The DEM represents topography as measured by the heterogeneously distributed 3D laser returns as a two dimensional grid of elevation values spaced at a constant resolution. Computing the DEM from the scattered data is a
computationally intensive, but relatively straightforward task. Many
algorithms are available for these computations (e.g., El-Sheimy et al.,
2005). Where the resolution is greater than the typical shot spacing,
the DEM preparation method typically used is local binning (e.g.,
El-Sheimy et al., 2005; Kim et al., 2006) in which the value at a DEM
node is a simple function (e.g., maximum, minimum, mean, and
inverse distance weighted—IDW) of the points within a specified
search radius (r). Increasing the search radius ensures that points
will be found (otherwise a null is assigned at the grid node), but
the value will represent a broader area