This PDF file contains the front matter associated with SPIE Proceedings Volume 8891 including the Title Page, Copyright information, Table of Contents, Introduction, and Conference Committee listing.

Current space borne synthetic aperture radar (SAR) systems are able to provide users with high resolution image data of around one meter. Focusing on systems operating in the X-band, this value is not the end of possible improvements in resolution. There still lies a great potential in an increase of bandwidth of the radar signal itself and also in a significant enlargement of the synthetic aperture. From the technical point of view this certainly is a challenge, but could be possible for future space borne SAR missions already with current state of the art hardware. As a matter of proof TerraSAR-X introduces a new staring spotlight image product that significantly improves the azimuthal resolution to around a quarter of a meter. The technical realization of the very high resolution SAR system is not the only obstacle to overcome. Especially the increase of Doppler bandwidth along the synthetic aperture requests special treatment and considerations in system design from a signal processing’s point of view. Challenges like orbital accuracy, tropospheric effects, approximations in SAR processing methods and depth-of-focus issues have to be addressed. In this paper, most of these challenges are studied separately by performing parametric simulations for single point targets and also for complex signatures of an airplane. A comparable SAR parameter set as used by the high resolution sliding spotlight mode and the new staring spotlight mode of TerraSAR-X are used for simulation.

Multi-Chromatic Analysis (MCA) of SAR images relays on exploring sub-band images obtained by processing portions of range spectrum located at different frequency positions. It has been applied to interferometric pairs for phase uwrapping and height computation. This work investigates two promising applications: the comparison between the frequency-persistent scatterers (PS_fd) and the temporal-persistent scatterers (PS), and the use of inter-band coherence of a single SAR image for vessel detection. The MCA technique introduces the concept of frequency-stable targets, i.e. objects exhibiting stable radar returns across the frequency domain which is complementary to that of temporal stability at the base of PS interferometry. Both spotlight and stripmap TerraSAR-X images acquired on the Venice Lagoon have been processed to identify PS_fd and PS. Different populations have been analyzed to evaluate the respective characteristics and the physical nature of PS_fd and PS. Concerning the spectral coherence, it is derived by computing the coherence between sub-images of a single SAR acquisition. In the presence of a random distribution of surface scatterers, spectral coherence must be proportional to sub-band intersection of sub-images. This model is fully verified when observing measured spectral coherence on open see areas. If scatterers distribution departs from this distribution, as for manmade structures, spectral coherence is preserved. We investigated the spectral coherence to perform vessel detection on sea background by using spotlight images acquired on Venice Lagoon. Sea background tends to lead to very low spectral coherence while this latter is preserved on the targeted vessels, even for very small ones. A first analysis shows that all vessels observable in intensity images are easily detected in the spectral coherence images which can be used as a complementary information channel to constrain vessel detection.

With the launch of the Italian constellation of small satellites for the Mediterranean basin observation COSMO-SkyMed and the German TerraSAR-X missions, the delivery of very high-resolution SAR data to observe the Earth day or night has remarkably increased. In particular, also taking into account other ongoing missions such as Radarsat or those no longer working such as ALOS PALSAR, ERS-SAR and ENVISAT the amount of information, at different bands, available for users interested in oil spill analysis has become highly massive. Moreover, future SAR missions such as Sentinel-1 are scheduled for launch in the very next years while additional support can be provided by Uninhabited Aerial Vehicle (UAV) SAR systems. Considering the opportunity represented by all these missions, the challenge is to find suitable and adequate image processing multi-band procedures able to fully exploit the huge amount of data available. In this paper we present a new fast, robust and effective automated approach for oil-spill monitoring starting from data collected at different bands, polarizations and spatial resolutions. A combination of Weibull Multiplicative Model (WMM), Pulse Coupled Neural Network (PCNN) and Multi-Layer Perceptron (MLP) techniques is proposed for achieving the aforementioned goals. One of the most innovative ideas is to separate the dark spot detection process into two main steps, WMM enhancement and PCNN segmentation. The complete processing chain has been applied to a data set containing C-band (ERS-SAR, ENVISAT ASAR), X-band images (Cosmo-SkyMed and TerraSAR-X) and L-band images (UAVSAR) for an overall number of more than 200 images considered.

Multidimensional Synthetic Aperture Radar (SAR) imaging is a technique based on coherent SAR data combi- nation for space (full 3-D) and space deformation-velocity (4-D) analysis. It extends SAR interferometry and differential interferometry concepts offering new options for the analysis and monitoring of ground scenes. In this paper, we consider the problem of detecting scatterers showing partial correlation properties induced by simulta- neous acquisitions from satellite formations or an uneven temporal distribution of satellite constellations. To this end, we design a decision rule accounting for the presence of partial coherent scatterers. At the analysis stage, we assess the performance of the new detector also in comparison with a previously proposed scheme, developed in the context of SAR tomography for fully coherent scatterers.

In this work, a Markov random field based phase locked loop is proposed for phase unwrapping. The neighboring pixels are used to update the phase estimate of the centering pixel. The performance of the proposed method is evaluated for both synthetic and real interferometric phase. For terrains with relatively low slopes, the phase unwrapping is done successfully. However, in case of high fringe frequency, the method fails to unwrap the whole phase gradient. Nevertheless, the noise suppression capability of phase locked loop is remarkable.

We present a preliminary simulation study of an interferometric SAR altimeter for the terrain-aided navigation application. Our simulation includes raw SAR data generation, azimuth compression, leading edge detection of the echo signal, maximum likelihood angle estimation and the Bayesian state estimation. Sour results show that radar altimeter performance can be improved with the feedback loop from the rear-end navigation part.

Synthetic Aperture Radar Interferometry (InSAR) is a technique for the generation of Digital Elevation Models (DEMs) of an observed scene. It exploits the phase difference (interferogram) of two SAR images relevant to the same area and acquired by two different look angles. To recover the topographic information from an InSAR data pair, it is necessary to evaluate a proper phase offset value to add to the unwrapped SAR interferogram. Generally, such a phase offset is accurately estimated by using Corner Reflectors (CRs) properly deployed over the illuminated area. Nevertheless, in some cases of practical interest, CRs cannot be used: in order to overcome this limit, different algorithms have been proposed in literature. In this paper, we present an algorithm aimed at estimating the InSAR phase-offset without using CRs. To this aim, we first present a theoretical analysis, validated by experiments carried out on simulated data, for the evaluation of the phase offset and, thereafter, we apply the proposed method on real data acquired by the X-band airborne OrbiSAR system. Comparisons with results achieved by using CRs properly deployed over the test site are also included.

SAR has several strong key features: fine spatial resolution/precision and high temporal pass frequency. Moreover, the InSAR technique allows the accurate detection of ground deformations. This high potential technology can be invaluable to study volcanoes: it provides important information on pre-eruption surface deformation, improving the understanding of volcanic processes and the ability to predict eruptions. As a downside, SAR measurements are influenced by artifacts such as atmospheric effects or bad topographic data. Correlation gives a measure of these interferences, quantifying the similarity of the phase of two SAR images. Different approaches exists to reduce these errors but the main concern remain the possibility to correlate images with different acquisition times: snow-covered or heavily-vegetated areas produce seasonal changes on the surface. Minimizing the time between passes partly limits decorrelation. Though, images with a short temporal baseline aren't always available and some artifacts affecting correlation are timeindependent. This work studies correlation of pairs of SAR images focusing on the influence of surface and climate conditions, especially snow coverage and temperature. Furthermore, the effects of the acquisition band on correlation are taken into account, comparing L-band and C-band images. All the chosen images cover most of the Yellowstone caldera (USA) over a span of 4 years, sampling all the seasons. Interferograms and correlation maps are generated. To isolate temporal decorrelation, pairs of images with the shortest baseline are chosen. Correlation maps are analyzed in relation to snow depth and temperature. Results obtained with ENVISAT and ERS satellites (C-band) are compared with the ones from ALOS (L-band). Results show a good performance during winter and a bad attitude towards wet snow (spring and fall). During summer both L-band and C-band maintain a good coherence with L-band performing better over vegetation.

Classical applications of the MTInSAR techniques have been carried out in the past on medium resolution data acquired by the ERS, Envisat (ENV) and Radarsat sensors. The new generation of high-resolution X-Band SAR sensors, such as TerraSAR-X (TSX) and the COSMO-SkyMed (CSK) constellation allows acquiring data with spatial resolution reaching metric/submetric values. Thanks to the finer spatial resolution with respect to C-band data, X-band InSAR applications result very promising for monitoring single man-made structures (buildings, bridges, railways and highways), as well as landslides. This is particularly relevant where C-band data show low density of coherent scatterers. Moreover, thanks again to the higher resolution, it is possible to infer reliable estimates of the displacement rates with a number of SAR scenes significantly lower than in C-band within the same time span or by using more images acquired in a narrower time span. We present examples of the application of a Persistent Scatterers Interferometry technique, namely the SPINUA algorithm, to data acquired by ENV, TSX and CSK on selected number of sites. Different cases are considered concerning monitoring of both instable slopes and infrastructure. Results are compared and commented with particular attention paid to the advantages provided by the new generation of X-band high resolution space-borne SAR sensors.

With the launch of Sentinel-1, advanced interferometric measurements will become more applicable then ever. The foreseen standard Wide Area Product (WAP), with its higher spatial and temporal resolution than comparable SAR missions, will provide the basement for the use of new wide scale and multitemporal analysis. By now the use of SAR interferometry methods with respect to risk assessment are mainly conducted for active tectonic zones, plate boundaries, volcanoes as well as urban areas, where local surface movement rates exceed the expected error and enough pixels per area contain a relatively stable phase. This study, in contrast, aims to focus on infrastructural sites that are located outside cities and are therefore surrounded by rural landscapes. The stumbling bock was given by the communication letter by the European Commission with regard to the stress tests of nuclear power plants in Europe in 2012. It is mentioned that continuously re-evaluated risk and safety assessments are necessary to guarantee highest possible security to the European citizens and environment. This is also true for other infrastructural sites, that are prone to diverse geophysical hazards. In combination with GPS and broadband seismology, multitemporal Differential Interferometric SAR approaches demonstrated great potential in contributing valuable information to surface movement phenomenas. At this stage of the project, first results of the Stamps-MTI approach (combined PSInSAR and SBAS) will be presented for the industrial area around Priolo Gargallo in South East Sicily by using ENVISAT ASAR IM mode data from 2003-2010. This area is located between the Malta Escarpment fault system and the Hyblean plateau and is prone to earthquake and tsunami risk. It features a high density of oil refineries that are directly located at the coast. The general potential of these techniques with respect to the SENTINEL-1 mission will be shown for this area and a road-map for further improvements is given in order to overcome limitations that refer to the influence of the atmosphere, orbit or DEM errors. Further steps will also include validation and tectonic modeling for risk assessment.

The paper investigates the potentialities of the COSMO/SkyMed (CSK) constellation for ground elevation measurement through conventional and multi-temporal SAR Interferometry (InSAR), with particular attention devoted to the impact of the improved spatial resolution with respect to the previous SAR sensors. The Atmospheric Phase Screen (APS) is wellknown to be the main source of errors for accurate topographic mapping through SAR interferometry, in case of monostatic sensors. Different strategies can be adopted to filter out this signal, ranging from the exploitation of the wellknown spatial and temporal statistics of the APS to the estimation of independent APS measurements through Numerical Weather Prediction (NWP) models. Their feasibility and the achievable accuracies are discussed here.

Three algorithms for the retrieval of soil moisture content (SMC) from MetOp-ASCAT acquisitions have been developed and compared. This activity has been carried out in the framework of the ASCAT Round Robin exercise that was supported by ESA as a part of the Climate Change Initiative Phase 1 Soil Moisture Project. The algorithms have been developed and tested using the ASCAT Round Robin data package (ASCAT – RRDP) that was distributed among the partners of the project for developing and validating the algorithms and that was composed by a selection of 150 test sites derived from the International Soil Moisture Network (ISMN) was used. The locations of the in situ soil moisture measurements used for the Round Robin represent a wide range of climate conditions, covering a wide variety of vegetation types and vegetation density classes.

Accurate information about soil moisture content (SMC) in mountain catchments is of great importance in hydrological applications, agriculture and climate change impact analysis. In the last two decades microwave remote sensing sensors such as Synthetic Aperture Radar (SAR) have been deeply exploited for surface SMC estimation. However, obtaining reliable predictions of fine-scale spatial and temporal patterns of SMC in mountain areas is still challenging due to the extreme variability in topography, soil and vegetation properties. In this contribution we analyze the spatial and temporal dynamic of surface SMC of alpine meadows and pastures with different techniques: (I) a network of fixed stations; (II) field campaigns with mobile ground sensors; (III) SMC retrieval from RADARSAT2 SAR images; (IV) simulations using the GEOtop 2.0 hydrological model. The strength and the weaknesses of the different estimation techniques are evaluated and the physical controls of the observed SMC patterns are analyzed. Results show that SAR SMC estimation corresponds well to the spatial ground surveys, but shows different patterns with respect to the model, especially for irrigated meadows. In fact, SAR patterns reflect vegetation, soil type and topography. Model output is in agreement with fixed stations observations, but it shows less spatial variability compared to SAR. Differences are likely due to the difficulties to know with sufficient spatial detail model parameters and irrigation amount. Therefore, results suggest that SAR products have a good ability to reproduce small-scale SMC patterns in mountain regions, thus complementing the ability of the hydrological model to predict temporal variations of SMC.

In this paper the results obtained from an experiment focused on the capabilities of GNSS-R sensors for land applications are described. This experiment was carried out within the framework of two ESA projects devoted to the investigation of GNSS-R signal over land: LEiMON(Land Monitoring with Navigation Signals) and GRASS (GNSS Reflectometry Analysis for BiomaSS Monitoring). The latter project consisted in the analysis of GNSS-R signal collected by a dual polarization sensor installed on an aircraft which flew over a test area in Italy with agricultural fields and poplar plots. It has been observed that the LR reflection coefficient was sensitive to changes in the surface soil moisture, with a total variation of about 6 dB between the dry and wet seasons, within an interval between -8/-17 dB.Whereas, the RR reflection coefficient was generally very low for all surfaces, in the range -20/-25 dB, with an increasing trend with incidence angle. LR reflection coefficient was directly related to the main parameters of soil and vegetation, namely soil moisture and vegetation biomass and rather good sensitivity to these parameters was observed. The sensitivity to soil moisture was of about 0.25dB/%soil moisture. These results have been compared with those obtained in the LEiMON project showing a good agreement. A clear correlation was also observed between LR reflection coefficient and poplar biomass, especially at steep incidence angles (17-23°). The observed sensitivity was of about 1.0dB/(50-100t/ha of dry biomass). These results have been subsequently compared with those obtained at the same frequency (L-band) with SAR sensors. From the comparison it was observed that the sensitivity of GNSS-R signal is generally lower than that one of SAR, except for the case of forest biomass. The obtained results suggest good prospects of GNSS-R especially for soil moisture and forest biomass monitoring, also in view of the increasing availability of GNSS constellations and the potential synergy with other Earth Observation sensors like SAR’s.

In the last decade there has been a considerable development of spaceborne SAR sensors. All the major space agencies are planning future SAR missions with polarimetric capabilities. However there is still a need to guide electromagnetic and statistics theories that take advantage of this kind of information towards operational applications. The use of contextual information is often required for automatic interpretation and target detection. The implementation of fast and reliable algorithms that exploit both polarimetric and contextual information can be limited by the increased dimensionality of the problem. Principal Component Analysis (PCA) is a data analysis technique that relies on a simple transformation of recorded observation, stored in a vector, to produce statistically independent variables. Non-Linear PCA is commonly seen as a non-linear generalization and extension of standard PCA. If non-linear correlations between variables exist, NLPCA will describe the data with greater accuracy and/or by fewer factors than PCA. In this work a combination of polarimetric and contextual information is performed using an Auto Associative Neural Network. A set of polarimetric input features were chosen together with contextual descriptors in order to produce an information set having lower dimensionality that can be exploited in a classification problem.

The latest generation synthetic aperture radar (SAR) systems allows providing emergency managers with near real time flood maps characterized by a very high spatial resolution. Near real time flood detection algorithms generally search for regions of low backscatter, thus assuming that floodwater appears dark in a SAR image. It is well known that this assumption is not always valid. For instance, vegetation emerging from floodwater may produce high radar returns because of the double bounce effect involving water surface and vertical stems. However, even mapping bare or scarcely vegetated inundated terrains, or crops totally submerged by water can turn out to be a difficult task. In the presence of wind that roughens the water surface, floodwater can appear bright in SAR images. Moreover, if X-band radars as TerraSAR-X or COSMO-SkyMed are used to map inundation, not only missed detection, but also false alarms may occur because of artifacts caused by heavy precipitating clouds that attenuate the radar signal. This paper proposes possible strategies to cope with flood mapping using SAR data in the presence of wind or heavy precipitation.

In Multi-Baseline SAR tomography it is necessary to process the acquired data by advanced signal processing techniques in order to adequately compensate the bad consequences of an under-sampled configuration. These techniques have to properly work on an environment characterized to have point targets, distributed targets and both of theme. This paper considers the Convex Optimization (CVX) tomographic solution in order to process multi-baseline data-sets collected in a Fourier under-sampled configuration in the above mentioned environment. The CVX and the Second Order Cone Programming Solution (SOCPs) have been tested by a generic log-barrier algorithm, through a successfully computational bottleneck Newton calculation. These techniques are validated on point targets, distributed targets and realistic forested environments.

A new technique for sub-pixel identification in Synthetic Aperture Radar (SAR) images is presented. First the sub-pixel combinations among we would like to distinguish are presented, and the theoretical basis of separation by using magnitude, polarization and phase of the return. The second part tests the separation ability by using SAR images of a sub-pixel shaped targets.

It's difficult for simple space-borne SAR to identify oil rigs effectively in single pass SAR image. One method of oil rig detection is using multi-temporal SAR images. Some papers had do some research on it. In this paper, we use interferometry SAR image to detect the oil rigs. We get the correlation coefficient image from the In-SAR data, and using the histogram of correlation coefficient image, A Constant False Alarm Rate (CFAR) algorithm is imported to detect the oil rigs. The Probabilistic Neural Network (PNN) is used to get the threshold of detection. The test result shows that sea coherent target detection in the coherent coefficient image is efficacious.

This paper investigates the ability to classify different vegetation types covering a semi-arid Mediterranean vegetated area using single-polarization SAR images from the TerraSAR-X (TSX-1) satellite. Based on statistical moments such as mean, standard deviation (STDEV), skewness and kurtosis, we found textural differences useful for the classification of the vegetation types. The research site, located near Zafit Hill, Israel, includes several different vegetation types such as pines and cypress forests with shrubs as an underlying vegetation layer (understory), olive orchards, eucalyptus clusters, natural grove areas with the presence of stones and smooth rocks, a wet cotton field, and smooth agricultural fields after harvest. In each vegetation type area, 40 equal polygons (10*10 pixels each) were identified on an optical image and defined on the TSX-1 image; 280 polygons in total were identified. The aforementioned statistical parameters were produced for each polygon, and co-variance matrices of combinations of two, three, or all four parameters together were produced. It was found that using the Mahalanobis distance of the mean-STDEV-skewness combination after applying a mode filter (5*5 in size) was the best way to classify the vegetation types in the research area.

Alluvial fans constitute important recorders of tectonic and climatic signals. Thus, determining the age of alluvial deposits is a common and pivotal component in many quantitative studies of recent tectonic activity, past climatic variations and landscape evolution processes. In this study we build on the established relation between surface age and surface roughness and examine the use of radar backscatter data as a calibrated proxy for constraining the age of alluvial surfaces in such environments. This study was conducted in the hyper-arid environment of the southern Arava rift valley north of the Gulf of Aqaba. ALOS-PALSAR L-Band dual-polarized (i.e., HH, HV) data with different incidence angles (24°, 38°) and resolutions (6.25m, 12.5m) were examined for 11 alluvial surfaces, for which surface ages ranging from 5-160 ka were previously determined. As expected, radar backscatter in such low-relief hyper-arid desert environments responded primarily to SR at pixel-scales and below. Nonetheless, measured backscatter values for single pixels were found to be unsuitable proxies for surface age because of the natural variability in SR across alluvial units of a given age. Instead, we found the statistical properties of radar pixel populations within a given unit to be the most effective proxies for surface age. Our results show that the mean backscatter value within representativeROI’s (region of interest) provided the best predictor for surface age: Lower mean backscatter values correlated well with older and smoother alluvial surfaces. The HHpolarized image with ~38° incidence angle and 6.25 m/pixel resolution allowed the best separation of surface ages. This radar-based approach allows us to quantitatively constrain the age of alluvial surfaces in the studied region at comparable uncertainty to that of “conventional” surface dating techniques commonly used.