G. Di Martino

A Novel Approach for Disaster Monitoring: Fractal Models and Tools

In this paper, we present a complete framework to support the monitoring of natural and man-made disasters by means of synthetic aperture radar (SAR) images. The fractal geometry is the most appropriate mathematical instrument in describing the irregularity of a natural observed scene, by means of few effective and reliable parameters. Therefore, fractal concepts can be used to model and identify geometrical changes that occurred in areas hit by disasters. We present an overall framework employing fractal-based models, algorithms, and tools to support the identification of natural area changes due to natural or man-made disasters. Such a framework includes an algorithm used to extract fractal parameters from a 2-D signal, a fractal interpolation tool, and a SAR raw-signal simulator. The combined use of these tools provides an innovative instrument for disaster monitoring applications. In this paper, we implement the fractal framework to obtain a relation between the fractal parameters of a SAR image and those of the relative imaged area. In addition, a case study is discussed, showing the potentiality of our framework for flooding detection

SAR Imaging of Fractal Surfaces

A complete theoretical model for synthetic aperture radar (SAR) imaging of natural surfaces is introduced in this paper. The topography of the natural scenes is described via models derived from fractal geometry; scattering evaluations are performed via fractal scattering models appropriate to the employed fractal scene description. Scattering contributions are combined according to the SAR image impulse response function. The power spectral density of appropriate cuts of the SAR image are evaluated in closed form in terms of the surface fractal parameters. Our theoretical model is here conceptually assessed, analytically derived, graphically validated, numerically verified, and also tested on simulated SAR images. The introduced model allows defining innovative postprocessing inverse techniques to retrieve fractal parameters directly from SAR images.

Angle Independence Properties of Fractal Dimension Maps Estimated From SAR Data

The extremely remarkable properties of angle independence exhibited by an innovative SAR product, the fractal dimension map estimated from a single SAR image, are discussed. The theoretical analysis is supported by a noticeable data set of actual SAR images acquired, with look angles varying from 20° to 45°, in the stripmap operational mode by the COSMO-SkyMed constellation. The behavior of the fractal dimension maps at different look angles is discussed for both natural and urban scenarios and emphasis is also posed on areas within the same image that, according to the scene macroscopic topography, are characterized by different incidence angles. The whole analysis is aimed at highlighting, on the one hand, the specific independencies of natural surface fractal dimension maps from the look angle and from the local incidence angle, which can be very useful in information extraction and SAR post-processing techniques and, on the other hand, the different fractal dimension maps behavior whereas urban areas are analyzed.

SAR simulation of ocean scenes covered by oil slicks with arbitrary shapes

The identification of oil slicks on the ocean surface from SAR data requires quantitative sound models accounting for the most important characteristics (ocean spectrum, slick viscosity, slick shape, and so on). In this paper we present the implementation of an innovative SAR raw signal and image simulator, which is able to reproduce images relative to ocean surfaces covered by oil slicks with arbitrary shapes. The attention is mainly focused on slicks with fractal contours. The fractal Weierstrass-Mandelbrot function is used to generate slicks with fixed fractal dimension. A box counting technique is employed to evaluate the fractal dimensions of the generated slicks and the corresponding SAR images. Radiometric properties of the area covered by oil are also estimated in order to show how the simulated data provide a powerful set for processing algorithms.

An end-user-oriented framework for the classification of multitemporal SAR images

ABSTRACT In this article, we present an end-user-oriented framework for multitemporal synthetic aperture radar (SAR) data classification. It accepts as input the recently introduced Level-1α products, whose peculiarities are a high degree of interpretability and increased class separability with respect to single greyscale images. These properties make the Level-1α products very attractive in the application of simple supervised classification algorithms. Specifically, (1) the high degree of interpretability of the maps makes the training phase extremely simple; and (2) the good separation between classes gives excellent results using simple discrimination rules. The end product is a simple, fast, accurate, and repeatable framework.

Characterization of local regularity in SAR Imagery by means of multiscale techniques: application to oil spill detection

Thanks to their capability to cover large areas, in all weather conditions, during the day as well as during the night, spaceborne Synthetic Aperture Radar (SAR) techniques constitute an extremely promising alternative to traditional surveillance methods. Nevertheless, in order to assure further usability of SAR images, specific data mining tools are still to be developed to provide an efficient automatic interpretation of SAR data. The aim of this paper is to introduce texture analysis performed in the framework of time - frequency theory, as a means to detect oil spills in the sea surface. In particular, an algorithm permitting a precise quantitative characterization of the border between the oil spill candidate and the sea, will allow a novel classification of oil spills and look-alikes.

Use of high-resolution SAR data for the monitoring of water resources in Burkina Faso

The integrated management of water resources is a crucial problem for improving the quality of life in Sub-Saharian Africa. Several satellites everyday acquire a huge amount of physical information that could be employed as a support for solving agriculture and water problems. In this paper we present a project devoted to exploit the use of high resolution synthetic aperture radar (SAR) images for water resource management at no cost for the users. A case study is developed in the Yatenga region, in the northern Burkina Faso, integrating hydrologic and remote sensing models in order to improve the capacity of predicting flood and drought events. Main attention is posed here on the innovative fractal techniques developed for the extraction of geometrical and physical parameters that can be used for calibrating hydro-geological models.

Monitoring of Flooding in Urban Areas

Urban areas are crowded environments, where a disaster can bring dramatic consequences, if not adequately forecasted and faced. Remote sensing instruments can be fruitfully used for both prediction and aid organization purposes. In particular, in this paper we present innovative synthetic aperture radar (SAR) techniques for the detection of a flooded area in urban settlements. A SAR raw signal simulator is presented and used, in order to improve the comprehension of the main physical phenomena and to plan the most adequate sensor characteristics for detection purposes. The single and multiple scattering phenomena, in conjunction with strong layover effects make the SAR images relative to urban areas extremely involved. The presented study is focused on a canonical environment, in order to provide a complete and powerful instrument for the comprehension of the complex texture of urban area SAR images.

The Effects of Finite Resolution on Radar Images of Fractal Profiles

In this paper a first step toward a complete model of the fractal imaging process is taken: for the sake of simplicity the mathematical details are here provided for a fractal profile with topological dimension equal to one. In particular, we show how the signal backscattered from a fractal profile modeled as a fractional Brownian motion (fBm) stochastic process is strictly linked to an associated fractional Gaussian noise (fGn) process. We compute in closed form the power density spectrum of the received signal in the simplified hypothesis of a linear dependence of the backscattered signal on the profile derivative process. Our results apply to physical fBm processes, as dictated by the low-pass filtering introduced by both the incident electromagnetic field wavelength and the finite sensor resolution. In the last section a numerical study of the above mentioned is also provided.

Fractal parameters and SAR images

The fractal geometry proved to be the most appropriate mathematical instrument in describing natural scenes, by means of few effective and reliable geophysical parameters. In this paper we describe a complete processing chain for the retrieving of SAR image fractal parameters and for change detection purposes. We present an overall framework employing fractal based models, algorithms and tools to support identification of natural area changes due to natural or man-made disasters. In addition, we test our fractal framework in a simulated disaster scenario. In particular, we consider the case of a simulated volcano eruption scenario to test the potentialities of our technique for lava flow detection. We also compare the performances of the proposed framework on different kinds of disasters to stress the significant differences in the parameters used for change detection purposes.