Resumen de Extracción automática de mdt a partir de datos procedentes de láser escáner (Espada)
Emilio Rodríguez Caballero, Sonia Chamizo de la Piedra, Yolanda Cantón Castilla, A. Afana, A. Solé i Benet
Terrestrial laser scanners (TLS) enable the acquisition of accurate tridimensional point clouds that allow the generation of high resolution digital terrain models (DTM) , permitting a wide range of applications in geomorphologic studies. These studies focus on the analysis of soil surfaces, where vegetation filtering is quite important. Filtering is a laborious and time-consuming process, cause for which numerous automatic methods have been developed to remove vegetation. Most methods have been developed to filter LiDAR data, which vary widely from TLS data. Application of LiDAR filtering to TLS data produces smoothed surfaces that do not retain their real morphology. We present a new methodology to filter TLS data, based on the morphology and spectral response of surfaces, which acts at different scales depending on the vegetation type. This methodology reduced the error generated by processed filtering compared with previous methodologies. This reduction is more important in non-vegetated areas, where other methodologies underestimate surface heights
