Mirka, Blair

Stow, Douglas

An, Li
Lewison, Rebecca

2020-06-26

Spring 2020

Thesis

56 pages

An important component of wildlife management and conservation is monitoring the health and population size of wildlife species. Monitoring the population size of an animal community can inform researchers about individual health, potential changes in habitat, and effectiveness of conservation efforts. Arboreal primates are difficult to monitor as their habitat is usually difficult to access and most primates have some degree of camouflage, making them hard to observe. Surveys conducted using unmanned aerial vehicles (UAVs) equipped with thermal infrared (TIR) cameras overcome these limitations by flying above the canopy and using the contrast between the warm body temperatures of animals and cooler background vegetation, limiting issues with impassible terrain and animal camouflage. I evaluated technical and procedural elements associated with conducting UAV-TIR surveys of arboreal monkeys. Initial tests were conducted to determine optimal flight parameters for detecting arboreal monkeys. Two imaging missions were then conducted and used for further analysis. One dataset was collected at Alvarado Creek, San Diego California, USA where human targets and heat packs were placed within a riparian environment. The second dataset was collected at Abenteuer Affenberg in Villach, Austria which is a forested primate research center housing 166 Japanese Macaques. This study shows that target detectability for both datasets depends on at least 4-7 °C difference between target and background temperatures, and selecting flight altitude and TIR camera focal length combinations that yield ground sampling distances no greater than half the size of the targets. Repeat Station Imaging (RSI) procedures using co-registered TIR image pairs facilitated the use of image differencing to detect targets that moved between repeated imaging passes. Through Structure from Motion (SfM), a point cloud showing area elevation data was generated from TIR imagery but lacked sufficient point density to determine target locations. However, x-y-z positions were effectively estimated with a 3D polygon mesh derived from the point clouds. Limited data collected on target location/elevation at the time of the surveys prevented a more robust assessment of target identification accuracy. Future research should include simultaneous, ground-level observations to increase confidence in the accuracy of these methods for arboreal animal surveys.

en_US

Geography (Geographic Information Science)

Geography

Arts and Letters

San Diego State University

Master of Science (M.S.) San Diego State University, 2020

http://hdl.handle.net/20.500.11929/sdsu:59969