Improving the continuous photogrammetric monitoring system
Keywords: Monitoring, Digital Photogrammetry, Vegetation masks, Precursor movements, Rockfalls detection, Cross correlation
Abstract. The basalt wall next to the Fluvià river, where Castellfollit de la Roca (Girona) is located, is known for its high frequency of rockfalls, which can affect existing buildings. To monitor it, a continuous photogrammetry system developed by the UPC's geomatics engineering group and the ICGC was installed in 2021. This system, consisting of three cameras, has since captured images that are sent to the server located at the UPC, where they are processed to detect movements in the rock massif over time. If a rockfall is detected, a three-dimensional model is generated to estimate the volume of rock mobilised.
This massif is characterised by the presence of changing vegetation and a high frequency of rainy and foggy days. These changes in the appearance of the rockfall decrease the sensitivity of the photographic system to the detection of premonitory movements or low-volume rockfalls. To mitigate these effects, improvements have been made to the matching algorithms. These improvements include the implementation of different procedures such as the automatic selection of the best quality daily image, radiance balancing techniques and the creation of dynamic masks adapted to vegetation changes.