1. State of the activity

The research activity at the Department of Information Engineering of the University of Pisa (Italy) has been mainly concerned with the definition of suitable processing procedures for estimating the two-dimensional sea wave spectrum from Synthetic Aperture Radar (SAR) images [1].

Classical methods proposed in the literature rely on the definition of a suitable analytical relationship between the sea-wave spectrum and the corresponding SAR image spectrum. This relationship accounts for the main effects contributing to the image formation process, as for instance velocity bunching, hydrodynamic phenomena, tilting of the local sea surface normal. In the most general case, the function relating the two-dimensional sea wave spectrum to SAR image spectrum is non-linear, rendering the inversion procedure a very complicated task. A solution to this latter problem can be obtained by resorting to an iterative technique, based on the minimization of a suitable functional. To this end, a sea wave first guess spectrum is assumed; then, this estimate is upgraded at each iteration till convergence is obtained. It is important to note that the convergence of the algorithm strongly depends on the first guess choice, which, in order to guarantee the effectiveness of the procedure, must be not too far from the actual sea wave spectrum. This guess is usually determined through hydrodynamic models or buoy measurements.

The availability of experimental data, contemporaneously recorded with in situ measurements, is of fundamental importance for testing and validating the aforementioned techniques. Moreover, this comparative analysis may also suggest optimization and extensions of these spectral estimation methods.

2. Significant results

At a first stage, these techniques are being applied to areas of the Mediterranean Sea. The attention is mainly devoted to the tuscan and ligurian coasts and in particular to La Spezia and Genoa Gulfs, where in situ measurements are available for our research group. These two gulfs represented a test-site during the last two SIR-C/X-SAR missions, but unfortunately no data relevant to this test-site were recorded during the missions. However, comparisons between experimental and numerical results have been performed using images recorded by the European satellite ERS-1.

A considerable part of the activity is devoted to the definition of efficient inversion algorithms making use of a numerical simulator previously implemented. This latter numerical code allows reconstruction of the SAR raw signal received by the spaceborne radar sensor in the presence of specific values of physical parameters as, for instance, local wind direction and intensity, and possible presence of swell. It can be used in the framework of an iterative inversion algorithm to provide a more accurate estimate of the SAR image spectrum generated from an assigned sea wave spectrum [2]. This evaluation is needed at each iteration step and is usually performed by an analytical transform. The introduction of this numerical tool may allow us to gain a deeper physical insight into several phenomena contributing to the SAR image formation process. It is important to note that this can be obtained at the expense of an increased computational cost which is essentially due to the time-varying character of the sea surface. This latter aspect makes numerical modelling very cumbersome. The simulation code implemented is capable of describing the case of a Real Aperture Radar (RAR) system as well. Consequently, to the end of validating the simulator, numerical tests have been carried out, comparing their outputs with experimental data presented in the literature [3], [4].

3. Future developments and data request

Concerning the next SIR-C/X-SAR mission, the acquisition of images of the sea surface by a two-antenna SAR system is of remarkable interest for the applications under consideration. Significant improvements in the above estimation procedures could be obtained with such a system [5]. By extending the numerical simulator implemented, the experiment could be simulated a priori to analyze the advantages of this operating configuration and to test the effectiveness of processing techniques. Later on, during the next mission, an in situ measurement campaign could be performed for validating both processing and numerical simulation. For these reasons, we are strongly interested, for the future developments of this research activity, in the acquisition of data relevant to the La Spezia and Genoa Gulf test sites during the next SIR-C/X-SAR mission in a two-antenna SAR configuration.

4. References

[1] Corsini, G., E. Ferretti, G. Manara, and G. Milillo, A. Monorchio, "Methods for directional sea wave spectrum retrieval from SAR images," to be published in AIT Italian Journal on Remote Sensing (in Italian).
[2] Corsini, G., G. Manara, and A. Monorchio, "Sea wave spectrum estimation from SAR images: a simulation based approach," 1995 International Geoscience and Remote Sensing Symposium (IGARSS '95) , Florence, Italy, July 10-14, 1995.
[3] Berizzi, F., G. Corsini, G. Manara, and A. Monorchio, "Simulation of electromagnetic backscattering data for sea clutter analysis," NATO AC243 Panel 3 RSG21 Conference on "Sea Surface Characteristics and Interaction with cm and mm Waves", Pisa, April 5-6, 1995.
[4] Corsini, G., G. Manara, and A. Monorchio, "Simulation of RAR reflectivity maps of the sea surface for remote sensing applications," 1996 International Geoscience and Remote Sensing Symposium (IGARSS '96) , Lincoln, Nebraska, May 27-31, 1996.
[5] Milman, A. S., A. O. Scheffler, and J. R. Bennett, "Ocean imaging with two-antenna radars," IEEE Trans. on Antennas Propagat. , Vol. AP-40, No. 6, June 1992.

Table of Contents

Converted to HTML by Alvin Wong, al.wong@jpl.nasa.gov

The Jet Propulsion Laboratory
4800 Oak Grove Drive
Pasadena, Cailfornia 91109