• 16:30 Spectral Method For Bright Scatterers Resolution Enhancement

  Weissgerber, Flora (1,2); Trouve, Nicolas (2); Koeniguer, Elise (2); Nicolas, Jean-Marie (1) 1: Institut Mines-Telecom ; Telecom ParisTech ; CNRS LTCI; 2: ONERA, the French Aerospace Lab

  Show abstract

  Urban area monitoring is a demanding task that can be supported by SAR imagery. Polarimetry improves target characterization but decreases the resolution. To over- come this difficulty, we designed a resolution enhancement solution based on a spectral approach. The developed algorithm aims to focus point targets while preserving their polarimetric responses. Results are illustrated on the city of Toulouse. The images used are 2007 PolInSAR ONERA acquisition. Isolated bright scatterers are correctly focused. They conserve the low resolution polarization features.

  
   

• 17:30 A Heterogeneity Index and Its Application to Parameter Estimation of Multilook PolSAR Clutter

  Liu, meng; Zhang, Hong; wang, chao; Xie, lei Institute of Remote Sensing and Digital Earth, CAS, China, People's Republic of

  Show abstract

  Fully polarimetric SAR (PolSAR) systems describe the interactions between the electromagnetic wave and the target area with all combination of transmit and receive polarization, and most of the recently launched PolSAR systems have the capability of polarimetric imaging. Due to the coherent interference of waves reflected from many elementary scatterers, the PolSAR image is highly contaminated by speckle noise, which complicates image interpretation and reduces the effectiveness of image segmentation and feature classification. Therefore, understanding SAR speckle statistics is essential for better information extraction by designing intelligent algorithms for speckle filtering, geophysical parameter estimation, and land-use, ground cover classification [1]. Recently, the speckle distribution models proposed for PolSAR data can be divided into two categories: homogeneous and heterogeneous clutter models. The typical representative of homogeneous clutter models for single look PolSAR data is a Gaussian distribution model, which has been used in low resolution images [1]. However, with the new generation of airborne and spaceborne SAR sensors, thinner spatial features can be observed from the high resolution of newly available PolSAR images. Recent studies show that the higher scene heterogeneity leads to inhomogeneous clutter modeling, particularly for urban areas [2]. One commonly used fully polarimetric inhomogeneous clutter model is the spherically invariant random vector (SIRV) distribution model. In this model, the multiplicative model has been employed for PolSAR scattering vector as a product between the square root of a scalar positive quantity (texture) and the description of an equivalent homogeneous surface (speckle) [3]. For data compression and preliminary speckle reduction, the PolSAR data are frequently multilook processed by spatial averaging at the expense of loss of spatial resolution. In this case, a more appropriate representation of the PolSAR data is the polarimetric covariance or coherency matrix. Under the homogeneous clutter assumption, the well-known Wishart distribution has been derived to describe the polarimetric coherency matrix [1]. And for “heterogeneous” or “textured” scenes, a multilook product clutter model, considered an extension of the SIRV distribution model in the single-look case, is presented to describe the coherency matrix of PolSAR clutter in [4]. When the PolSAR statistical model is determined, its parameters should be estimated based on the method of moments or the maximum likelihood methodology. Recently, the method of log-cumulants is also included as a possible parameter estimation approach [5]. Although the heterogeneous model exhibit better performance than the homogeneous model in high-resolution PolSAR clutter fitting, the application using the homogeneous distribution model is approximately 50–90 times faster than that of using the heterogeneous clutter model [4][6].This is due to the high complexity of calculation. Therefore, the choice of algorithm has to be comprised between the quality of the result and the time consumption of the application. The objective of this paper is to propose an optimal estimation strategy for multilook PolSAR clutter fitting. First, a heterogeneity index is derived to quantify the heterogeneity degree of the PolSAR data. It is expected that the heterogeneous areas will give a higher index than other areas. Then, the probability density function (PDF) of the proposed index is derived through the general asymptotic expansion theory for the large-sample distribution. And a constant false alarm rate (CFAR) method based on this asymptotic PDF is applied to divide the PolSAR image into two categories: the heterogeneous areas and the homogeneous areas. For the heterogeneous areas, it is better to estimate clutter parameters using the multilook product clutter model. And it is reasonable to use the Wishart distribution to fit the homogeneous areas. To illustrate the capability of the estimation strategy, a Radarsat-2 fully polarimetric SAR image taken at Shihu Lake area in Suzhou, China, acquired on August 29, 2014, was used for our experiment. And the generalized Gamma distribution has been used to represent the texture of the multilook product clutter model. Compared with the method only based on heterogeneous model, it is shown that the proposed strategy can significantly reduce the high computation complexity. REFERENCES [1] J. S. Lee and E. Pottier, Polarimetric Radar Imaging: From Basics to Applications. Florida: CRC, 2009. [2] M. S. Greco and F. Gini, "Statistical Analysis of High-Resolution SAR Ground Clutter Data," IEEE Trans. Geosci. Remote Sens., vol. 45, no. 3, pp. 566-575, 2007. [3] G. Vasile, F. Pascal, J. Ovarlez, P. Formont, and M. Gay, "Optimal Parameter Estimation in Heterogeneous Clutter for High-Resolution Polarimetric SAR Data," IEEE Geosci. Remote Sens. Lett., vol. 8, no. 6, pp. 1046-1050, 2011. [4] M. Liu, H. Zhang, C. Wang, and F. Wu, "Change Detection of Multilook Polarimetric SAR Images Using Heterogeneous Clutter Models," IEEE Trans. Geosci. Remote Sens., vol. 52, no. 12, pp. 7483-7494, Dec. 2014. [5] V. A. Krylov, G. Moser, S. B. Serpico, and J. Zerubia, "On the Method of Logarithmic Cumulants for Parametric Probability Density Function Estimation," IEEE Geosci. Image Process., vol. 22, no. 10, pp. 3791-3806, Oct. 2013.. [6] X.-V. Phan, L. Bombrun, G. Vasile, and M. Gay, "Heterogeneous Clutter Models for Change Detection in PolSAR Imagery," in Proc. PolInSAR, Frascati, Italy, 2011.

  
   

• 17:10 Investigation of Variations in the Equivalent Number of Looks for Polarimetric Channels

  HU, Dingsheng (1,2); Anfinsen, Stian Normann (1); Tao, Ding (1); Qiu, Xiaolan (2) 1: University of Tromso, Department of Physics and Technology, Earth Observation Laboratory, Norway; 2: Institute of Electronics, Chinese Academy of Sciences, Beijing, China

  Show abstract

  This paper investigates deviations in the estimated equivalent number of looks (ENL) that are observed when processing polarimetric synthetic aperture radar images of ocean surfaces. Even for surface that appear to be homogeneous, the estimated ENL is significantly different in cross-polarimetric (cross-pol) and co-polarimetric (co-pol) channels. We have formulated two hypotheses for the cause of this. The first hypothesis is that the co-pol channel is more sensitive to fine-scale oceanographic features, and the image sample is therefore more heterogeneous results in this case, resulting in a lowered ENL value due to image texture. The other hypothesis is that the cross-pol measurements of ocean are close to the noise floor, which also affects the ENL estimate, and ultimately requires a different signal model and estimator. In the investigation we employ different approaches to ENL estimation. One approach is to estimate each polarimetric channel separately; another is based on the full covariance matrix and combines the statistical information of all polarimetric channels, as proposed in [1]. We also use a recent texture invariant estimator proposed in [2]. Note that various factors, like mixture of classes, texture, and noise level, can cause variances of the ENL estimated from each channel, which leads us to question the validity of using the full polarimetric information to estimate a global ENL, in terms of accuracy and rationality. In the paper, we investigate typical scenarios in real PolSAR images, and analysis the reasons causing the differences between channels in each case. [1] S. N. Anfinsen, A. P. Doulgeris, and T. Eltoft, "Estimation of the Equivalent Number of Looks in Polarimetric Synthetic Aperture Radar Imagery," Geoscience and Remote Sensing, IEEE Transactions on, vol. 47, pp. 3795-3809, 2009. [2] T. Liu, H.-g. Cui, Z.-m. Xi, and J. Gao, "Texture-Invariant Estimation of Equivalent Number of Looks Based on Trace Moments in Polarimetric Radar Imagery," Geoscience and Remote Sensing Letters, IEEE, vol. 11, pp. 1129-1133, 2014.

  
   

• 16:50 Precipitating cloud effects on the radar polarimetric signature at Ka band

  Mori, Saverio (1); Marzano, Frank Silvio (1); Pierdicca, Nazzareno (1); Polverari, Federica (1); Mereu, Luigi (1); Rommen, Bjorn (2) 1: University "La Sapienza" Rome (IT), Italy; 2: ESA-ESTEC, Nordwjick , The Netherland

  Show abstract

  In recent years operational missions and many investigations have assessed the potentialities of spaceborne Synthetic Aperture Radars (SARs) operating at X band and above to monitoring the Earth surface. The most appealing aspect of these instruments is the capability of observing the Earth at very high spatial resolution (of the order of meters) and with a well-known sensitivity to water content, ground roughness, ground displacement (i.e., interferometry applications) and so on. These characteristics make this kind of instrument very suitable for monitoring floods, earthquakes, volcanoes, urban areas, infrastructures, land use and marine surfaces and to produce Digital Elevation Models (DEMs) with a resolution of few meters (e.g. COSMO-SkyMed, TerraSAR-X/TanDEM-X systems). Moreover, fully-polarimetric SAR systems allow the complete characterization of target scattering properties to be measured, improving target classification capabilities. The experience gained throughout the mentioned missions has stimulated the development of SAR systems operating at higher frequencies, such as Ka-Band (around 35 GHz) and above, subject of many investigations and even airborne prototype development. The possibility to implement a single pass satellite interferometer thanks to the small along track baseline required at this frequency is very attractive. In this respect sensitivity to atmospheric effects is a major concern. Numerous works in the last years have assessed the sensitivity of spaceborne SARs operating at frequencies above C band. In particular atmospheric precipitations have demonstrated to condition SAR response echoes, both in amplitude and in phase. Also ground resolution can be reduced due to the turbulent flow of hydrometeors during precipitations. The impact of precipitation on SAR slant-view imagery is due to a combination of surface and volumetric backscattering, coupled with path attenuation and with a significant dependence on frequency, polarization, spatial distribution of hydrometeors and their electromagnetic properties. A valid aid to analyze and characterize SAR response to precipitation is represented by forward modeling, where a known scenario is used to simulate SAR echoes. In this work we will introduce a simulation framework developed to characterize precipitating clouds effects on spaceborne X- and Ka-Band SAR systems. The proposed model framework account for the SAR slant observing geometry and it is able to characterize the polarimetric SAR response echoes both in amplitude and phase when a precipitating cloud is present. In particular, in this work we have derived the Normalized Radar Cross Sections (NRCSs) and the complex correlation coefficients among copolarized returns, which are the most significant elements of the covariance matrix measured by polarimetric radar, starting from the covariance matrix of the surface target and the properties of the superimposed scattering atmosphere. Polarimetric SAR NRCSs can be model through the combination of two contributes, a surface one, accounting for ground backscattering, and a volume one, accounting for the clouds reflectivity and extinction; both contribution are weighted by the two-way path attenuation. Radar observed complex correlation coefficient, useful to characterize the signal differential phase shift (i.e., between horizontal and vertical polarizations), can be modelled through the correlation coefficient of the ground target and hydrometeors reflectivity and specific differential phase, both weighted by the two way path attenuation. The target scenario is simulated in a realistic way, both for atmospheric and surface contributions. The System for Atmospheric Modeling (SAM) high-resolution mesoscale model is used to extract the three-dimensional distribution of atmospheric hydrometeors, both precipitating (rain, snow, graupel) and non-precipitating (cloud ice and cloud liquid water), expressed by their water content. Scattering properties of the considered hydrometeors have been characterized through HESS (Hydrometeor Ensemble Scattering Simulator) T-Matrix simulations. In this work we have used SAM datasets consisting of volumes of 64 x 64 km2 and 30 km height, ground resolution of 250 m and varying vertical resolution, ranging from less than 250 m up to 3 km height, and until 800 m near 30 km height. Dataset dimensions and resolution are comparable with those achievable from spaceborne SAR observing a turbulent medium. The whole dataset have been systematically analyzed and a significant number of vertical sections (each of them representing a simulation case study) have been selected among the available SAM 3-D structures, as representative of the variety of clouds systems (e.g. convective or stratiform). The objective is the characterization of precipitating clouds effects on spaceborne Ka-Band SAR polarimetric response. The scenario is completed by the surface scattering model. The framework can be easily configured to simulate bare soils, using for instance Semi-Empirical Models, as well as surface canonical targets, such as dihedrals or volume scattering, for an ease analysis of the effect of the atmosphere on the polarimetric signature and a comparison with literature works. The work has been accomplished in the framework of an ESA project (Contract ESTEC N. 4000109477/13/nl/lvh) aiming at supporting instrument parametric analyses and establishing system requirements, such as SNR, NESZ, transmitting power, antenna gain, image ambiguities and so on, of a spaceborne Ka polarimetric/interferometric radar. Preliminary analysis confirms the sensitivity of SARs X-band and Ka-band instruments even to moderate atmospheric precipitations. Ka-Band shows a much greater sensitivity to atmospheric effects respect to X-Band, as expected. Rain path attenuation dramatically reduces NRCS to instrument sensitivity limit. Nevertheless Ka-band correlation coefficient appears less influenced by oblate particles (raindrops) and more by spherical frozen ones. An important sensitivity to frozen particles can be observed also in NRCS. The study represent a novel attempt to characterize the atmospheric effects on SAR data, not only in terms of path delay or attenuation, but investigating the degree of modification of the complete polarimetric signature of the surface target.

  
   

• 17:50 Round Table Discussion

  Show abstract

  Round Table Discussion 01

  
   

Highlights

In May 2013, the Biomass mission concept was selected to become the next in the series of satellites developed to further our understanding of Earth.

Released: 21/06/2013

REPLAY EARTH EXPLORER 7 USER CONSULTATION MEETING (PART 1)

Released: 11/03/2013

POLinSAR 2013 Workshop

Released: 22/02/2013