Earth Observation Using SAR and Social Media Images

Abstract

Earth Observation (EO) is mostly carried out through
\ncentralized optical and synthetic aperture radar (SAR)
\nmissions. Despite the controlled quality of their products,
\nsuch observation is restricted by the characteristics of the
\nsensor platform, e.g. the revisit time. Over the last decade,
\nthe rapid development of social media has accumulated
\nvast amount of online images. Despite their uncontrolled
\nquality, the sheer volume may contain useful information
\nthat can complement the EO missions, especially the SAR
\nmissions.
\nThis paper presents a preliminary work of fusing social
\nmedia and SAR images. They have distinct imaging
\ngeometries, which are nearly impossible to even coregister
\nwithout a precise 3-D model. We describe a general
\napproach to coregister them without using external 3-D
\nmodel. We demonstrate that, one can obtain a new kind of
\n3-D city model that includes the optical texture for better
\nscene understanding and the precise deformation retrieved
\nfrom SAR interferometry.