Generate Digital Elevation Models Using Laser Altimetry (LIDAR) Data Literature Survey

Laser altimetry (LIDAR) data must be processed to generate a digital elevation model (DEM), a digital map of the terrain surface. The raw LIDAR points can be processed and then gridded to form a map of the ground surface. For example, a waveform generated by binning a set of points over a limited region can be decomposed into a set of Gaussian components. The component with the lowest mean represents the ground elevation. The data contributing to the lowest component is then gridded to form the final DEM. The LIDAR data can be gridded to form different digital images and then the images are processed to create a map of the ground surface. A common and simple technique uses varying window sizes and thresholds at each pixel of the image. This can produce good results when properly tweaked but with very little statistical justification. Anisotropic Diffusion Pyramids' (ADPs) ability to preserve edge information while smoothing homogenous regions can be useful for segmenting buildings or areas of vegetation which need to be removed.