15 Ship Surveillance with High Resolution TerraSAR-X Satellite in African Waters
309
the radial wave motions generally (Hasselmann et al. 1985). This leads, among other
effects, to image smearing and to a loss of information beyond the so-called azimuth
cut-off wavelength (Alpers and Brüning 1986). For ERS and ENVISAT SAR, this
corresponds typically to wavelengths shorter than about 200 m in the along track. As
motioned in introduction, the improved parameters of TS-X results in cut-off of about
100 m (azimuth) and 30 m (range) direction, the nonlinear effects are significantly
reduced.
First guess information, needed for algorithms to derive sea state from SAR image
spectra like PARSA (Schulz-Stellenfleth 2003), are not easily available. The use of
empirical functions is here needed. The empirical X-WAVE.2 model for obtaining
integrated wave parameters (wave height and period) has been developed for X-Band
data (Lehner et al. 2011, 2012) without needing a priori information. The algorithm
is based on analysis of image spectra and uses parameters fitted with collocated hind
cast wave model results of the German Weather Services (DWD). For the validation
and tuning of the X-WAVE algorithm, different data sets were used:
• Collocated buoy measurements,
• WaMoS II—Wave and Current Monitoring System (Borge et al. 2004, Dankert
et al. 2005),
• Radar altimeter data,
• DWD wave hind cast model resulTerraSAR (Bruck and Lehner 2010, Pontes et al.
2010).
Comparison of TS-X derived significant wave height with the significant wave height
obtained by the buoy located at Ekofisk oil platform shows a correlation of 0.83, for
National Data Buoy Center (NDBC) buoy 44066 the scatter index is 0.19, correlation is 0.95, mean square error 0.89. More details for parameter estimation and
comparisons are given in (Bruck et al. 2010, 2011; Lehner et al. 2012). Figure 15.23
shows an example of sea state parameter estimation carried out near Nigerian coast
from TanDEM-X Stripmap image acquired on February 19, 2012 with an incidence
angle of 23.9
◦ at centre coordinates.
15.6.3 Surface Current obtained from TerraSAR-X Dual Receive
Antenna Mode
A method to estimate the sea surface speed of the SAR cross-track velocity component
from TerraSAR-X image has been developed. The technique exploits the phase of
the Along Track Interferometry (ATI) acquisition. The ATI image pair is multiplied
in order to obtain the interferogram which is processed in order to obtain the speed.
The application of several filters, linear and non-linear is needed to get rid of the
SAR image noise and to deal with the surface scattering which moves. The measured
currents are derived by using TerraSAR-X Dual Receive Antenna (DRA) Stripmap
data. Due to the Pulse Repetition Frequency (PRF) a pre-filtering is needed to remove
the ambiguities which strongly affect these data. The demonstration on DRA data is
309
the radial wave motions generally (Hasselmann et al. 1985). This leads, among other
effects, to image smearing and to a loss of information beyond the so-called azimuth
cut-off wavelength (Alpers and Brüning 1986). For ERS and ENVISAT SAR, this
corresponds typically to wavelengths shorter than about 200 m in the along track. As
motioned in introduction, the improved parameters of TS-X results in cut-off of about
100 m (azimuth) and 30 m (range) direction, the nonlinear effects are significantly
reduced.
First guess information, needed for algorithms to derive sea state from SAR image
spectra like PARSA (Schulz-Stellenfleth 2003), are not easily available. The use of
empirical functions is here needed. The empirical X-WAVE.2 model for obtaining
integrated wave parameters (wave height and period) has been developed for X-Band
data (Lehner et al. 2011, 2012) without needing a priori information. The algorithm
is based on analysis of image spectra and uses parameters fitted with collocated hind
cast wave model results of the German Weather Services (DWD). For the validation
and tuning of the X-WAVE algorithm, different data sets were used:
• Collocated buoy measurements,
• WaMoS II—Wave and Current Monitoring System (Borge et al. 2004, Dankert
et al. 2005),
• Radar altimeter data,
• DWD wave hind cast model resulTerraSAR (Bruck and Lehner 2010, Pontes et al.
2010).
Comparison of TS-X derived significant wave height with the significant wave height
obtained by the buoy located at Ekofisk oil platform shows a correlation of 0.83, for
National Data Buoy Center (NDBC) buoy 44066 the scatter index is 0.19, correlation is 0.95, mean square error 0.89. More details for parameter estimation and
comparisons are given in (Bruck et al. 2010, 2011; Lehner et al. 2012). Figure 15.23
shows an example of sea state parameter estimation carried out near Nigerian coast
from TanDEM-X Stripmap image acquired on February 19, 2012 with an incidence
angle of 23.9
◦ at centre coordinates.
15.6.3 Surface Current obtained from TerraSAR-X Dual Receive
Antenna Mode
A method to estimate the sea surface speed of the SAR cross-track velocity component
from TerraSAR-X image has been developed. The technique exploits the phase of
the Along Track Interferometry (ATI) acquisition. The ATI image pair is multiplied
in order to obtain the interferogram which is processed in order to obtain the speed.
The application of several filters, linear and non-linear is needed to get rid of the
SAR image noise and to deal with the surface scattering which moves. The measured
currents are derived by using TerraSAR-X Dual Receive Antenna (DRA) Stripmap
data. Due to the Pulse Repetition Frequency (PRF) a pre-filtering is needed to remove
the ambiguities which strongly affect these data. The demonstration on DRA data is