UAV photogrammetric survey of the Poggio Baldi Landslide (Santa Sofia, North Apennine, Italy)

On 19th March 2010, a re-activation of an ancient landslide occurred in Poggio Baldi, a small village located near Corniolo, in the municipality of Santa Sofia (FC, Italy). Its previous catastrophic activation is dated back to March 1914. The 4 million m3 landslide is considered a real hazard, because during its last re-activation produced several serious damages on some private constructions and on the 310 Statal Road, but caused also a natural dam on the Bidente River. With a height-range between 810 – 480 m asl, the Poggio Baldi landslide involved the “Marnoso – Arenacea Romagnola” Formation, arenite-pelite alternation [Burdigaliansup – Langhian]. Started as a structurally controlled rockslide, it then evolved as a partially confined flow-like landslide affecting the material originally mobilized by the 1914 landslide. The landslide area is now made of a high vertical rock cliff up to 100 m high and about 250 m large, a rock debris on the order of 1.5*105 m2 at the toe of the cliff and an extensive mixed rock-earth debris mobilized by the 2010 landslide. Thanks to the joint effort of NHAZCA S.r.l., spin-off of the University of Rome “Sapienza”, the Department of Earth Sciences of the University of Rome “Sapienza” and Visiva4D, the experimental site of the Poggio Baldi landslide was surveyed on 12th April 2016 by a UAV flight. The employed UAV was a DJI Phantom 4 equipped with GPS and IMU apparatus. 12 MPx camera, with a 1/2,3’’ sensor size, equipped with a 20 mm lens, was used. Four flights were performed in order to generate accurate DSM of the entire site, by acquiring about 950 images. Due to the wide coverage, caused by the mountainous area, no satellite signals were available. For this reason all the flights were executed in manual mode. The orientation of the UAV flights were carried out by means of 12 ground control points, measured with a GeoMax Zenith 25 Pro GPS/GNSS Base and RTK Rover. The collected images was processed by suitable Structure from Motion (SfM) algorithms in order to achieve an accurate digital elevation model (DEM) of the whole landslide. The comparison of the acquired DEM with previous available DEM, achieved by Aerial and Terrestrial LiDAR surveys, allowed to evaluate the quality of the overall UAV DEM and, then, to compute the morphological changes occurred in the landslide area.