| Description |
The use of unmanned aerial vehicles, or UAVs, has grown widely as they have grown less costly and more available to the public. One area of growth is in the remote sensing of snow and ice. In particular, structure from motion photogrammetry is used to determine snow depths through differential mapping of surface elevations, with one flight being done without snow and one with. The difference is then calculated from the two. In this paper I will present a snow depth analysis with SfM differential mapping to investigate how these snow depth differences can change with a change in scale. The flight was conducted at Alta, Utah, in May of 2017, in addition to a snow free flight. Using the Agisoft Metashape Professional and ArcGIS Pro software snow depth images at different resolutions were mapped, then differences were assessed using histograms of the distributions of values. Based on the snow depth maps and related histograms, it can be seen that, depending on the coarseness, there can be subtle and very obvious changes in detail when measuring the difference in elevation between the snow covered images and the snow free image. The 0.1 m resolution was the least detailed and therefore appeared to have the lowest accuracy compared with the other resolutions. in measuring the snow depth. The 0.05 meter resolution was more detailed and had a greater standard deviation, meaning there was more variation in the values mapped at this resolution. The 0.01 meter resolution was the closest to the default resolution output by Agisoft metashape, with a highly detailed orthomosaic and histogram. Each of these has positives and negatives when taken into consideration, but the 0.01 is the best suited for measuring something as detailed as snow depth with structure from motion. With these results, it is important to understand how changes in scale affect the values of the differences as snow depth mapping with drones becomes more common. |
