A comparative Analysis of Grain Size distributions and Cross Sections using an Oblique Photoset on the Nisqually Riverbed Adjacent to Longmire In Mount Rainier National Park
MetadataShow full item record
Collecting cobble counts and gathering cross sections out in the field can be a tedious and time consuming process. An attempt was made at expediting this process using structure from motion technology to create an orthophoto and a digital surface model of the Nisqually riverbed adjacent to Longmire in Mount Rainier National Park. An oblique photoset was gathered of the Nisqually riverbed using a telescoping pole with a digital camera and high-precision GPS mounted at the end of it. This photoset was then used to create a point cloud, an orthophoto, and a digital surface model using Pix4D. Automated cobble counts were gathered using two different Matlab scripts; DigitalGrainSize, and BASEGRAIN. DigitalGrainSize proved to be fairly accurate and may act as a replacement if grain sizes 11 mm and below are not relevant to a study. An automated grain size distribution may be even more accurate if a higher resolution digital surface model is produced or if a single photo is used instead of an orthophoto. BASEGRAIN did not perform as well and did not detect both smaller and larger grain sizes. Cross sections were derived from the digital surface model and have a high resolution when compared to 1 m resolution lidar in the same area. Channels that are only active at higher flows can be seen clearly in the digital surface model cross sections as well. The only drawbacks are that vegetation, and water are included in the digital surface model, so it cannot measure beneath the water’s surface as opposed to a total station, and the elevation was approximately 60 feet lower than actual elevation. This was likely due to a GPS error. I believe that these two applications show promise, especially if these techniques are refined.