The Role of Resolution in the Estimation of Fractal Dimension Maps From SAR Data

Gerardo Di Martino,Daniele Riccio,Ivana Zinno,Antonio Iodice,Giuseppe Ruello

doi:10.3390/rs10010009

Abstract

This work is aimed at investigating the role of resolution in fractal dimension map estimation, analyzing the role of the different surface spatial scales involved in the considered estimation process. The study is performed using a data set of actual Cosmo/SkyMed Synthetic Aperture Radar (SAR) images relevant to two different areas, the region of Bidi in Burkina Faso and the city of Naples in Italy, acquired in stripmap and enhanced spotlight modes. The behavior of fractal dimension maps in the presence of areas with distinctive characteristics from the viewpoint of land-cover and surface features is discussed. Significant differences among the estimated maps are obtained in the presence of fine textural details, which significantly affect the fractal dimension estimation for the higher resolution spotlight images. The obtained results show that if we are interested in obtaining a reliable estimate of the fractal dimension of the observed natural scene, stripmap images should be chosen in view of both economic and computational considerations. In turn, the combination of fractal dimension maps obtained from stripmap and spotlight images can be used to identify areas on the scene presenting non-fractal behavior (e.g., urban areas). Along this guideline, a simple example of stripmap-spotlight data fusion is also presented.

Highlights

Cosmo/SkyMed synthetic aperture radar (SAR) multi-operational capabilities allow the observation of a scene at different spatial scales, i.e., with different levels of detail
We dictated consideron images obtained stripmap andwhich enhanced spotlight scales in the estimation are the one hand byinthe resolution, is related to so that theon area by the spotlight is a subset of the areaisimaged theoperational minimum modes, spatial scale, and thecovered other hand by the size ofimages the sliding window, which related in mode.scale
This fact, in principle, stripmap image was acquired on 18 August 2011 and the enhanced spotlight image on 22 August can lead to different values of the fractal dimension on the two maps, for the reasons discussed in the

Summary

Introduction

Cosmo/SkyMed synthetic aperture radar (SAR) multi-operational capabilities allow the observation of a scene at different spatial scales, i.e., with different levels of detail. In this context, stripmap images, or even better, scansar images, are well suited for large-scale applications, whereas spotlight products can be fruitfully used for local-scale analyses, able to provide more detailed information on the observed scene. The combined use of multi-operational data, can be set in the framework of a multi-scale approach, in which coarse resolution images can provide the necessary regional-scale survey of the area of interest, while high resolution images can be used to refine the analysis on a local scale.

Objectives

Methods

Results