3D MODELING OF SAND DUNE SURFACES USING CLOSE RANGE DIGITAL PHOTOGRAMMETRY: A PROGRESS REPORT

We have been using close range digital photogrammetry to record the three-dimensional configuration of sand dune surfaces at scales from 5 to 100 m. The unique characteristics of dune environments emphasize both strengths and weakness of digital photogrammetry as a method for gathering 3D positional data. Strengths include that (1) points are measured remotely, in contrast to GPS and some laser survey methods, thereby leaving the highly mobile surface undisturbed; (2) data gathering is more rapid than in laser survey methods; and (3) compared to laser scanning, photogrammetry requires less and lighter equipment that is easier to hike into remote dune locations. Limitations include that (1) at our scale of observation sand dunes can be visually homogenous, providing fewer uniquely identifiable homologous points needed to render a 3D surface; (2) photogrammetry is a line-of-sight method, so topographic irregularities or foreground vegetation can obscure background features; (3) photogrammetry works best with views normal to the surface being modeled, but dune topography may limit views to relatively low angles; and (4) the mobile nature of the landscape requires extra effort to establish stationary control points for time series comparisons. Of the limitations, the homogeneity of dune surfaces when viewed at a distance and the effects of low angles of view appear to most seriously impede accurate construction of 3D dune models. A future partial solution to the first of these may be to place targets to provide additional homologous points to improve the density of surface modeling, although at the cost of disturbing the surface during target placement and retrieval. Pole aerial photography may provide a more complete remedy to the problems of foreground obstructions and low angles of view.

The current limitations on dune photogrammetry make it difficult to achieve high enough resolution to investigate questions requiring high 3D accuracy such as measuring the effects of single storm events over large areas, but the method can still be useful to measure sand movement within smaller, less homogenous fields of view, to measure larger scale dune migration over longer time periods, and to monitor changes in dune vegetation.