This PDF file contains the front matter associated with SPIE Proceedings Volume 11154, including the Title Page, Copyright Information, Table of Contents, Author and Conference Committee lists.

Multi-temporal SAR interferometry (MTInSAR) techniques are able to derive displacement maps and displacement time series over coherent objects on the Earth, used for monitoring either geophysical ground deformation or structural instabilities. Nowadays, several datasets are available at different wavelengths, spatial resolutions, and revisit time, collectively covering long time periods (even more than 20 years). In particular, short revisit times, by improving the temporal sampling, make theoretically possible to catch high-rate and non-linear kinematics, which typically characterize warning signals like, for instance, those related to pre-failure of artificial infrastructures. However, MTInSAR algorithms generally fit the displacement signal by using a linear model, which is computationally convenient, and also, more robust than higher-order models for what concerns the errors affecting the InSAR phase. Moreover, the analysis of the MTInSAR products is often performed by only considering the mean displacement rate, computed over the monitoring period, which is the information typically displayed on the displacement maps. However, this approach suffer for two main limitations: 1) by analysing time series covering long time periods (as more and more common with the datasets available nowadays), it is more probable that nonlinear displacements occur; 2) the mean displacement rate, though useful for distinguishing between stable and unstable areas on the ground, often mask interesting nonlinear signals, useful for early warning. Therefore, in order to fully exploit the content of MTInSAR products, methods are needed for automatically identifying relevant changes along displacement time series, and, consequently, classifying the targets on the ground according to their kinematic regime. This also allows performing a more reliable ground deformation spatial analysis, by distinguishing among spatial patterns of different kinematics. Recently, approaches have been proposed to tackle this problem, which use different strategies. This work presents some general considerations and a possible new procedure based on statistical characterization of displacement time series, which allows, under certain constraints, recognising automatically different kinematic classes, and estimating the relevant parameters useful for target characterization in time. The theoretical formulation of the proposed method is introduced, then the algorithm performance is evaluated under simulated scenarios, by varying simulation parameters such as type of kinematic model, temporal sampling rate, number of observations, and signal to noise ratio. Finally, some results of the analysis of displacement time series derived by MTInSAR processing of real datasets are presented.

Multi-temporal SAR interferometry (MTInSAR), by providing both mean displacement maps and displacement time series over coherent objects on the Earth, allows analysing wide areas, identifying critical ground instabilities, and studying the phenomenon evolution in a long time-scale. This technique has been proved to be very useful for detecting and monitoring also slope instabilities. Nowadays, several satellite missions are available providing InSAR data at different wavelengths, spatial resolutions, and revisit time. The high-resolution X-Band COSMO-SkyMed constellation, acquires data with spatial resolution reaching metric values, and provides revisit times of up to a few days, leading to an increase in the density of the measurable targets, thus improving the monitoring of local scale events as well as the detection of non-linear displacements. The recent Sentinel-1 C-band mission from the European Space Agency (ESA), provides a spatial resolution comparable to the previous ESA missions, but a nominal revisit time reduced to 6 days. By offering regular global-scale coverage, better temporal resolution and freely available imagery, Sentinel-1 improves the performance of MTInSAR for ground displacement investigations. In particular, the short revisit time allows a better time series analysis by improving the temporal sampling and the chances to catch pre-failure signals characterised by high rate and non-linear behaviour signals. Moreover, it allows collecting large data stacks in a short time period, thus improving the MTInSAR performance in emergency (post-event) scenarios. These characteristics are very promising for early warning of slope failure events and monitoring subsequent displacements trends. In this work, we present the results obtained by using both COSMO-SkyMed and Sentinel-1 data for investigating the ground stability of hilly villages located in Southern Italian Apennine (Basilicata region). In the area of interest, several landslides occurred in the recent past (e.g. on Montescaglioso in 2013) and also very recently (e.g. on Pomarico in 2019), causing damages to houses, commercial buildings, and infrastructures. SAR datasets acquired by COSMO-SkyMed and Sentinel-1 from both ascending and descending orbits have been processed by using SPINUA MTInSAR algorithm, in order to exploit the potentials of these two satellite missions to investigate ground displacements related to the slope instabilities. Mean velocity maps and displacement time series have been analysed looking, in particular, to non-linear trends that are possibly related to relevant ground instabilities and, thanks to the high spatial resolution, useful in terms of early warning, in the case of rigid soil masses. Results are presented and discussed in relation to the well know events occurred in the area of interest.

Synthetic Aperture Radar (SAR) constitutes a fundamental asset for wide-areas monitoring with high-resolution requirements. The first SAR sensors have given rise to coarse coastal and maritime monitoring applications, including oil spill, ship and ice floes detection. With the upgrade to very high-resolution sensors in the recent years, with relatively new SAR missions such as Sentinel-1, a great deal of data providing a stronger information content has been released, enabling more refined studies on general targets features and thus permitting complex classifications, as for ship classification, which has become increasingly relevant given the growing need for coastal surveillance in commercial and military segments. In the last decade, several works focused on this topic have been presented, generally based on radiometric features processing; furthermore, in the very recent years a significant amount of research works have focused on emerging deep learning techniques, in particular on Convolutional Neural Networks (CNN). Recently Capsule Neural Networks (CapsNets) have been presented, demonstrating a notable improvement in capturing the properties of given entities, improving the use of spatial informations, in particular of spatial dependence between features, a severely lacking feature in CNNs. In fact, CNNs pooling operations have been criticized for losing spatial relations, thus special capsules, along with a new iterative routing-by-agreement mechanism, have been proposed. In this work a comparison between Capsule and CNNs potential in the ship classification application domain is shown, by leveraging the OpenSARShip, a SAR Sentinel-1 ship chips dataset; in particular, a performance comparison between capsule and various convolutional architectures is built, demonstrating better performances of CapsNet in classifying ships within a small dataset.

Compact active transponders have been recently proposed as a powerful and cost effective solution to monitor motion of critical geological areas. Continuous monitoring of the ground deformation can be pursued by the recourse to SAR interferometry; precise measurements can be obtained when time coherence persists over long times on the inspected zone at least in a few points (permanent scatterers). Sometimes the introduction of coherent radar targets can be a good solution to overcome the problem of phase instability, especially in vegetated and/or unurbanized areas. In the past, passive devices such as corner reflectors were used in such areas but their drawbacks due to their large size and weight and loss of time coherence due to geometric deformations and material degradation has limited their use. The availability of smaller, lighter, and less complex radar transponders is a recently presented as an alternative to passive devices, but also such compact and low cost devices can suffer some drawbacks, such as misplacement of the focused source on the ground due to the internal HW delay and distortion of many ground points due to the strong impulse response of the transponder that creates the typical cross pattern on the focused image. It would be desirable to have the phase precise information on the ground facilitated by the presence of compact active transponders with the possibility of switching on and off the visibility of the transponder on the acquired raw image. In this paper we propose a noise like pure transmitter to be used as a phase reference system, emitting a pseudo noise known signal at the background noise level. The presence of the transmitter can be extracted properly focusing the pseudo-noise transmitted sequence, stored on the sensor on-board memory. The correlation based range processing allows to focus the single transmitter, neglecting the influence of other ground returns, and the azimuthal chirp and range walk are preserved in the range focused transmitter signal, as they result dependent on the Doppler shift due to the SAR platform relative motion with respect to the ground. Once extracted the phase information of single point scatterer, the raw data image can be obtained by a standard focusing algorithm. In this case, the presence of the transmitter is not evident in the scene as it is seen as a noise source by the focusing algorithm. There are several problems to be addressed in the definition of the system but some simplification of the required hardware can be pursued. Preliminary simulation results of the proposed approach are presented in the paper.

A key success criterion for any long-term ESDR (Earth System Data Record) is how well it retains internal consistency upon input from multiple data sources. Our approach to achieving internal consistency of a soil moisture ESDR is based on mature data records derived from established cross-calibration procedures leveraged from the GPM mission. The X-band soil moisture estimates in our soil moisture ESDR will be derived from the GPM Level 1C Common Calibrated Brightness Temperatures (GPMXCAL, Biswas et al., 2014). This product suite contains cross-calibrated brightness temperatures from all radiometers considered in the GPM constellation using the GMI as the calibration reference. The double-difference cross-calibration procedures deployed in the GPMXCAL development have been extensively published, evaluated and refined throughout the development and operation of GPM (Berg et al., 2016). The result is a stable and consistent dataset among radiometers (GMI, AMSR2, WindSat, AMSR-E, and TMI) whose X-band cross-calibrated brightness temperatures cover a period of more than two decades (1997–present), providing the basis for producing a long-term consistent multi-satellite soil moisture ESDR. We then apply the same set of inversion algorithms, physical models, and ancillary data to the cross-calibrated brightness temperatures above. This gives us another means to reinforce ESDR consistency because for a given frequency, cross-calibrated brightness temperatures, regardless of their underlying sources, are all now subject to the same analytical inversion procedures to produce soil moisture estimates.

Global Navigation Satellite System Reflectometry (GNSS-R) is a valuable tool for the remote sensing of forest biomass, which is essential for the understanding of the hydrological and carbon cycles. The experiments carried out from 2008 to 2012 during the ESA projects GRASS and Leimon demonstrated that the GNNS-R signal is correlated to the Aboveground Biomass (AGB) with a monotonical decrease when AGB increases: the sensitivity to AGB observed during the GRASS experiment was around 1.5 dB/(100 t/ha) with a correlation coefficient of 0.91. Moreover, a saturation effect was not observed within the AGB values considered in the analysis. This does represent a major improvement with respect to conventional monostatic radars, as the backscattering coefficient at L-band is reported to saturate within the 100–150 t/ha range, depending on the type of forest. This study aims at exploiting the sensitivity of the GNSS-R signal to forest biomass and to assess the possibility of estimating globally the above parameter using data from the satellite missions TechDemoSat-1 (TDS-1), launched by Surrey Satellite Technology Ltd. in 2014, and the NASA’s Cyclone GNSS (CyGNSS), launched in 2017. For evaluating the sensitivity of the data of these two sensors to forest biomass, some test areas have been identified worldwide, as representative of the most important forest types, from boreal to equatorial forests. For these areas, two TDS parameters, namely Reflectivity and Signal Noise Ratio (SNR), have been analysed and compared with the Woody Volume (t/ha) derived from ALOS data by using the ANN inversion algorithm proposed by Santi et al. (2017). A further sensitivity analysis was carried out by comparing TDS Reflectivity and ALOS backscattering to the Vegetation Optical Depth (VOD) obtained from SMAP (Soil Moisture Active Passive) sensor for all the available areas. The same comparison has been repeated at a global scale by using CyGNSS data. In addition, TDS and CyGNSS sensitivities to AGB have been also evaluated by comparison with the improved pan-tropical biomass map (t/ha) proposed by Avitabile et al. (2016) Based on the obtained results, the potential of both TDS and CYGNSS in estimating the forest biomass has been exploited, by implementing and testing retrieval algorithms based on Artificial Neural Networks (ANN). The analysis is still in progress, however the obtained results in the retrieval of both VOD and AGB from data from TDS-1, ALOS2, and CyGNNS, are encouraging, as it has been pointed out by correlation coefficients (R) ranging between 0.81 and 0.92 and RMSE between 0.1 and 0.17 for VOD retrievals and RMSE ≃76 t/ha for AGB retrieval. References Santi E., S. Paloscia, S. Pettinato, G. Fontanelli, M. Mura, C. Zolli, F. Maselli, M. Chiesi, L. Bottai, G. Chirici, 2017, The potential of multifrequency SAR images for estimating forest biomass in Mediterranean areas, Remote Sensing of Environment, 200 (2017), pp. 63–73. Avitabile V, Herold M, Heuvelink G, Lewis SL, Phillips OL, Asner GP et al. (2016). An integrated pan-tropical biomass maps using multiple reference datasets. Global Change Biology, 22: 1406–1420. doi:10.1111/gcb.13139.

Fire has a vast influence on the climatic balance, and the Global Climate Observing System (GCOS) considers it an Essential Climate Variable (ECV). Remote sensing data is a powerful source of information for burned area detection and thus for estimating greenhouse gases (GHGs) emissions from fires. Currently, most burned area products are based on optical images. However, cloud cover independent Synthetic Aperture Radar (SAR) datasets are increasingly exploited for burned area mapping. This study assessed temporal indices based on temporal backscatter coefficient to understand their suitability for burned area detection. The analysis was carried out using the random forests machine learning classifier, which provides a rank for each independent variable used as input. Depending on land cover type, soil moisture, and topographic conditions, remarkable differences were observed between the temporal backscatter based indices.

A fundamental element of Copernicus, the EU's Earth Observation and monitoring programme, is the development and operation of an independent, dedicated and sustained space-based observation infrastructure. Six series of so-called Sentinel satellites ensure continuity until 2026-2028 (First Generation). The “CSC Expansion” programme includes the new missions that have been identified by the European Commission (EC) as priorities for implementation in the coming years. Among the new missions presently under preliminary design phase, the Copernicus L-band SAR will support key European policy objectives through the filling of observation gaps in the current Copernicus satellite constellation and will provide enhanced continuity for diverse operational services, supporting the down-stream commercial and institutional users. The mission is presently named “Radar Observing System for Europe L” (ROSE-L), to indicate its capability to work in synergy with other Copernicus SAR-based space elements operating at a different frequency (particularly Sentinel-1 operating at C-band) as an element of a System of Systems (SoS). In the frame of the work carried out to preliminarily design the system, the ESA Mission Advisory Group (MAG) helps to advice on the mission requirements. The requirements are mainly inspired by the Copernicus service needs, that are being collected by the MAG looking at previous studies, at the EC Copernicus Service evolution document, as well as at the outcome of specialised workshops in diverse application fields. Due to the longer wavelength and enhanced penetration capabilities, L-Band SAR observations from space provide additional information that cannot be gathered by other means, for instance on volumetric targets (vegetation, forest, ice, etc…) or on moisture content. ROSE-L is capable to address important measurement gaps from space, leveraging unique information on land, disasters and geohazards and the cryosphere. This includes the collection of ground motion and high-resolution soil moisture information on vegetated areas, the provision of vegetation biomass data, improved land cover. The quantification of the vegetation scattering contribution is another key factor to ensure a better coherence of the signal for interferometric applications. The unique interaction mechanisms of L-band longer wavelength will also bring additional information when exploiting multifrequency datasets, e.g., for sea ice observations. Being an element of a complex mission system, ROSE-L will support the overall continuity of the Copernicus observations, e.g., improving their accuracy, the products quality, the temporal and spatial resolution of collected data. A summary of what ROSE-L will be able to provide to the user community will be given during the conference, discussing what unique information can be gathered at L-band and which mission performance parameters (resolution and coverage) are considered necessary to fulfill the user needs.

Landslide events are triggered by heavy rain or earthquake each year around the world. The remote sensing technique is an effective tool for disaster mapping, monitoring, and early warning. In particular, synthetic aperture radar (SAR) has great potentials due to its all-weather with day and night time imaging capabilities. Thus, rapid damage detection of landslide events using SAR data is expected. However, the technique for landslide detection has not been established. In recent years, convolutional neural networks (CNNs) have been widely developed in semantic segmentation. The CNNs based on the U-Net, in particular, has a capability of fewer training images for semantic segmentation. In this study, we demonstrate a landslide detection using the CNNs based on the U-Net with SAR data. The landslides induced by the 2018 Hokkaido Eastern Iburi earthquake was selected for a case study. ALOS-2/PALSAR-2 data were acquired on 29 March 2018 and 13 September 2018 in ascending orbit and were acquired on 23 August 2018 and 6 September 2018 in descending orbit. For pre-processing, radiometric calibration and ortho-rectification were performed in each PALSAR-2 data. Firstly, we performed a conventional method, which is the backscattering coefficient difference (BCD) for landslide detection. The highest value of the F-measure (31.9%) was obtained within a 7 x 7 window size with the slope angle in the ascending orbit dataset. Secondly, we performed a CNNs method based on the U-Net for landslide detection. The highest value of the F-measure (79.9%) was obtained using the CNNs method in the ascending orbit dataset. In conclusion, the proposed CNNs method is more effective than the threshold method for landslide detection using pre- and post-event ALOS-2/PALSAR-2 data.

DInSAR technique has been long used to observe and monitor ground surface changes over large areas. Multi- temporal SAR image datasets are used for such investigations. To eliminate the issues of signal decorrelation associated with DInSAR based processing, Persistent Scatterer InSAR (PSInSAR) technique has come in a big way in the recent years where deformation studies at small scales and fine accuracy (mm-level) have been at- tempted and measured. In this paper, ground deformation of the area near Naples, Italy has been estimated for the year 2014 using a spatial correlation based PSInSAR method. Here, Cosmo-SkyMed (CSK) SLC (ascending orbit, single polarisation, stripmap Himage mode) images of Very High Resolution (VHR) were used in this study due to its capability to detect small-scale ground deformation signal over an urban area. The PSInSAR processing, used here, involves two stage selection of PS points or stable scatterers, with the coregistered SLCs and differential interferograms, using amplitude and phase analysis. The PSInSAR results were validated using time series data of two continuous GPS (cGPS) stations over the same period for the study area. The mean deformation rate over the study area was observed to be varying from -15 to +18 mm/year along line-of-sight (LOS) of Cosmo-SkyMed. Comparison between PSI derived deformation time series and cGPS measurements re- veal a good correlation with minimal discrepancies. Additionally, distinct differences were observed between the PS based LOS displacements with those obtained from cGPS based observations in case of one station compared to the other.

The main objective of this work is to estimate Snow Water Equivalent (SWE) by jointly exploiting the information derived from X-band Synthetic Aperture Radar (SAR) imagery acquired by the Italian Space Agency COSMO-SkyMed satellite constellation in StripMap HIMAGE mode and manual SWE ground measurements. The idea is to verify the sensitivity of the backscattering coefficient at X-band to the SWE and, by means of a Support Vector Regression (SVR) algorithm, to estimate the SWE for the South Tyrol region, north-eastern Italy. The regressor is trained by exploiting about 1,000 simulated backscattering coefficients corresponding to different snowpack conditions, obtained with a theoretical model based on the Dense Media Radiative Transfer theory - Quasi-crystalline approximation Mie scattering of Sticky spheres (DMRT-QMS). Then, the performance is evaluated on the backscattering values derived from COSMO-SkyMed satellite images and using the corresponding ground measurements of SWE as references. The results show a correlation coefficient equal to 0.6, a bias of 10.5 mm and a RMSE of 51.8 mm between estimated SWE values and ground measurements. The limited performance could be related to the DMRT-QMS theoretical model used for the simulations that results to be very sensitive to snow grain size and may have generated a training dataset only partially representative of satellite derived backscattering coefficients used for testing the algorithm.