MODELING SAND DUNE SURFACES USING CLOSE-RANGE DIGITAL PHOTOGRAMMETRY

We use digital photogrammetry to record features of sand dunes at scales ranging from centimeters to tens of meters. Close-range digital photogrammetry is an efficient technique for recording the disposition of unstable landscape surfaces that would be subject to disturbance using GPS or other methods of direct measurement. Starting with two or more photographs of a dune surface, photogrammetric software can simultaneously generate xyz coordinates for many (>100,000) points throughout the scene. These point coordinates form vertices in a triangulated irregular network (TIN) that recreates the surface as a 3-dimensional computer model. A digital image draped over the TIN creates a photorealistic rendering of the surface, allowing measurement of features such as crossbeds.

To measure sand movement over time we need stable landmarks that are reliably locatable within the shifting dune landscape yet minimize chances for human disturbance. We have developed a system of two-piece marker pins placed within or around each site. The pins' bases are hollow metal tubes, buried even with the dune surface, with removable tops consisting of either a plastic ball attached to a laboratory stopper or a ball attached to a post that fits into the base. For each data gathering session we locate the bases with a metal detector, affix their respective tops, take digital photographs, recollect the tops, and rebury the bases. Pins work best in areas of lower disturbance within the dune system, where bases are less likely to be lost due to sand burial or moved during slumping of dune slopes.

Several factors can limit the photogrammetric modeling of sand dunes. Photogrammetry works best with lines of sight nearly normal to the modeled surface, but dune topography may restrict camera stations to more oblique views. Vegetation movement can distort distance measures and hinder point-matching between photographs. Also, whereas at close range sand presents many unique points that can be matched between photos, at greater distances dune surfaces may appear homogeneous, reducing the density of points that the software can recognize and use to create models. Despite these limitations, however, in many situations close-range digital photogrammetry is an effective means for studying dune features and dynamics.