This research aims to exploit the potentialities of multi-mission SAR data at X-, C- and L-band for the monitoring of snowpack and alpine soils. The snow parameters as snow water equivalent, snow liquid water content and snow metamorphism have been monitored and different methods are proposed for their retrieval. In order to gather consistent datasets, experimental activities have been conducted in two selected sites in Northern Italy, which are covered by alpine snow during winter and spring periods and are in some cases characterized by the presence of permafrost. Microwave responses of snow and soil have been then simulated by using electromagnetic (i.e., AIEM, Oh, SFT and DMRT-QCA), and physical models (SNOWPACK). Finally, machine learning approaches, as Artificial Neural Networks and Random Forest, were implemented for retrieving snow parameters; whereas interferometric techniques were used in case of snow and soil displacement as rock glaciers. Preliminary and consistent results have been obtained in terms of estimate of snow parameters and soil displacement. This multi-frequency/multi-mission approach enhances the ability of SAR sensors to monitor and analyze snow dynamics, contributing to improved decision-making in various domains.

The research results described in this paper have been obtained in the framework of the 2019–2022 ALGORITMI project between the Italian Space Agency (ASI) and the Institute of Applied Physics of the National Research Council (CNR-IFAC). The focus of the research was the development of innovative algorithms for the estimation of geophysical parameters of soil, snow, and vegetation with the aim of monitoring soil, snow cover and agricultural crop conditions. The estimation of soil moisture, vegetation biomass, snow water equivalent, and crop classification was improved by using retrieval algorithms based on machine- learning approaches and temporal series of SAR images from COSMO-SkyMed (CSK) and Sentinel-1 (S-1) missions, along with optical images from Sentinel-2. This paper provides an overview of the most recent and valuable results obtained during the project. In particular, the validation of soil moisture provided R=0.89 and RMSE=0.025 m3/m3by integrating data from S-1 and CSK and that one of snow water equivalent gave R=0.85 with RMSE=86.24 mm (CSK HIMAGE) and R=0.86 with RMSE=71.59 mm (CSK PP). Early mapping results showed an almost monotonic progression in overall accuracy over time higher than 90% by increasing the available images.

This study aims at estimating the dry snow water equivalent (SWE) by using X-band SAR data from the COSMO-SkyMed (CSK) satellite constellation. Time series of CSK acquisitions have been collected during the dry snow period in the Alto Adige test site, in the Italian Alps, during the winter seasons from 2013 to 2015 and from 2019 to 2021. The SAR data have been analyzed and compared with the in-situ measurements to understand the X-band SAR sensitivity to SWE, which has been further assessed by Dense Media Radiative Transfer (DMRT) model simulations. The sensitivity analysis provided the basis for addressing the SWE retrieval from the CSK data, by exploiting two different machine learning (ML) techniques, namely Artificial Neural Networks (ANN) and Support Vector Regression (SVR). To ensure a statistical independence of training and validation processes, the algorithms are trained and tested using SWE predictions of the fully distributed snow model AMUNDSEN as reference data and are subsequently validated on the experimental dataset. Due to its influence on the CSK estimates, the effect of forest canopy was accounted for in the analysis. Depending on the algorithm, the validation resulted in a correlation coefficient 0.78 ≤ R ≤ 0.91, and a Root Mean Square Error 55.5 mm ≤ RMSE ≤ 87.4 mm between estimated and in-situ SWE. Further analysis and validation are needed; however, the obtained results seem suggesting the Cosmo-SkyMed constellation as effective tool for the retrieval of the dry snow water equivalent in alpine areas.

The monitoring of snow conditions in Alpine areas to support water management and avalanche warning applications would require the estimate of snow parameters, such as the snow water equivalent (SWE). In this research, COSMO-SkyMed (CSK) X-band SAR data were exploited to estimate the SWE. In-situ snow measurements (depth, density, snow grain radius, temperature) collected in South Tyrol (Italy), were used to simulate the X-band backscatter with the Dense Medium Radiative Transfer (DMRT) electromagnetic model. Two SWE retrieval algorithms based on machine learning approach were implemented. The algorithms are based on Artificial Neural Networks (ANN) and Support Vector Regression (SVR) and have been trained with both experimental data and DMRT model simulations. These algorithms were applied to a selection of CSK StripMap HIMAGE HH-polarized scenes collected over the test area. The obtained results are promising and they confirm the potential of SAR data at X-band to retrieve snow parameters, although the algorithm validation should be improved in the future, with more consistent measurement dataset.

The characterization of snow conditions and the estimation of snow water equivalent (SWE) are the main goals of this paper, achieved through the exploitation of multi-frequency SAR data at both C- and X-bands from Sentinel-1 (S-1) and COSMO-SkyMed (CSK) satellites, respectively. Dry/wet snow conditions have first been assessed using C-band S-1 images. Subsequently, a sensitivity analysis was carried out by using datasets of in-situ snow measurements (i.e. snow depth, density, snow grain radius, temperature and wetness) collected in South Tyrol region, in north-eastern Italy. Simulations based on the Dense Medium Radiative Transfer (DMRT) forward electromagnetic model were considered to interpret and assess the experimental findings. Two retrieval algorithms for SWE estimation from X-band SAR data were implemented. These algorithms are based on machine learning approaches, i.e. Artificial Neural Networks (ANN) and Support Vector Regression (SVR). The training of the algorithms accounts for experimental data and DMRT model simulations and, then is applied to a selection of X-band CSK StripMap HIMAGE scenes collected over the test area. The results are promising, and pave the way for further analysis and validation to exploit the potential of SAR for snow parameter retrieval.

The present work exploits multi-frequency SAR and optical imagery in order to assess soil and vegetation water content values in agricultural areas in Italy and Argentina. Sentinel-1, Sentinel-2, ALOS-2, RADARSAT-2, and COSMO-SkyMed imagery were used in a retrieval algorithm based on Support Vector Machine (SVR) approach. Preliminary results indicate that the integration of optical and SAR data (C and L band) by the mean of machine learning techniques leads to an accurate retrieval of soil moisture (RMSE = 4%). Moreover, a preliminary test for the retrieval of Vegetation Water Content (VWC) indicates that ALOS-2 L-Band backscattering, combined with L-Band simulated backscattering, leads to a reliable model to compute VWC, without any optical information.

This research aimed at exploiting the joint use of machine learning and polarimetry for improving the retrieval of surface soil moisture (SMC) from SAR acquisitions at C- and X-band.The study was conducted on an alpine test area in Italy and two agricultural areas in Canada, for which series of Radarsat-2 (RS2) and COSMO-SkyMed (CSK) images were available along with direct measurements of SMC from in-situ stations. The analysis confirmed the sensitivity of SAR backscattering (σ°) from both sensors to the SMC variations, with similar correlations (R ≃0.5). The comparison of SMC with the Compact Polarimetric (CP) parameters, computed from the RS2 acquisitions by Radarsat Constellation Mission (RCM) data simulator pointed out that the right and left polarized signals and the Shannon entropy intensity also have some sensitivity to SMC variations, with R ≃0.4 for all the three parameters.Based on these results, two different machine learning (ML) algorithms, namely Support Vector Regression (SVR) and Artificial Neural Network (ANN) have been implemented and tested on the available data. On the South Tyrol test area, both SVR and ANN tested with different combinations of RS2 and CSK data were able to retrieve SMC with a RMSE between 4% and 6% of SMC and R between 0.78 and 0.88, depending on the combination of inputs. The ANN algorithm based on CP data was tested on the Canada areas, being able to estimate SMC with a RMSE between 2% and 5% of SMC and R between 0.85 and 0.96.

This research aims at exploiting the integration of C- and X-band SAR data for the monitoring of Soil Moisture Content (SMC). Time series of Radarsat2 (RS2) and COSMO-SkyMed (CSK) images are collected on two test areas, located in Italy and in Canada. The backscattering sensitivity to SMC measured by in-situ stations is investigated considering the available sensor frequencies and polarizations. In addition, for exploiting the potential of fully polarimetric acquisitions of RS2, simulated Compact Polarimetric (CP) data are computed by using a Radarsat Constellation Mission (RCM) data simulator, and their sensitivity to the target SMC is examined. Based on the experimental findings, two machine learning (ML) approaches to the SMC retrieval, namely Support Vector Regression (SVR) and Artificial Neural Network (ANN) are implemented and tested on the two areas. Looking at the preliminary results, the integration of X- and C-band images does provide valuable information for the retrieval of SMC, while the simulated CP parameters exhibit a certain sensitivity to SMC. On the South Tyrol test area, both SVR and ANN tested with different combinations of RS2 and CSK data were able to retrieve SMC with a RMSE between 2% and 4% of SMC and correlation coefficient R between 0.85 and 0.97, depending on the combination of inputs. The application of the ML algorithms to the other available images on the Mazia test area and the implementation of the ML retrieval algorithm using CP data are still under investigation.

This research aims at analyzing the integration of C and X band data collected from Radarsat2 (RS2) and COSMO-SkyMed (CSK) systems on two Italian test areas, located in South-Tyrol and in Tuscany, close to Florence, to estimate soil moisture (SMC, in %) and vegetation biomass (PWC, in kg/m2). Two retrieval approaches based on Support Vector Regression (SVR) and Artificial Neural Network (ANN) have been applied to these areas. Looking at the preliminary results, it has been noted that the integration of X and C band images could provide valuable information for the retrieval of SMC, even though further investigations should be carried out on a larger time-series and set of samples. On the South Tyrol test area, SVR methods provided an accuracy in the estimate of SMC with determination coefficient, R2> 0.85 and root mean square error, RMSE2from 0.35 to 0.8 and RMSE= from 6 to 2 (% SMC), depending on the polarization combinations considered as input. X band allowed instead retrieving the PWC of cereal fields with a satisfactory accuracy (R2=0.94 and RMSE=0.35 Kg/m2).

With this work, we present a method for the detection of alpine permafrost surface deformations by using DInSAR (Differential SAR Interferometry) technique, integrating RADARSAT-2 and COSMO-SkyMed data through Support Vector Machine (SVM). On our test dataset, the combination of the two sensors produces an increase of classification accuracy equal to 8.5% with respect to the case in which only one sensor is employed, leading to an overall accuracy equal to 86.9%. We are also showing here how COSMO-SkyMed data time series acquired in the snow-free period are well suited to estimate surface deformations on some particular alpine rock glaciers. The velocities estimated with the SBAS (Small BAseline Subset) algorithm are well correlated with velocity measurements obtained by means of a ground based total station, showing a root mean squared error (RMSE) and R square value equal to 3.5 cm and to 0.67 respectively.

This research aims at analyzing the integration of C and X band data collected from Radarsat2 (RS2) and COSMO-SkyMed (CSK) systems on some test areas in Italy, in order to estimate the main geophysical parameters of soil and vegetation, such as soil moisture and vegetation biomass. A check of the sensitivity of SAR signal to the soil parameters was first carried out on both test sites. Over the South-Tyrol area a 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 carried out. From these preliminary results it can be concluded that X-band images combined with C-band images could provide valuable information for the retrieval of SMC, even though further investigations should be carried out on a larger time-series and larger set of samples.

In mountain areas, soil moisture is a key parameter for both agricultural management and natural hazard support. This paper presents an approach for retrieval of soil moisture content (SMC) from different satellite sensors with a specific focus on mountain areas. The experimental analysis was carried out on images acquired over the Südtirol/Alto Adige Province (Italy) during 2010-2011 from the RADARSAT2 in quad-pol mode and Envisat ASAR in Wide Swath mode in VV polarization. The methodology for soil moisture retrieval is based on the Support Vector Regression (SVR) method specifically trained to be able to consider topographic effects of the mountain areas. The comparison with ground measurements collected during field campaigns indicates an RMSE value of around 5% of SMC% while the comparison with fixed ground stations reports an error of around 9% of SMC%. Comparing RADARSAT2 and ASAR SMC, both datasets reveal very similar distributions of SMC values. The cumulative histogram curve for the two datasets shows a slight underestimation of SMC in the ASAR product. This could be ascribed to the reduced resolution of ASAR WS and the use of VV polarization.

This paper presents the results concerning the application of COSMO-SkyMed SAR images to mountain areas for the challenging retrieval of some key parameters of the hydrological cycle: soil moisture content (SMC), snow depth (SD), and snow water equivalent (SWE). The results obtained so far are encouraging. Regarding SMC, the results indicate that, in spite of the low penetration capability of X-band wavelength, the SAR signal is well related to soil moisture variations, even in presence of vegetation. As far as snow cover is concerned, from the analysis of data it has been observed that the backscattering coefficient remains almost constant until the SD of dry snow accumulated on soil is higher than 50-60 cm and increases rapidly as SD rises up to 150 cm. The use of an inversion algorithm for the retrieval of SWE, based on Artificial Neural Networks, showed a determination coefficient higher than 0.8, with an associated probability value of 95%.

The atmospheric phase screen is one of the main error sources that affect the precise phase measurements in many fields of earth remote sensing. The atmospheric effects can be mitigated if a precise knowledge of the power spectral density of the process is available and if same sample observations can be retrieved on a sparse grid. At smaller scales, where interactions are no longer isotropic, the behaviour is not easily predicted by the ultimate dissipative behaviour of turbulence eddies. We start by assuming that the propagation of the electromagnetic wave in the lower atmosphere can be represented by a ray travelling in a layered medium where the refractive index is constant along each layer. In a turbulent atmosphere, the interaction among eddies induces a diffusion process that propagates with different rates in both vertical and horizontal direction with the final effect of ruling the number of effective layers in the atmosphere. In this way, the overall path travelled by the electromagnetic wave is governed by the accumulated number of such effective layers whose interactions play a primary role in our model. A good model for the interactions among different layers is the linear interaction model. The power spectrum of the process can be found by solving a differential equation with given initial conditions. It can be demonstrated that an asymptotic power law decay is found under binomial competitive interactions and that, at a smaller scale, the behaviour observed in the observed data is naturally induced by the interaction process itself. The model predictions have been tested using samples of the atmospheric phase screen extracted from Synthetic Aperture Radar interferograms. To this end, the model parameters have been estimated from the data set and the predicted spectrum has been compared with the measured one.

Segmentation algorithms are often used in many image processing applications like compression, restoration, content extraction, and classification. In particular as for content extraction works carried out in the past decade have demonstrated that multi-frequency fully polarimetric SAR observations are particularly interesting, thanks to physical properties of the backscattered signal at various frequencies and polarizations. To achieve a good classification, the main difficulty is that SAR images are often embedded in heavy speckle. Segmentation of multi/hyperspectral (optical) imagery is obtained by means of algorithms based on image models, which exploit the spatial dependencies of land-covers. Unfortunately, speckle noise hides such spatial dependencies in observed SAR data. With the aim of investigating on a content extraction algorithm capable of discriminating cover classes present in the observed SAR image, heterogeneity features are used here to emphasize spatial dependencies in the data. Thus, observed pixel values are mapped into features, that take "similar" values on "similar" textures. This allows for using the same procedure of the optical case. Obviously, homogeneity/heterogeneity feature and segmentation quality are fundamental for classification accuracy. Here, the problem is tackled through the joint use of information theoretic SAR features and of a segmentation algorithm based on Markov Random Fields (MRFs).

Segmentation algorithms are often used in many image processing applications like compression, restoration, content extraction, and classification. In particular as for the content extraction, works carried out in the past decade have demonstrated that multi-frequency fully polarimetric SAR observations content are particularly interesting, thanks to physical properties of the backscattered signal at various frequencies and polarizations. To achieve a good classification, the main difficulty is that SAR images are often embedded in heavy speckle. Segmentation of multi/hyperspectral (optical) imagery is obtained by means of algorithms based on image models, which exploit the spatial dependencies of landcovers. Unfortunately, speckle noise hides such spatial dependencies in observed SAR data. With the aim of investigating on a content extraction algorithm capable of discriminating cover classes present in the observed SAR image, homogeneity/heterogeneity features are used here to emphasize spatial dependencies in the data. Thus, observed pixel values are mapped into features, that take "similar" values on "similar" textures. This allows for using the same procedure of the optical case. Obviously, homogeneity/heterogeneity feature and segmentation quality are fundamental for classification accuracy. Here, the problem is tackled through the joint use of information-theoretic SAR features and of a segmentation algorithm based on Markov Random Fields (MRFs).

In this paper the application of a transform coding technique, based on overcomplete independent component analysis (ICA), for the compression of single look intensity synthetic aperture radar (SAR) images is explored. The method has the advantage of representing the image through almost statistically independent coefficients, with an assigned distribution, so that a scalar entropy constrained quantizer, optimized for the coefficients statistics, can be used. Numerical results on ERS-1 data are presented.

The forest cover classification is extremely important for land use planning and management. In this framework, the application of pixel based classifications of middle resolution images is well assessed while the usefulness of segmentation processes and object classification is still improving. In this paper, a method based on tree-structured Markov random field (TS-MRF) is applied to Landsat TM images in order to assess the capability of the TS-MRF segmentation algorithm for discriminating forest-non forest covers in a test area located in the Eastern Italian Alps of Trentino. In particular, the regions of interest are selected from the image using a two step process based on a segmentation algorithm and an analysis process. The segmentation is achieved applying a MRF a-prior model, which takes into account the spatial dependencies in the image, and the TS-MRF optimisation algorithm which segments recursively the image in smaller regions using a binary tree structure. The analysis process links to each object identified by the segmentation a set of features related to the geometry (like shape, smoothness, etc.), to the spectral signature and to the neighbour regions (contextual features). These features were used in this study for classifying each object as forest or non-forest thought a simple supervised classification algorithm based on a thresholds built on the feature values obtained from a set of training objects. This method already allowed the detection of the forest area within the study area with an accuracy of 90%, while better performances could be achieved using more sophisticated classification algorithm, like Neural Networks and Support Vector Machine.