Use of a Light Uav and Photogrammetric Techniques to Study the Evolution of a Landslide in Jaén (Southern Spain)

Abstract

ABSTRACT:This paper presents a methodology for slope instability monitoring using photogrammetric techniques with very high resolution images from an unmanned aerial vehicle (UAV). An unstable area located in La Guardia (Jaen, Southern Spain), where an active mud flow has been identified, was surveyed between 2012 and 2014 by means of four UAV flights. These surveys were also compared with those data from a previous conventional aerial photogrammetric and LiDAR survey. The UAV was an octocopter equipped with GPS, inertial units and a mirrorless interchangeable-lens camera. The flight height was 90 m, which allowed covering an area of about 250 x 100 m with a ground pixel size of 2.5 cm. The orientation of the UAV flights were carried out by means of ground control points measured with GPS, but the previous aerial photogrammetric/LiDAR flight was oriented by means of direct georeferencing with in flight positioning and inertial data, although some common ground control points were used to adjust all flights in the same reference system. The DSMs of all surveys were obtained by automatic image correlation and then the differential models were calculated, allowing estimate changes in the surface. At the same time, orthophotos were obtained so horizontal and vertical displacements between relevant points were registered. Significant displacements were observed between some campaigns (some centimeters on the vertical and meters on the horizontal). Finally, we have analyzed the relation of displacements to rainfalls in recent years in the area, finding a significant temporal correlation between the two variables.

Analysis of Landslides Based on Displacements of Lines

“…This process includes the planning of the flight (repeated in both epochs), the orientation of the photographs, the obtaining of the Digital Surface Model (DSM) and subsequently, the obtaining of an orthoimage, according the methodology followed in previous works (Fernández et al, 2015). The DSMs and the orthoimages of both epochs are the final products derived from the photogrammetric process used in this study.…”

“…In this zone there are several active landslides (Fernández et al, 2015). In this case we analysed two active sections (Figure 3) or landslides which correspond to mud flows, according to the classification of Varnes (1978).…”

“…UAS represent a very promising tool for landslide recognition (Scaioni et al, 2014). In this context there are several studies that have used photogrammetric products obtained from UAS flights in order to study and analyse landslides, as in natural slopes (Niethammer et al, 2012;Turner et al, 2015;Fernández et al, 2015) such as road embankments (Carvajal et al, 2011). These studies have mainly analysed DSMs or discrete points extracted from landslides at a determined moment or comparing them between several epochs.…”

“…A further step is given by photogrammetric and remote sensing techniques (Metternich et al, 2005;Fernández et al, 2009;Scaioni et al, 2014) and LiDAR (Jaboyedoff et al, 2012). From traditional aerial photogrammetry (Brunsden and Chandler, 1996;Walstra et al, 2004;Prokešová, et al, 2010) to the current development of Unmanned Aircraft Systems (UAS) (Stumpf et al, 2012;Niethammer et al, 2012;Turner et al, 2015;Fernández et al, 2015), the use of photogrammetric flights, many times combined with LiDAR data (Dewitte et al, 2008;Corsini et al, 2009;Fernández et al, 2012), for modelling landslides has suffered an enormous increment because of the high geometric accuracy achieved, the semantic information of images, the easy accessibility in the field and the reduction of costs in the case of UAS. Furthermore, the photogrammetric techniques allow a DSM of the terrain to be obtained and also orthoimages.…”

“…These studies have mainly analysed DSMs or discrete points extracted from landslides at a determined moment or comparing them between several epochs. The first and simplest approach used by several authors has been the comparison of DSMs corresponding to several epochs (Stumpf et al, 2012;Niethammer et al, 2012;Turner et al, 2015;Fernández et al, 2015) in order to analyse the evolution of a landslide. In other cases, DSM have been used to identify point features automatically; in this sense, a common issue of these studies is the use of points (extracted from DSMs or orthoimages), in a manual (Fernandez et al, 2015) or automatic way (Lucieer et al, 2014;Turner et al, 2015) for modelling the displacements.…”