric camera. Three levels of post-processing were explored (1) Level 1 standalone processing, (2) Level 2 PPK processing and (3) Level 3 PPK and GCP processing to achieve the best possible accuracy without the need for ground control points (GCPs). In addition to evaluating the accuracy of UAV derived elevation data, a two-part multivariate statistical analysis incorporating logistic regression and k-means clustering was performed to analyze the relationship, if any, between in-field slope and UAV-derived slope.
This research provides a methodology that can be consistently applied at the River Ridge Ranch study area to generate highly reliable high spatial and temporal resolution data.
Results demonstrate that UAVs with direct georeferencing methods outperform that of indirect georeferencing methods, but the most accurate products can be generated through pairing both indirect and direct methods. By using both indirect and direct georeferencing, this study was able to produce high spatial resolution imagery and derived terrain surfaces with vertical accuracies of 91 centimeters. This study also found that UAV-derived elevation accuracies varied based on terrain and cover type with low slope (RMSE: 1.13 m, MAD: 1.01 m) and open canopy points (RMSE: 1.09 m, MAD: 1.02 m) proving to be less accurate than severe slope (RMSE: 0.94 m, MAD: 0.78 m) and closed canopy points (RMSE: 0.84 m, MAD: 0.71 m). The study was unable to use UAV-derived slope values to predict human slope classification categories identified in the field. The human perspective on in-field slope gradients bears further exploration. This may be possible through further exploration of and improvements in UAV-derived slope data.
|Commitee:||Wechsler, Suzanne P., Winslow, Scott|
|School:||California State University, Long Beach|
|School Location:||United States -- California|
|Source:||MAI 81/3(E), Masters Abstracts International|
|Subjects:||Geography, Remote sensing|
|Keywords:||Accuracy, DEM, DTM, Geography, Remote sensing, UAV|
Copyright in each Dissertation and Thesis is retained by the author. All Rights Reserved
The supplemental file or files you are about to download were provided to ProQuest by the author as part of a
dissertation or thesis. The supplemental files are provided "AS IS" without warranty. ProQuest is not responsible for the
content, format or impact on the supplemental file(s) on our system. in some cases, the file type may be unknown or
may be a .exe file. We recommend caution as you open such files.
Copyright of the original materials contained in the supplemental file is retained by the author and your access to the
supplemental files is subject to the ProQuest Terms and Conditions of use.
Depending on the size of the file(s) you are downloading, the system may take some time to download them. Please be