A Crowd-Sensing System for Geomatics Applications

Risk prevention is recognized as one of the most critical aspects of the policies of environmental monitoring. Because of the limited resources and the large amount of structures used for erosion control and slope protection, the Civil Protection and the Italian Forestry Carabinieri are not able to supervise them directly, with enough frequency. The present work is aimed to develop an innovative technique for periodically monitoring those structures, combining Mobile Crowd-Sensing (MCS) technology with photogrammetry and GIS. The experiments were performed in the Nature Reserve of Tirone (a protected natural area located inside the Vesuvius National Park in Naples) by analysing the metric reconstruction of two structures (a small weir and a log crib wall), before and after an accident, artificially generated for simulating a hydrogeological event or an act of vandalism, in order to evaluate GCPs influence. The procedure was split into four main phases: periodic acquisition of sets of photos with common smartphones and their transmission via the Internet; elaboration of the threedimensional model starting from a subset of selected pictures; comparison between the generated and the previous model; database update and programming of the subsequent monitoring. The accuracy of photogrammetric reconstructions was evaluated comparing the reconstruction with and without Ground Control Points (GCPs). The results show the models extracted without GCPs are satisfactory, since they allow to retrieve dimensional information of the examined constructions and to detect any instability. Models, generated using GCPs, are more detailed, but the processing and operational time is strongly higher.