1 / 15

OIL SLICKS DETECTION USING GNSS-R

OIL SLICKS DETECTION USING GNSS-R. E. Valencia, A. Camps, H. Park , N. Rodríguez-Alvarez, X. Bosch-Lluis and I. Ramos-Perez Remote Sensing Lab, Dept. Teoria del Senyal i Comunicacions, Universitat Politècnica de Catalunya and IEEC-CRAE/UPC Tel. +34 93 405 46 64, E-08034 Barcelona, Spain.

keltons
Download Presentation

OIL SLICKS DETECTION USING GNSS-R

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. OIL SLICKS DETECTION USING GNSS-R E. Valencia, A. Camps, H. Park , N. Rodríguez-Alvarez, X. Bosch-Lluis and I. Ramos-Perez Remote Sensing Lab, Dept. Teoria del Senyal i Comunicacions, Universitat Politècnica de Catalunya and IEEC-CRAE/UPC Tel. +34 93 405 46 64, E-08034 Barcelona, Spain. E-mail: valencia@tsc.upc.edu

  2. Introduction GNSS-R imaging of the ocean surface Oil Slicks modelling Simulation results Conclusions 0. Outline

  3. 1. Introduction (1/3) • GNSS-R has stood as a powerful remote sensing technique. • In ocean applications, proposed to retrieve sea surface roughness (scatterometry). • Current approaches use the delay waveform or the DDM: • Fit a model tuned by the desired parameter. • Use an empirical direct relationship among the desired parameter and a GNSS-R observable (e.g. scatterometric delay or volume of the normalized DDM). • These approaches retrieve a single roughness descriptor for the whole glistening zone (averaged retrieval).

  4. 1. Introduction (3/3) • Glistening zone may present non-homogeneous sea state conditions. • DDM can be regarded as a blurred image of the scattering coefficient distribution in the Delay-Doppler (DD) domain. • Proposal: • Application: Oil slicks detection using GNSS-R. • Added value to other existing sensors (SAR, optical imagers, microwave radiometers, etc.) • Completeness of the spectral signature. • A method to retrieve σ0 from measured DDMs is presented and evaluated. • “Use the Delay-Doppler Map to retrieve the • scattering coefficient distribution over the observed surface”

  5. 2. GNSS-R imaging of the ocean surface (1/4) • A method for using GNSS-R measurements to retrieve the scattering coefficient distribution over the ocean has been recently proposed in: • E. Valencia, A. Camps, et al. • “Ocean surface’s scattering coefficient retrieval • by Delay-Doppler Map inversion,” • IEEE Geoscience and Remote Sensing Letters, July 2011. • This method is based on the inversion of the measured DDM expressed as the 2D convolution in the delay-Doppler (DD) domain of the WAF with a simple function (computationally inexpensive):

  6. 2. GNSS-R imaging of the ocean surface (2/4) • The first step to retrieve σ0 is to remove the WAF effects on the measured DDM by deconvolution (conceptually): • Inverse filter not suitable due to noise amplification effects. • Many deconvolution methods available in the image processing literature. • Deconvolution is the key step of the proposed method.

  7. 2. GNSS-R imaging of the ocean surface (3/4) • After deconvolution, the scattering coefficient distribution can be retrieved from Σ by simple operations using known terms. • However, 2 xy coordinates contribute to each DD point. • 2 ambiguous spatial regions. • Function Σ is a linear combination of both regions’ contributions. • Each region has an associated Jacobian function: J1 and J2. • Ambiguity has to be resolved, presented approach based on the measurement setup. • Other approaches are possible to resolve the ambiguity.

  8. 2. GNSS-R imaging of the ocean surface (4/4) • Spatial filtering of only one of the ambiguous zones is proposed by antenna beam pointing away from the specular point . • SAR-like setup. • A beam-forming system is needed • planned in future missions • (PARIS-IoD). • Using this spatial filtering scheme, • Σ is only contributed by a single space • region with an associated Jacobian. • At this point σ0 in the DD domain can be derived: • The final distribution over the xy surface domain is obtained by direct coordinates correspondence:

  9. 3. OilSlicksmodelling • Mean Squared Slopes modelled by the relationships with the wind speed derived by Cox and Munk, 1954. • After applying the empirical modification proposed by Katzberg et al. 2006, these MSS have been used to compute the corresponding σ0 distribution (KA-GO). • The DDM is computed by extending to the Doppler domain the Zavorotny and Voronovich, 2000 model for the waveform. Corresponding σ0 distribution of an ocean surface with an homogeneous 12 m/s wind speed and oil slick presence

  10. 4. Simulation results (1/4) • A first evaluation of the method has been performed by setting a LEO receiver simple scenario (without loss of generality). • Hardware effects not considered and the PARIS-IoD antenna pattern’s parameters have been used (D = 23 dBi, HPBW = 35º). Iso-range and iso-Doppler lines Antenna pattern projection on the observed surface Specular point: (x,y) = (0,0)

  11. 4. Simulation results (2/4) • With the defined scenario and scattering coefficient distribution, a reference DDMxy has been computed by the classical double integration over the surface xy domain. • As in Valencia et al. 2011, noise has been added with a SNR similar to that of available space measurements (UK-DMC).

  12. 4. Simulation results (3/4) • A CLS frequency-based method has been used for deconvolution: • P : Fourier transform of the smoothing criterion function. • Trade-off definition/noise by tuning the γ parameter. • After deconvolution, an estimation of the Σ (Δτ,ΔfD) is obtained. • The σ0distribution in the DD domain is derived by applying the appropriate compensation factors (Jacobian term, antenna pattern, propagation factors, etc.). • The last step of the method is to map from the DD to xy domain by direct coordinates correspondence (spatial ambiguity overcome by spatial filtering in the measurement).

  13. 4. Simulation results (4/4) • Scattering coefficient distribution retrieved with low error. • Oil slick is clearly detected. • Main error due to deconvolution artifacts. • Deconvolution process identified as the key step of the proposed method. Error map w.r.t. the input distribution σ0 retrieved in the xy domain

  14. 5. Conclusions • GNSS-R powerful remote sensing technique suitable for ocean scatterometry. • Ocean σ0 imaging from GNSS-R measurements proposed, with application to oil slicks detection. • DDM deconvolution to σ0 mapping within the glistening zone: • Antenna beam steering to eliminate the mapping ambiguity. • WAF deconvolution needed. • Compensation for the Jacobian and other known terms. • Deconvolution process identified as the key step of the method and further work is required. • Proposed method successfully evaluated with a simulation example for a LEO receiver, including noise. • The presence of an oil slick on an homogeneous 12 m/s wind speed ocean surface is clearly observable in the retrieved image.

  15. Thank you for your attention

More Related