Master of Science
Urban forests are invaluable to residents of cities who have limited access to green space, but these forests require continuous monitoring to ensure their health. With limited resources to conduct frequent monitoring on foot, remote sensing and Geographical Information Systems (GIS) could offer a timely and cost-effective solution. Shelby Farms Park in the city of Memphis, Tennessee, boasts 1500 acres of forest, but a 2010 ground survey found Chinese Privet (Lingustrum sinense) was impeding forest health. In this study UAS imagery and publicly-available LiDAR were used to monitor the forests of this urban park and detect change. Two different methods of supervised classification were used on UAS imagery to try and detect privet. Support vector mean (SVM) classification had a Kappa value of 0.087, and Random Trees classification had a Kappa value of 0, so neither method would be recommended for future use without improvements. The metrics tree height, tree count, and canopy cover were extracted from 2011 and 2017 LiDAR. Differences in tree heights between 2011 and 2017 were found to be significant with a p-value <0.01, but there was no correlation between tree growth and the prevalence of privet. Fusion of 2021 UAS digital surface models (DSM) and 2017 LiDAR digital elevation models (DEM) allowed for tree measurements comparable to LiDAR. While classification of UAS imagery to detect privet was unsuccessful, there are still opportunities for improvement that could make these methods useful in the future. The fusion of UAS DSM’s with LiDAR DEM’s allow for frequent forest monitoring in years that publicly-available LiDAR was not accessible.
Dissertation or thesis originally submitted to ProQuest.
Lennon, Aileen, "Using Open-access LiDAR and UAS Imagery to Survey the Forests of Shelby Farms Park" (2022). Electronic Theses and Dissertations. 3291.