Geo-referencing techniques of 3d models (SfM): Case study of mud volcano Village Santa Barbara, Caltanissetta (Sicily)

In the last 2 years we have performed a topographic and geodetic monitoring on an active mud volcano located in central Sicily (Caltanissetta, Terra Pelata locality). The target was to understand how the volcano is evolving morphologically and kinematically. Topographic and geodetic surveys were performed periodically (every 3 months) by using a suite of double-frequency GNSS receivers that where been set-up to monitor the kinematics and to geo-reference cartographic products (DTM, DSM and orthophotos). Digital maps were realized through photogrammetry technics and 3D surface modelling by adopting Structure for Motion (SfM) algorithm (James et al. 2017). Following the aim, morphological variations due to active tectonics and/ or ground movements which deform the area were achieved. The monitoring network has been deployed following a twofold target. The first one, is to observe the deformation and kinematics of the area surrounding the mud volcano. To do this, we fixed a stable benchmark by building a GNSS base and surveys were hence performed with a static D-GNSS. This configuration allows us to obtain a sub-centimetre precision in the point location. Other six points were also built and connected to the base. The second was to georeferencing, with high-precision, a series of ground control point (from now on GCPs) to positioning 3D surface model. In order to geo-reference the 3D models the receivers were configured as follow: –– RTK mode in the first four surveys; –– In the fifth, Static Ultra Rapid mode; –– In the sixth, we tested the precision of geo-referencing by comparing RTK data with Static Ultra Rapid. –– In the seventh, we coupled TST (Theodolite Total Station) and GNSS (RTK). All these techniques highlight the follow results: –– by using RTK mode derived coordinates, show errors of 0.05 m: –– by using Static Ultra Rapid mode, error in data fitting in in the order of one decimetre; –– by using TST coupled with GNSS, the error in data fitting are below 0.03 m. Models were georeferenced according to all applied techniques, while quality of method was tested during the monitoring period to improve model’s generation and their positioning. Our data show that the combination of TST with GNSS provides a greater precision in geo-referencing digital maps even if it requires a long acquisition time. The results, coming from the monitoring, allow us to quantify volumetric and altimetry changes of the mud volcano which are in the order of ± 1 dm during inflection or deflection phases. Our data suggest that the surface of mud volcano is experiencing a general trend of uplift in the order of one decimetre. The methods applied in this work to improve the positioning of objects in remote sensing techniques such as rock clusters, landslides, active structures, coastal morphology (Fonstad et al., 2013). Fonstad, M.A., Dietrich, J.T., Courville, B.C., Jensen, J.L. & Carbonneau, P.E. (2013): Topographic structure from motion: a new development in photogrammetric measurement. Earth Surface Processes and Landforms, 38(4), 421-430.