Anna Roser, Peter Olsoy, and Trevor Caughlin
Accurate and comprehensive monitoring, where information can be collected across multiple scales and be spatially referenced on a continual basis, is needed to create better models for sagebrush restoration efforts. By using remote sensing techniques like unoccupied aerial vehicles (UAV), researchers can collect in an afternoon flight equivalent data to field-based methods that would take days to accumulate. In this study, we used Agisoft Metashape software to process UAV imagery taken at the Soda common garden in 2019 and again in 2020. Our objective was to test the impacts of several image processing parameters on final products including point clouds. We found that changes to the parameters in Agisoft Metashape did not produce any large differences in point cloud products. However, we did find a noticeable difference in the quality of images from flights in June 2019 and September 2020. Because the images were taken at different times of year, we found the software had difficulty detecting the sagebrush in the 2020 images due to the lack of leaves, and the longer shadows cast in the fall, which resulted in a lower quality point cloud. Based on these results, our next steps will focus on testing other parameters to improve the final products generated from UAS flights in both spring and fall seasons.
Marie, Valorie; Clifton, Stacey; Roser, Anna; Olsoy, Peter; Zaiats, Andrii; and Caughlin, Trevor, "Comparison of Image Processing Methods for Better Point Clouds of Sagebrush" (2021). 2021 Undergraduate Research Showcase. 27.