First Experiment of Long-Range Panoramic Images on a High-Precision Geodetic Reference Frame

Abstract

[EN] Geomonitoring of rockfalls and landslides is increasingly carried out by solutions that integrate different geomatics techniques to provide quickly 3D point clouds or models that are required to be rigorously in the same reference system. Methods based on remote sensing such as terrestrial laser scanning or photogrammetry need precise ground control, which is usually provided by means of geodetic surveys. However, when the study area is large with strong limitations due to particular orography, those geodetic techniques cannot always grant accurate target points optimally distributed within the monitored object, and only an external reference frame is available to provide absolute orientation to those 3D point clouds or models. In that case, terrestrial photogrammetry shows clear advantages compared to terrestrial laser scanning. Still, it may require a large number of high-quality images taken from well-distributed stations, thus hampering the desired fast data collection. A possible alternative to cope with this problem is the use of the panoramic photogrammetry method by using robotic devices like Gigapan along with a systematic collection procedure from stable stations of a reference frame whose coordinates are accurate and well-controlled. This contribution describes an experiment conducted in Cortes de Pallás (Spain), where an existing infrastructure consisting of 10 pillars and 15 check points is annually monitored at millimetric level, to investigate the potential of long-range panoramic photogrammetry as applied to deformation monitoring. The panoramic images were taken from 7 selected pillars using the Gigapan robotic device. The accuracy of oriented panoramic images, the performance of the method in long-ranges (500-1000 m), and the consistency with the geodetic techniques in the 15 check points were analyzed. Finally, some conclusions about the suitability of panoramic photogrammetry as applied to long-range deformation monitoring are drawn.