Surface Displacement Measurement from Remote Sensing Images

Ewelina Rupnik

Abstract: Image matching is a well-known problem in the Earth sciences that features in two main metrology applications: the computation of 3D scene geometry using images acquired from different viewpoints, and the computation of surface deformations using images acquired from similar viewpoints. Images taken at different times and/or from different viewpoints are subject to radiometric and geometric deformations. To exploit such images for metrology, the deformations must be adequately modelled. This chapter will address the most common approaches of deformation modelling, including normalized template matching, imposing an "a priori" on deformations (i.e., regularization), as well as multi-resolution techniques. Finally, because image matching relies on the known geometry of the imaging sensor, we also briefly introduce the most common types of sensor geometries, their mathematical representations and the frequently met sensor's geolocation errors.