This PDF file contains the front matter associated with SPIE Proceedings Volume 9642 including the Title Page, Copyright information, Table of Contents, Introduction, and Conference Committee listing.

In this work, a multitemporal algorithm (MLTA), originally conceived for the C-band radar aboard the Sentinel-1 satellite, has been updated in order to retrieve soil moisture from L-Band radar data, such as those provided by the NASA Soil Moisture Active Passive (SMAP) mission. Such type of algorithm may deliver frequent and more accurate soil moisture maps mitigating the effect of roughness and vegetation changes, which are assumed to occur at longer temporal scales with respect to the soil moisture changes. Within the multitemporal inversion scheme based on the Bayesian Maximum A Priori (MAP) criterion, a dense time series of radar measurements is integrated to invert a forward backscattering model which includes the contribution from vegetation. The calibration and validation tasks have been accomplished by using the data collected during the SMAP Validation Experiment 12.The SMAPVEX12 campaign consists of L-Band images collected by the UAVSAR sensor, in situ soil moisture data and measurements of vegetation parameters, collected during the growing season of several crops (pasture, wheat, soybean, corn, etc.). They have been used to update the forward model for bare soil scattering at L-band with respect to the Oh and Sarabandi model previously used at C band. Moreover, the SMAPVEX12 data have been also used to tune a simple vegetation scattering model which considers two different classes of vegetation: those producing mainly single scattering effects, and those characterized by a significant multiple scattering involving terrain surface and vegetation elements interaction.

This paper investigates the ability of the Information Theoretic Snow Detection Algorithm (ITSDA) in detecting changes due to snow cover between summer and winter seasons on large area images acquired by COSMO-SkyMed constellation. ITSDA is a method for change detection in multitemporal SAR images, which has been recently applied by the authors to a subset of Cosmo-SkyMed data. The proposed technique is based on a nonparametric approach in the framework of Shannon’s information theory, and in particular it features the conditional probability of the local means between the two images taken at different times. Such an unsupervised approach does not require any preliminary despeckling procedure to be performed before the calculation of the change map. In the case of a low quantity of anomalous changes in relatively small-size images, a mean shift procedure can be utilized for refining the map. However, in the present investigation, the changes to be identified are pervasive in large size images. Consequently, for computational issues, the mean shift refinement has been omitted in the present work. However, a simplified implementation of mean shift procedure to save time will be possibly considered in future submissions. In any case, the present version of ITSDA method preserve its characteristics of flexibility and sensibility to backscattering changes, thanks to the possibility of setting up the number of quantization levels in the estimation of the conditional probability between the amplitude values at the two acquisition dates.

Land-cover classification in Synthetic Aperture Radar (SAR) images has significance in both civil and military remote sensing applications. Accurate classification is a challenging problem due to variety of natural and man-made objects, seasonal changes at acquisition time, and diversity of image reconstruction algorithms.. In this study, Feature Preserving Despeckling (FPD), which is an edge preserving total variation based speckle reduction method, is applied as a preprocessing step. To handle the mentioned challenges, a novel feature extraction schema combined with a super-pixel segmentation and dictionary learning based classification is proposed. Computational complexity is another issue to handle in processing of high dimensional SAR images. Computational complexity of the proposed method is linearly proportional to the size of the image since it does not require a sliding window that accesses the pixels multiple times. Accuracy of the proposed method is validated on the dataset composed of TerraSAR-X high resolutions spot mode SAR images.

This research aims at investigating the backscatter sensitivity at C and X band to the characteristics of agricultural surfaces and analyzing the integration of these data collected from Radarsat2 (RS2) and COSMO-SkyMed (CSK) systems on tree agricultural test areas in Italy (San Pietro Capofiume, in Emilia Romagna, Sesto Fiorentino, in Tuscany, and Mazia Valley, in South Tyrol).

A preliminary test of the sensitivity of SAR signal to the soil and vegetation characteristics was first carried out by also comparing data from previous experiments. From these results, it can be concluded that X-band data are mainly sensitive to vegetation structure and biomass, and to soil moisture of bare or slightly vegetate soils, whereas C-band images could provide valuable information for the retrieval of soil moisture, even in vegetation covered soils.

Two retrieval algorithms were implemented for estimating the main geophysical parameters, namely soil moisture content (SMC) and vegetation biomass (PWC) from these sensors. Over Sesto Fiorentino area, an algorithm based on Artificial Neural Network (ANN) technique was implemented for estimating both SMC of bare or scarcely vegetated soil and vegetation biomass of wheat crops at X band. On the South-Tyrol area, a SMC retrieval approach based on the Support Vector Regression methodology, which was already tested in this area using C-band data from ENVISAT/ASAR data, was adopted. This algorithm integrated data at both X and C bands showing encouraging results, even though further investigations shall be carried out on a larger time-series and larger set of samples.

This study presents the use of multi-temporal Synthetic Aperture Radar (SAR) images for detection of land cover changes in the eastern part of the Island of Crete (Greece). For this purpose, fourteen Envisat ASAR acquisitions from July 2004 to December 2006 were calibrated and registered.

We applied a temporal filter and spatial averaging to the backscatter intensity to reduce the noise. Furthermore, we used the concept that the changes between different backscatter intensity observations can show changes on the target dielectric properties. In order to detect changes due to geometrical characteristics of land cover types, we created coherence maps using twenty-seven interferometric pairs with proper spatial and temporal baselines. In all calculations, layover and shadow effects, as well as the sea, were masked by using information from the digital elevation model of the area. The observed changes in the coherence values were analyzed with respect to different decorrelation factors that can contribute to the loss of coherence.

Our results present the different backscatter values for several land cover types (farmland, olive groves, forests, etc.). In addition, some land cover types such as olive groves show variations of backscatter signal due to the density and height of trees. Furthermore, olive groves show good coherence in interferograms with short time intervals. All interferometric pairs have low coherence in farmland because of the rapid growth of plants. Finally, the maps of backscatter temporal changes and coherence changes were superimposed and compared to auxiliary data such as multi-temporal optical satellite imagery (i.e. Landsat/ETM, Terra/Aqua MODIS) and thematic land cover maps (Corinne 2000). We found that changes are mostly due to plant growth and man-made activity.

This ongoing study shows the potential of SAR in providing complementary information such as changes in dielectric and geometric properties to optical data in land cover dynamics monitoring.

Huynen decomposition prefers the world of basic symmetry and regularity (SR) in which we live. However, this preference restricts its applicability to ideal SR scatterer only. As for the complex non-symmetric (NS) and irregular (IR) scatterers such as forest and building, Huynen decomposition fails to analyze their scattering. The canonical Huynen dichotomy is devised to extend Huynen decomposition to the preferences for IR and NS. From the physical realizability conditions of polarimetric scattering description, two other dichotomies of polarimetric radar target are developed, which prefer scattering IR, and NS, respectively, and provide two competent supplements to Huynen decomposition. The canonical Huynen dichotomy is the combination of the two dichotomies and Huynen decomposition. In virtue of an Adaptive selection, the canonical Huynen dichotomy is used in target extraction, and the experiments on AIRSAR San Francisco data demonstrate its high efficiency and excellent discrimination of radar targets.

For the purpose of getting sensitive information relevant to civil or military security, high-resolution airborne Synthetic Aperture Radar (SAR) provides the possibility to organize missions at short notice regardless of the daylight and of the weather conditions. The use of compact millimeter-wave FMCW SAR systems allows reaching these goals more safely and at lower cost using unmanned lightweight platforms. As a counterpart these platforms are relatively unstable, making the data-processing more difficult. In order to reach optimum focusing quality also in unfavorable flight conditions or for highly non-linear tracks we developed a fast Time-Domain Processor that relies on parallelization using the GPU resources. A production areal processing rate as high as 6 km²/h using 20 cm ground pixel spacing on a single PC station was achieved. The processing quality and efficiency is demonstrated using real data from the MIRANDA35 Ka-band SAR system.

Due to the improved spatial resolution, Earth observation (EO) data, either from Synthetic Aperture Radar (SAR) or optical sensor, provide the opportunity to assess earthquake damage of individual buildings. However, the operational use of EO data for earthquake damage mapping is basically limited to the visual inspection of Very High Resolution (VHR) optical imagery. In this work we investigate the feasibility of a damage assessment product at single building scale from a pair of VHR SAR images acquired before and after a seismic event. We perform the change analysis using the Kullbach-Leibler divergence and the intensity ratio and then we associate detected changes to a building map provided as GIS layer. Finally the expected SAR signature of a collapsed building is considered to identify severely damaged buildings. In order to test the proposed methodology we use Spotlight COSMO-SkyMed SAR imagery of L’Aquila (Italy) collected before and after the earthquake occurred on April 6, 2009. A macroseismic survey on the whole central area of L’Aquila city based on the European Macroseismic Scale 1998 is used to assess the capability of VHR SAR images to map damage.

Accurate geolocation of SAR data is nowadays strongly required because of the increasing number of high resolution SAR sensors available as for instance from TerraSAR-X / TanDEM-X and COSMO-SkyMed space-borne missions. Both stripmap and spotlight acquisition modes provide from metric to sub metric spatial resolution which demands the ability to ensure a geolocation accuracy of the same order of magnitude. Geocoding quality depends on several factors and in particular on the knowledge of the actual values of the satellite position along the orbit, and the delay introduced by the additional path induced by changes in the refractivity index due to the presence of the atmosphere (the so called Atmospheric Path Delay or APD). No definitive results are reported yet in the scientific literature, concerning the best performances achievable by the COSMO-SkyMed constellation in terms of geolocation accuracy. Preliminary studies have shown that sub-pixel geolocation accuracies are hardly achievable with COSMO-SkyMed data. The present work aims at inspecting the origin of the geolocation error sources in COSMO-SkyMed Single-look Complex Slant (SCS) products, and to investigate possible strategies for their compensation or mitigation. Five different test sites have been selected in Italy and Argentina, where up to 30 corner reflectors are installed, pointing towards ascending or descending passes. Experimental results are presented and discussed.

With the continuous image product acquisition of satellite missions, the size of the image archives is considerably increasing every day as well as the variety and complexity of their content, surpassing the end-user capacity to analyse and exploit them. Advances in the image retrieval field have contributed to the development of tools for interactive exploration and extraction of the images from huge archives using different parameters like metadata, key-words, and basic image descriptors. Even though we count on more powerful tools for automated image retrieval and data analysis, we still face the problem of understanding and analyzing the results. Thus, a systematic computational analysis of these results is required in order to provide to the end-user a summary of the archive content in comprehensible terms. In this context, visual analytics combines automated analysis with interactive visualizations analysis techniques for an effective understanding, reasoning and decision making on the basis of very large and complex datasets. Moreover, currently several researches are focused on associating the content of the images with semantic definitions for describing the data in a format to be easily understood by the end-user. In this paper, we present our approach for computing visual analytics and semantically querying the TerraSAR-X archive. Our approach is mainly composed of four steps: 1) the generation of a data model that explains the information contained in a TerraSAR-X product. The model is formed by primitive descriptors and metadata entries, 2) the storage of this model in a database system, 3) the semantic definition of the image content based on machine learning algorithms and relevance feedback, and 4) querying the image archive using semantic descriptors as query parameters and computing the statistical analysis of the query results. The experimental results shows that with the help of visual analytics and semantic definitions we are able to explain the image content using semantic terms and the relations between them answering questions such as what is the percentage of urban area in a region? or what is the distribution of water bodies in a city?

Synthetic aperture radar (SAR) systems produce high resolution, two dimensional imaging of areas of environmental interest. SAR interferometry and tomography enables these techniques to extend to three dimensional imaging by exploiting multiple SAR images with diversity in space and time. These techniques require accurate phase information over multiple images as the data is extremely sensitive to deviations from the reference track, therefore to enable interferometry and tomography an accurate autofocus solution is required. This paper investigates phase errors resulting from navigational uncertainties in multipass spotlight SAR imaging and uses techniques from the field of compressive sensing to achieve an autofocus solution. The proposed algorithm builds on previous autofocus work by expanding it to the multipass case and jointly recovers phase errors for all images simultaneously, making it extremely useful for interferometry and tomography techniques. The algorithm described uses pixels that are stable in all SAR images to gain an autofocus solution as these are the pixels that are the focus for analysis using tomography. This is unlike conventional autofocus, which just works on an image-by-image basis. The tools of compressive sensing can be used to concurrently select pixels for bright image elements that are stable and coherent over all images, as these pixels are sparse in the image domain, and calculate the phase errors present in each pass. Using the multipass data after autofocus, height distributions for scatterers in single pixels are determined for simulated forest scenes at X-band. The performance of the autofocus algorithm is examined through numerical simulations and is also applied to real data collected from Selex ES’s airborne, X-band, experimental SAR system. The experimental results demonstrate that the algorithm effectively achieves an autofocus solution. By finding the vertical distribution of two scatterers in a single pixel over simulated forestry images we can determine if compressive sensing can be utilized to gain information on scatterers below the canopy at X-band in future studies.

Volcanic Island can be affected by instability phenomena such as landslide and partial collapse events, even in quiescent period. Starting from data collected by an aerial laser scanning survey at cm-level accuracy), a GIS based approach was implemented in order to perform a landslide-susceptibility analysis. The results of this analysis were compared and integrated with data derived from Differential Synthetic Aperture Radar Interferometry (DinSAR) analysis able to identify the most active areas and quantify the on-going deformation processes. The analysis is focused on the on the active volcanic edifice of Vulcano Island and in some areas of Lipari island, both include in the Eaolian Islands in Sicily (Italy). The developed approach represent a step-forward for the compilation of hazard maps furnishing in an overall contest, updated and georeferenced quantitative data, describing the morphology and the present behaviour of the slopes in the area of investigation.

A new approach to Persistent Scatterer Interferometry (PSI) data processing and analysis implemented in the PSI chain of the Geomatics (PSIG) Division of CTTC is used in this work. The flexibility of the PSIG procedure allowed evaluating two different processing chains of the PSIG procedure. A full PSIG procedure was implemented in the TerraSAR-X dataset while a reduced PSIG procedure was applied to the nine Sentinel-1 images available at the time of processing. The performance of the PSIG procedure is illustrated using X-band and C-band Sentinel-1 data and several examples of deformation maps covering different types of deformation phenomena are shown.

Remote sensing data play an important role for the environmental monitoring because they allow to provide systematic information on very large areas and for a long period of time. Such information must be analyzed, validated and incorporated into proper modeling tools in order to become useful for performing risk assessment analysis. These approaches has been already applied in the field of natural hazard evaluation (i.e. for monitoring seismic, volcanic areas and landslides). However, not enough attention has been devoted to the development of validated methods for implementing quantitative analysis on civil structures.

This work is dedicated to the comprehensive utilization of ERS / ENVISAT data store ESA SAR used to detect deformation trends and perform back-analysis of the investigated structures useful to calibrate the damage assessment models. After this preliminary analysis, SAR data of the new satellite mission (ie Cosmo SkyMed) were adopted to monitor the evolution of existent surface deformation processes and to detect new occurrence. The specific objective was to set up a data processing and data analysis chain tailored on a service that sustains the safe maintenance of the built-up environment, including critical construction such as public (schools, hospital, etc), strategic (dam, highways, etc) and also the cultural heritage sites.

The analysis of the test area, in the southeastern sector of Roma, has provided three different levels and sub-levels of products from metropolitan area scale (territorial analysis), settlement scale (aggregated analysis) to single structure scale (damage degree associated to the structure).

Noise radar has been applied in many fields since it was proposed more than 50 years ago. However, it has not been applied to interferometric SAR imaging yet as far as we know. This paper introduces our recent work on interferometric noise radar. An interferometric SAR system was developed which can transmit both chirp signal and chaotic noise signal (CNS) at multiple carrier frequencies. An airborne experiment with this system by transmitting both signals was carried out, and the data were processed to show the capability of interferometric SAR imaging with CNS. The results shows that although the interferometric phase quality of CNS is degraded due to the signal to noise ratio (SNR) is lower compared with that of chirp signal, we still can get satisfied DEM after multi-looking processing. Another work of this paper is to apply compressed sensing (CS) theory to the interferometric SAR imaging with CNS. The CS theory states that if a signal is sparse, then it can be accurately reconstructed with much less sampled data than that regularly required according to Nyquist Sampling Theory. To form a structured random matrix, if the transmitted signal is of fixed waveform, then random subsampling is needed. However, if the transmitted signal is of random waveform, then only uniform subsampling is needed. This is another advantage of noise signal. Both the interferometric phase images and the DEMs by regular method and by CS method are processed with results compared. It is shown that the degradation of interferometric phases due to subsampling is larger than that of amplitude image.

In recent years, multi-input and multi-output (MIMO) radar has attracted much attention of many researchers and institutions. MIMO radar transmits multiple signals, and receives the backscattered signals reflected from the targets. In contrast with conventional phased array radar and SAR system, MIMO radar system has significant potential advantages for achieving higher system SNR, more accurate parameter estimation, or high resolution of radar image. In this paper, we propose a new MIMO SAR system based on Alamouti space-time coding scheme and orthogonal frequency division multiplexing linearly frequency modulated (OFDM-LFM) for obtaining higher system signal-to-noise ratio (SNR) and better range resolution of SAR image.

Frequency-Modulated Continuous-Wave Synthetic Aperture Radar (FMCW SAR) is a promising compact remote imaging sensor. Its detection and refocusing of the ground moving targets is discussed in this paper in order to provide FMCW SAR system with simultaneous moving targets indication application. A moving target indication method by focusing is derived and presented. This method is derived based on range migration algorithm taking into account the moving target range cell walk. To discriminate the target optimally, the concept of relative motion is extended to FMCW SAR. Due to continuous transmitting and receiving at all time, the standard “stop-go” approximation is not valid here. By searching the relative velocity, the moving target can be focused accurately, equivalent as a fixed target focused in squint mode. The maximum output corresponds to the best matched parameters. Thus the matched targets are detected and the others suppressed.

The purpose of this paper is to compare the performance of two feature extraction methods when applied on high resolution Synthetic Aperture Radar (SAR) images acquired with the new ESA mission SENTINEL-1 (S-1). The feature extraction methods were previously tested on high and very high resolution SAR data (imaged by TerraSAR-X) and had a good performance in discriminating between a relevant numbers of land cover classes (tens of classes). Based on the available spatial resolution (10x10m) of S-1 Interferometric Wide (IW) Ground Range Detected (GRD) images the number of detectable classes is much lower. Moreover, the overall heterogeneity of the images is much lower as compared to the high resolution data, the number of observable details is smaller, and this favors the choice of a smaller window size for the analysis: between 10 and 50 pixels in range and azimuth. The size of the analysis window ensures the consistency with the previous results reported in the literature in very high resolution data (as the size on the ground is comparable and thus the number of contributing objects in the window is similar). The performance of Gabor filters and the Weber Local Descriptor (WLD) was investigated in a twofold approach: first the descriptors were computed directly over the IW GRD images and secondly on the sub-sampled version of the same data (in order to determine the effect of the speckle correlation on the overall class detection probability).

Recently, remote sensing is considering one of the most important tools in studies of water scattering and water characterization. Traditional methods for monitoring pollutants depended on optical satellite rather than Radar data. Thus, many of Water Quality Parameters (WQP) from optical imagery are still limited. In this paper, a new approach based on the TerraSAR-X images has been presented which it is used to map the region of interest and to estimate physical and chemical WQPs. This approach based on a Small Perturbation Model (SPM) for the electromagnetic scattering is applied by using the Elfouhaily spectrum. A series of inversions have been included in this model started by finding the reflectivity from backscattering coefficients which are calculated from SAR images. Another inversion has been applied to find dielectric constant from the calculation models of the reflectivity (in HH and VV polarizations). Then, a Stogryn Debye formulation has been used to estimate temperature and salinity of water surface from SAR images. After many derivations we got a new model able to estimate temperature and salinity directly from backscattering coefficients obtained from radar images. Inorganic chemical parameters which are represented by Total Dissolved Salts (TDS) and the Electrical Conductivity (EC) are estimated directly from salinity. A tow dataset of instu data have been used to validate this work. The validation included a comparison between parameters measured in situ and those estimated from Terra SAR-X image.

This paper proposes a robust matching method for the multi-sensor imagery. Firstly, the SIFT feature matching and relaxation matching method are integrated in the highest pyramid to derive the approximate relationship between the reference and slave image. Then, the normalized Mutual Information and multi-grid multi-level RANSAC algorithm are adopted to find the correct conjugate points. Iteratively perform above steps until the original image level, the facet- based transformation model is used to carry out the image registration. Experiments have been made, and the results show that the method in this paper can deliver large number of evenly distributed conjugate points and realize the accurate registration of optical and SAR multi-sensor imagery.

The paper discusses the major modes of mapping the underlying surface using algorithms synthetic aperture radar antenna system onboard. The causes of the appearance of systematic errors in the mathematical modeling of the synthetic aperture antenna device. A method for reducing these methodological errors, is to increase the number of active reflectors radar signals. It is shown that in this case must take into account correlations between the echoes of individual reflectors. The mathematical model takes into account the spatial and temporal correlation characteristics of the observed signals. The mathematical model is implemented as a simulation algorithm of unsteady, non-Gaussian processes with the specified correlation-spectral characteristics of a given density and distribution of samples of these processes, simulating fluctuations in the amplitude of the observed echo in each bin. The possibility of a significant acceleration of the process simulation using the simplified forms correlations in the form of exponential curves.

In this study, a method for reducing the filtering effects on the interferometric phase signal is proposed. Theoretical analysis showed that while noise reduction is maximized after filtering, the loose of interferometric phase signal is also maximized. This state has been also verified by observations on SAR interferometric data where pixels with high coherence value, which are assumed to contain a lot of information, presented lower coherence values after SAR image filtering.

The proposed method performs interferometric phase modeling. The method recovers the signal after the interferometric filtering for the pixels that loss of information is observed. The selection of these pixels is based on the decrease of their coherence value after the filtering. Signal recovery is associated to the preservation of the initial values for these pixels. Consequently, the method prevents the decrease of the coherence values for these pixels.

Performance of the method depends on the performance of the used filter; however, it always improves the interferometric results. Since the phase signal is the basis for the DEM production, its preservation improves all the steps of the interferometric procedure, especially the phase unwapping. Effects of the method on the final interferometric product, the DEM, are also evident.

The proposed method was evaluated using real interferometric data. Experiments showed that the applied filters within this study, did not always improve the accuracy of the produced DEM. Sub-images for which filtering does not improve their mean coherence value have been selected and the proposed method has been applied. For these sub-images, coherence values and RMS errors of the produced DEMs showed that the method improves the results of the interferometric procedure. It compensates the negative effects of the filtering for these sub-images and leads to the improvement of the DEM accuracy in the majority of the cases.

Urban villages comprise mainly low-rise and congested, often informal settlements surrounded by new constructions and high-rise buildings whereby structures can be very different between neighboring areas. Monitoring urban villages and analyzing their characteristics are crucial for urban development and sustainability research. In this study, we carried out a combined analysis of multispectral GaoFen-1 (GF-1) and high resolution TerraSAR-X radar (TSX) imagery to extract the urban village information. GF-1 and TSX data are combined with the Gramshmidt spectral sharpening method so as to provide new input data for urban village classification. The Grey-Level Co-occurrence Matrix (GLCM) approach was also applied to four directions to provide another four types (all, 0°, 90°, 45° directions) of TSX-based inputs for urban village detection. We analyzed the urban village mapping performance using the Random Forest approach. The results demonstrate that the best overall accuracy and the best producer accuracy of urban villages reached with the GLCM 90° dataset (82.33%, 68.54% respectively). Adding single polarization TSX data as input information to the optical image GF-1 provided an average product accuracy improvement of around 7% in formal built-up area classification. The SAR and optical fusion imagery also provided an effective means to eliminate some layover, shadow effects, and dominant scattering at building locations and green spaces, improving the producer accuracy by 7% in urban area classification. To sum up, the added value of SAR information is demonstrated by the enhanced results achievable over built-up areas, including formal and informal settlements.