Rafael Zandona Schneider

Polarimetric and interferometric characterization of coherent scatterers in urban areas

In this paper the concept of point-like coherent scatterers (CSs) in urban areas is introduced. The detection of CSs in single and quad-polarized images is addressed and applied on polarimetric and interferometric high-resolution airborne SAR data at L-band. For the detection of CSs two different approaches are proposed and their individual performance is analyzed. The properties of the detected CSs and their polarimetric and interferometric characteristics are assessed. The role of polarimetry on the detection of the CSs is evaluated and the potential of extracting the orientation and dielectric properties of individual CSs is finally addressed.

Estimation and correction of ionospheric induced phase errors in SAR images using Coherent Scatterers

Coherent Scatterers (CSs) are usually detected in SAR data using spectral correlation techniques, by evaluating multilook images in the range direction. The first part of the paper demonstrates an alternative way to detect CSs by exploiting their inherent strong phase stability along frequency. The main advantage of the technique is the absence of a spatial average (necessary in the original procedure), which avoids loss of resolution. The second part of this work illustrates the use of CSs for the estimation and correction of phase errors along the synthetic aperture. Such source of errors may be originated by different effects as platform motion and atmospheric or ionospheric disturbances. For the correction of one and two-dimensional phase errors addressed here, a CSs based autofocus procedure was implemented and applied to simulated and real data.

Polarimetric interferometry over urban areas: information extraction using coherent scatterers

In this paper the detection of point-like coherent scatterers (CSs) in single and quad-pol images is addressed and applied on polarimetric and interferometric high-resolution SAR data. For the detection of CSs two different approaches are described, both based on spectral correlation. The effect of multi acquisitions on the number and quality of CSs is verified and the results are discussed. The role of polarimetry on the detection is evaluated and the potential of polarimetric based information extraction as the estimation of the orientation angle – is addressed.

Entropy among a sequency of SAR images for change detection

The coherence map between two images is frequently used to detect changes in a sequence of temporal SAR images and to evaluate the interferometric image quality. Recently, an alternative method for change detection was proposed: the entropy image. This paper performs the entropy calculation among several temporal SAR images. With this method, it is possible to calculate the entropy between two images considering intermediary, previously, or later acquired images, which contain information about the change process. It is shown that this method can improve the change detection performance in low coherence areas (using intermediary images) or in high coherence areas (using former or further images). Simulated images were used to show the performance of the method and to compare this method with coherence and entropy only between two images.

Pol-dinSAR: polarimetric SAR differential interferometry using coherent scatterers

We investigate the impact of polarimetry and polarimetric effects on the Differential Interferometry (DinSAR) technique for the coherent backscattering reciprocal case, with the aim to obtain better accuracy and to extract additional information about the scatterer. Emphasis is given to the effects of Faraday and scatterer line-of-sight (LOS) rotations, and to the determination of the intrinsic scatterer scattering mechanism.

Characterisation of Coherent Scatterers in Urban Areas by Means of Angular Diversity

The detection of scatterers with pointwise response in SAR images of urban areas using the so-called Coherent Scatterers (CSs) technique was in previous works introduced (1), (2). In this paper, the potential to exploit sub-aperture images from wide azimuth angular apertures (as it is the case for air- borne systems) to estimate CSs backscattering radiation patterns is addressed. The backscattering radiation pattern is modeled assuming a simple canonical class of scatterers. The model is then inverted using the estimated radiation patterns of individual CSs and their geometrical parameters are evaluated and related to CSs polarimetric properties. Two experiments using dihedral corner reflectors with differentt azimuth orientation and Line of Sight (LOS) rotation angles were performed, in order to verify the agreement with the theory. The data used are at L-band from the airborne E-SAR system of the German Aerospace Center (DLR) and the testsite is the Munich city in Germany. I. INTRODUCTION The Coherent Scatterers (CSs) technique to detect scatterers with point-like behaviour in SAR images of urban areas was introduced in (1), (2). Polarimetric and interferometric characterisation of CSs has been discussed, as well as the potential of physical and geometrical information extraction, as the estimation of the scatterers Line Of Sight (LOS) rotation angle and CSs dielectric properties. In (3) the backscattering radiation pattern in the azimuth direction of the so-called Permanent Scatterers (PSs) was estimated using ERS-2 image series and taking advantage of its loss of the on-board gyroscopes. This allowed to acquire images of the same scene with different azimuth aspect angles. A backscattering pattern model for canonical PSs was used to invert PSs geometrical parameters. In this work, azimuth backscattering radiation patterns of CSs are estimated using multi-look images from wide azimuth bandwidth systems. The backscattering radiation patterns are modeled as in (3) and the model parameters are inverted. Their relationship to the scatterers type estimated through polarimetry is investigated. Finally, two experiments using dihedral corner reflectors were performed, in order to verify the agreement with the theory.

On the accuracy of scatterers LOS rotation estimation procedures in radar polarimetry

The Line-of-Sight (LOS) rotation angle of scatterers can be estimated using polarimetric radar measurements. Exact estimation is unrealistic since noise and clutter are almost always present in the data. The paper presents the derivation of an approximated distribution, mean bias and variance analysis, of the scatterers LOS rotation angle using polarimetry, when estimated for the high-coherent (quasi-deterministic) scattering case. Example of scatterers fullfilling such condition are the so-called Coherent Scatterers (CSs) [1], [2].Different procedures for the estimation of the rotation angle of scatterers around the radar Line-Of-Sight using polarimetric data have been proposed in the literature. In [1], a deep analysis of the fundamentals of the procedures has been carried out and some common aspects among the procedures were demonstrated. Particularly, it has been shown that many procedures are in fact the same and can be divided into just two classes. The two classes are fundamentally different based on two different principles: The Rotation Transformation Based Routines (RTBR) and the Consimilarity Transformation Based Routines (CTBR). For a certain group of scatterers the two classes lead to the same rotation angle, but for most general scatterers the two classes of LOS rotation estimation procedures lead to different rotation estimation.