Authors: Hunter Gordon, Jace Riley, McCade Larsen
Mentors: Aaron Davis
Insitution: Utah Tech University
Tracking animal populations is key to ensuring that populations are healthy and thriving. Current data collection methods, such as radio tagging, manned aerial flyovers, and camera traps, are not only time-consuming and expensive but also fail to provide accurate population estimates. This interdisciplinary research project aims to produce a more accurate and less expensive data collection system for large game populations that could potentially be used to monitor other animal populations. The planned procedure for data collection is to attach a thermal imaging device to an unmanned aerial vehicle (UAV) and perform aerial transects throughout the observation area. Using a thermal camera will allow the UAV to fly at a higher altitude, lessening the disturbance to the animals themselves. The imaging device will take images at a set time interval along transects that can be stitched into a complete data set for the area. By using image processing techniques and deep learning models, the images will be processed to show the location as well as population counts of the animals in the area. Results from similar experiments have shown that using UAVs to collect population data not only provides more accurate data but also requires less time overall to collect the data.[1] This experiment expands upon those findings by creating automatic image processing and analysis software to increase the ease of use, allowing for more data to be collected and analyzed in the same time period.