2D Horizontal Landslide Displacement Estimation by Multi-temporal Image Correlation Techniques

Geomatic techniques based on digital images have become a very useful tool for landslide monitoring, quantifying landslide movements over a long period of time. One of these techniques is based on digital image correlation. Digital image correlation estimates 2D displacements by means of statistical techniques, automatically matching identical points in the two digital images. It may be applied using remote sensing or aerial images. Digital image correlation has been applied to multi-temporal aerial photographs at the Cerro Gordo landslide (Granada, Spain) which is badly affecting an urbanized area. This landslide consists of a main deep slide and several minor shallow slides, located around the head scarp or affecting some parts of the main landslide body. Slides involve existing landfills used for urban works as well as in situ materials (colluvial deposits and weathered schists). The period under study (2002–2010) has been divided into four subperiods of approximately two years each. Results provide horizontal displacement in the landslide area and surroundings which allow to assess the landslide movement evolution at different stages according to triggering factors and structural measures that were carried out to stabilize the affected area. Over the period of 2002–2004, displacement rates varied from 2 to 6 cm, meanwhile for the periods of 2004–2006 and 2006–2008 displacements continued with 4–6 cm rates. Then, the period of 2008–2010 generally shows a higher activity with more areas showing 6 cm displacement rates. Surface observations of these periods go along with both location and direction of displacements as shown by the results of digital image correlation techniques. Image correlation results have provided very useful information to understand the complex movements of the landslide as well as the effectiveness of the structural measures implemented.