Solid Earth Applications

Courtesy of NASA/JPL

research group coordinator Prof. Giorgio Franceschetti

The research objective is twofold

For the first time, possibilities for Synthetic Aperture Radar (SAR) to survey data useful for management and planning of urban areas are to be explored.
	The morphostructure of some areas where measurements of on-ground facts are available will be analyzed.

Urban areas

How a Synthetic Aperture Radar works is familiar to specialists in the sector. For those coming to this remote-sensing sensor for the first time, SAR's operation can be schematically explained as follows:

the SAR is a radar that transmits an electromagnetic signal; this signal propagates from the satellite carrying the sensor to the area of Earth under observation; the signal, like a wave of the sea hitting a rock, is reflected from the Earth area observed, and part of the reflected signal goes right back to the sensor that transmitted it.

The idea suggests itself that the signal that comes back to the sensor should retain traces, that is, have characteristics that depend on the configuration of the Earth area observed. Accordingly, by receiving the return signal at the sensor and appropriately processing it one can in fact manage to obtain images of the scene observed that can be seen as microwave "photographs".

Some websites in the world are rich in these microwave photographs (NASA, ESA) that show how this sort of sensor can manage to supply images that show aspects that complement optical imagery, but with much more interesting resolution (in this connection see the optical images from  Ikonos sensor).

In these images it is easy to recognize seas, lakes, rivers, roads. It is somewhat harder to manage to recognize mountains with their shapes; built-up areas, which appear very bright, can also be recognized, but it is very hard to identify individual buildings..

The previous diagram helps us to understand why: in the case of buildings, the signal is reflected twice. The first, for instance, is from the side walls of each building, which are smooth and reflect the signal, as a mirror does with a light ray, towards the ground; the second reflection comes from the ground where, if the paving is smooth enough, almost all the signal is directed back to the sensor. The twofold reflection thus has the power to concentrate the signal almost entirely back in the direction it was transmitted from, which explains why urban areas appear very bright in SAR images. At the same time the signal received does retain traces, unfortunately overlapping, of the building and the surrounding paving. Just as with a photo frame exposed twice with two different subjects, it is hard to separate these two contributions and reconstruct even one of the two original images.

The research the group proposes will seek specifically to study these reflection effects, understand the origin of the overlapping contributions and identify techniques enabling more comprehensible microwave images of urban areas to be made.

The group is organized as follows: the universities of Naples and Florence swill study the mechanisms of interaction between the SAR signals and urban areas, create theoretical models and carry out simulations of the type of city image to be expected from X-SAR/SRTM; the university of Pavia will use the models proposed and seek to extract information on built zones by working on the data actually gathered by X-SAR/SRTM..

For those wanting further details

The research in the area of remote surveying will be done by the universities of Naples, Florence and Pavia. A model of the electromagnetic signal diffracted from buildings will be developed and inserted into a simulator of the RAW signal.

The results of the simulation will be used to define the information that can be extracted from SAR images and from interferograms of urban areas. The identification of constructions in urban areas will be checked on, and the quality of DEMs in built-up areas assessed.

The university of Pavia will have available direct-scattering models it will use to train neural networks specialized for inverting the diffracted field. Automatic and semi-automatic tools will be designed to invert interferometric measurements on urban areas. This will be used to validate X-SAR/SRTM data from densely built environments, taking into account the materials, shape and density of buildings, and to derive from the data themselves the greatest possible amount of information on complex, crowded environments like cities.

To this end, among tools used will be neural networks specialized for inversion of the diffracted field, and instruments will be developed to interpret the images, making use of correlation of the radar measurements to identify such structures as buildings and streets. Data from the Pavia group's previous experience in this area will be taken as a basis. The test area should be the city of Pavia itself, for which three-dimensional cartography is already available.

 Morphology

What morphological information can be derived from the X-SAR/SRTM data?
Three geological areas of particular importance will be analyzed

Northern Ethiopia, one of the areas of worst territorial degradation due to natural factors and human settlement; The area running from Scutari (Albania) to the Eastern Peloponnese (Greece), for the search for oil; Vesuvius, an area at high volcanic risk

In each of these areas geological data will be surveyed in association with the passage of X-SAR/SRTM to verify which are recognizable on the SAR images.