Digital Elevation Models (DEMs) are raster datasets representing the bare-earth terrain surface. Bridges and culverts, being structures above or within the terrain, would ideally be removed from a DEM to accurately reflect the underlying topography. However, the presence of these structures within DEM data often persists due to limitations in data processing techniques and source data resolution. For example, if a bridge spans a significant distance but the DEM’s resolution is coarse, the bridge’s representation may blend with the surrounding terrain during processing, making its removal difficult without introducing artificial voids or inaccuracies.
Retaining bridges and culverts in DEMs can be beneficial in specific contexts. For hydraulic modeling, for example, accurate representation of water flow requires accounting for these structures, as they influence water conveyance. Furthermore, in some applications, maintaining a complete and unmodified representation of the original data is crucial for historical record-keeping or change detection analyses. Removing bridges and culverts might inadvertently erase valuable information about the built environment over time. Historically, processing power and automated algorithms were less sophisticated, contributing to the challenge of reliably extracting these features from DEMs.