Mapping Drainage Structures Using Airborne Laser Scanning by Incorporating Road Centerline Information

Wide-area drainage structure (DS) mapping is of great concern, as many DSs are reaching the end their design life and information on location usually absent. Recently, airborne laser scanning (ALS) has been proven useful for DS through manual methods using ALS-derived digital elevation models (DEMs) hillshade images. However, slow labor-intensive. To overcome these limitations, this paper proposes an automated algorithm (DSMA) classified ALS point clouds road centerline information. The DSMA begins with removing ground points within buffer centerlines; size varies according to different classes. An ALS-modified DEM (ALS-mDEM) then generated from remaining points. A network (DN) derived ALS-mDEM. Candidate obtained by intersecting DN centerlines. Finally, a refinement 15 m placed around each candidate prevent duplicate being in close proximity. total area 50 km2, including urban site rural site, Vermont, USA, was used assess DSMA. Based functional classification scheme Federal Highway Administration (FHWA), regarding FHWA roads public data portal. non-FHWA roads, i.e., private streets, impervious surface land cover dataset. benchmark dataset gathered transport agency Vermont further augmented Google Earth Street View images authors. one-to-one correspondence between mapped two sites established. positional accuracy assessed computing Euclidian distance DS. mean were 13.5 15.8 m, respectively. F1-scores calculated prediction accuracy. For 0.87 0.94 0.72 0.74