EXTRACTION OF TRUE DEPRESSIONS IN DIGITAL ELEVATION MODEL (DEM) DERIVED FROM LIDAR

Digital elevation models (DEMs) typically have a combination of artifacts and actual topographic depressions and many of these local minima (depressions/pits) are the result of errors in the DEM production process as opposed to a few which are real. The pits or depressions are usually considered as spurious features in low-resolution DEMs that hinder flow routing in landscape evolution and hydrological modeling studies and are removed by filling them up to the lowest elevation of surrounding pixels in order to have a “depressionless” DEM. This process is referred to as pit filling. The occurrence of digital depressions can only be confirmed as existing features in the landscape through ground inspection. Even depression validation through field checking by itself is an imperfect process because of the varying accuracy level associated with the ability to identify depressions in the field.

Recent advancement in the field of digital photogrammetry and laser altimetry has shown increasing capabilities of representing actual depressions in the landscape because of high spatial resolution. This study attempts to locate natural depressions in a DEM derived from LIDAR data of Nixa sinkhole plain, Missouri having a spatial resolution of 1 meter. The objectives of this study are to -1) extract true depressions (a subset of which are sinkholes) from DEM alone , 2) compare them with the depressions that are evident from the aerial photo interpretation and ground checking, and 3) find out the level of accuracy of the method. The result from this study will show whether high resolution DEM can be used independently to extract the near perfect terrain features or not.