| Paper: | TU-P1.1 |
| Session: | Advanced Radiometry |
| Time: | Tuesday, March 27, 13:20 - 13:40 |
| Presentation: |
Oral
|
| Topic: |
Advanced radiometer techniques: |
| Title: |
LOW SIDE LOBE LEVEL IMAGE RECONSTRUCTION IN MICROWAVE INTERFEROMETRY |
| Authors: |
Manuel Martin-Neira; European Space Agency | | |
| | Ana Olea; Airbus Defence and Space - Madrid | | |
| | Andres Solana; Airbus Defence and Space - Madrid | | |
| | Josep Closa; Airbus Defence and Space - Madrid | | |
| | Javier Herreros; Airbus Defence and Space - Madrid | | |
| | Marta Garcia; TRYO Aerospace - Madrid | | |
| | Alejandro Garcia; TRYO Aerospace - Madrid | | |
| | Ignasi Corbella; Polytechnic University of Catalonia - Barcelona | | |
| | Francesc Torres; Polytechnic University of Catalonia - Barcelona | | |
| | Nuria Duffo; Polytechnic University of Catalonia - Barcelona | | |
| | Raúl Díez-García; European Space Agency | | |
| Abstract: |
The over-8-year data record available from the SMOS mission has allowed, through intensive analysis, the better understanding of the origin of the spatial ripple present in the brightness temperature images produced by its payload, MIRAS, and in general, by any two-dimensional microwave aperture synthesis radiometer. The source of the spatial ripple can be attributed to the combined effect of dissimilarities between antenna patterns, element spacing when it does not respect the Nyquist alias-free criterion for 2 dimensional sampling, and array geometry, closed arrays leading to lower ripple.
Spatial ripple may still be present even when having perfect knowledge of the instrument and in the absence of any calibration errors depending on similarity, spacing and array geometry, and for this reason it is sometimes referred to as ‘noise floor’. The noise floor is moreover dependent on the scene content, in particular in the alias regions, and presents a real limitation in the quality of the images an aperture synthesis radiometer can deliver. Therefore the demonstration, by measurements performed on real hardware, that an interferometer could be built with significantly lower spatial ripple, is of importance for future radiometer missions using aperture synthesis.
The reduction of spatial ripple can be achieved by narrowing the element spacing, making the element patterns more similar or having a closed array shape, or a combination of these. ESA initiated an activity to test the performance improvement when element spacing is reduced and their patterns made more similar, leaving the shape of the array aside for a separate assessment.
Shortening the element spacing has a strong impact in instrument complexity (the number of receiver and correlator increases) and performance (sensitivity degrades as the collecting area reduces). Moreover there is always a physical limit to the spacing reduction. For example, the LICEF antennas of MIRAS which are spaced by 0.875λ can be brought together until they touch each other, reducing the spacing down to a minimum of 0.767λ. This is the spacing under consideration for a potential SMOS follow-on mission before having to redesign the current antenna.
In what concerns pattern similarity, theoretically, in order to make the element patterns identical, the boundary conditions around every antenna element must be the same. Assuming an hexagonal geometry of MIRAS that condition leads to the need of surrounding every antenna element by other identical ‘dummy’ elements arranged according the underlying hexagonal grid. A ‘dummy’ element is defined as an additional element identical to an active one except that it is terminated with match loads (instead of receiving chains) and used for the purpose of making the antenna patterns of active elements more similar to one another. From the experience gained during the antenna ground characterization of MIRAS, the radius of the scattering area of one particular element spans up to 3 times the element spacing, this which sets the amount of dummy antennas that is needed to surround every active element.
In conclusion, by narrowing the antenna spacing and ensuring any active elements is surrounded by 3 other elements in any direction of the hexagonal grid, the noise floor should be substantially reduced. The results of the experimental activity ESA is carrying out on this subject are the subject of this paper.
|
