A functional analytic method to derive displacement vector fields from satellite image sequences

Over the past 20 years the maximum cross-correlation method has been employed both operationally and for research in geophysical sciences. Some weaknesses of the well-understood and established technique have remained, which suggest that a non statistical approach is required to overcome these deficiencies. Presently a new functional analytic approach is adopted from computer vision to derive displacement vector fields from satellite image sequences. Case studies of artificially created images with defined motion patterns as well as displacement vector fields derived from Meteosat water vapour pictures and NOAA-AVHRR sea surface temperature pictures reveal the applicability of the method. Due to the basic constraints, the method is particularly suitable for fuzzy greyvalue structures. Nevertheless, modifications concerning the atmospheric and oceanographic kinematics are still required to allow also strong shear flow to be accounted for.