SAR Observations Research Articles

Retrospect on the Ground Deformation Process and Potential Triggering Mechanism of the Traditional Steel Production Base in Laiwu with ALOS PALSAR and Sentinel-1 SAR Sensors.

The realization of a harmonious relationship between the natural environment and economic development has always been the unremitting pursuit of traditional mineral resource-based cities. With rich reserves of iron and coal ore resources, Laiwu has become an important steel production base in Shandong Province in China, after several decades of industrial development. However, some serious environmental problems have occurred with the quick development of local steel industries, with ground subsidence and consequent secondary disasters as the most representative ones. To better evaluate possible ground collapse risk, comprehensive approaches incorporating the common deformation monitoring with small-baseline subset (SBAS)-synthetic aperture radar interferometry (InSAR) technique, environmental factors analysis, and risk evaluation are designed here with ALOS PALSAR and Sentinel-1 SAR observations. A retrospect on the ground deformation process indicates that ground deformation has largely decreased by around 51.57% in area but increased on average by around -5.4 mm/year in magnitude over the observation period of Sentinel-1 (30 July 2015 to 22 August 2022), compared to that of ALOS PALSAR (17 January 2007 to 28 October 2010). To better reveal the potential triggering mechanism, environmental factors are also utilized and conjointly analyzed with the ground deformation time series. These analysis results indicate that the ground deformation signals are highly correlated with human industrial activities, such underground mining, and the operation of manual infrastructures (landfill, tailing pond, and so on). In addition, the evaluation demonstrates that the area with potential collapse risk (levels of medium, high, and extremely high) occupies around 8.19 km2, approximately 0.86% of the whole study region. This study sheds a bright light on the safety guarantee for the industrial operation and the ecologically friendly urban development of traditional steel production industrial cities in China.

Advantages of High-Temporal L-Band SAR Observations for Estimating Active Landslide Dynamics: A Case Study of the Kounai Landslide in Sobetsu Town, Hokkaido, Japan

Estimating landslide dynamics is vital for the prevention of landslide disasters. Differential interferometric synthetic aperture radar (DInSAR) based on L-band SAR satellites is an effective tool for estimating the dynamics of forested landslides that occur in Japan. High-temporal L-band SAR observations have been planned for the future. Thus, it is necessary to further investigate the specific advantages of high-temporal L-band SAR observations for estimating landslide dynamics. In this study, we used DInSAR data with different time windows to identify active landslides in Hokkaido, Japan. This study is the first attempt to demonstrate the advantages of high-temporal L-band SAR observations for estimating active landslide dynamics. We successfully observed the dynamics of two active landslides, Kounai-1 and Kounai-2, using DInSAR over a time window of 14 days. We present the first spatial observation of the dynamics of Kounai-1 and Kounai-2. In addition, we discuss the dynamics of Kounai-1 and Kounai-2 based on interferograms, and our results suggest that both landslides are subunits of the same landslide, called the Kounai landslide. These results indicate that high-temporal L-band SAR observations can mitigate cycle slips and enable the estimation of active landslide dynamics.

Integrating multi-source remote sensing data for mapping boreal forest canopy height and species in interior Alaska in support of radar modeling

Vegetation information is essential for analyzing aboveground biomass and understanding subsurface characteristics, such as root biomass, soil organic matter, and soil moisture conditions. In this study, we mapped boreal forest canopy height (FCH) and forest species (FS) distributions in the Delta Junction region of interior Alaska, by integrating multi-source remote sensing observations within a machine learning framework based on the extreme gradient boosting technique. Model inputs included multi-frequency (C-/L-/P-band) SAR observations from Sentinel-1, UAVSAR (Uninhabited Aerial Vehicle SAR) and AirMOSS (Airborne Microwave Observatory of Subcanopy and Subsurface), and Sentinel-2 optical reflectance data. LVIS (Land Vegetation and Ice Sensor) LiDAR measurements (RH98) and Tanana Valley State Forest timber inventory data were used as respective canopy height and species ground truth data. The combination of multi-source datasets produced the best model performance (RMSE 1.62 m for FCH, and 84.27% overall FS classification accuracy) over other models developed from single source observations. The resulting FCH and FS maps using multi-source datasets were derived at 30 m spatial resolution and showed favorable agreement with plot level field measurements from the Forest Inventory and Analysis record. The model results also captured characteristic differences in stand structure between dominant species and from post-fire vegetation succession. Our results show the potential of multi-source remote sensing observations, including low frequency microwave sensors, for monitoring boreal forest complexity and changes due to global warming.

Integrated PSInSAR and GNSS for 3D Displacement in the Wudongde Area

The major limitation of persistent scatterer interferometric synthetic aperture radar (PSInSAR) is that it detects only one- or two-dimensional displacements, such as those in the line of sight (LOS) and azimuth directions, by repeat-pass SAR observations. Three-dimensional (3D) displacement reflects the actual sliding surface and failure mechanism of a slope. To transform LOS deformation into a reliable 3D displacement, a new approach for obtaining the 3D displacement is proposed herein based on the slope deformation (Dslope). First, the deformation value calculated using the Global Navigation Satellite System (GNSS) as a constraint is used to eliminate the residual deformation of PSInSAR. Then, Dslope is obtained from the relationship between DLOS and the slope angle extracted from the digital elevation model (DEM). Finally, according to the geometric relationship between Dslope and DLOS, a novel approach for calculating 3D displacement is proposed. When comparing the 3D displacement extracted by the proposed method and that from GNSS data in Jinpingzi landslide, the root-mean-square error (RMSE) values were ±2.0 mm, ±2.8 mm, and ±2.6 mm in the vertical, north, and east directions, respectively. The proposed method shows high accuracy in 3D displacement calculation, which can help to determine the failure mechanism of a landslide. This method can be widely used in landslide monitoring in wide areas.

Monitoring Yanwan deep-seated toppling deformation with the impact of water-level fluctuation by SAR observations

Monitoring Yanwan deep-seated toppling deformation with the impact of water-level fluctuation by SAR observations

Exploring optimal combinations of multi-frequency polarimetric SAR observations to estimate forest above-ground biomass

ABSTRACT The penetration capability of electromagnetic wave signal into forest increases with increasing wavelength. SAR data at each frequency senses different components of forest structure. Therefore the biomass distributed at various tree components could be estimated using different radar frequencies. Additionally, the synthesis of multiple SAR frequencies could improve the accuracy in retrieving forest above-ground biomass (AGB). Taking advantage of available X-, C-, L-, and P-band quad-polarimetric SAR images of airborne or spaceborne for the test site located at Genhe national forest scientific field station, we used a Genetic Algorithm and Support Vector Regression optimization algorithm (GA-SVR) to explore the sensitivity of polarimetric observations at various frequencies to forest AGB and effectiveness of AGB retrievals using single-frequency, dual-frequency, triple-frequency, and quad-frequency SAR observation combinations. We found that: (i) Most of the polarimetric observations are sensitive to forest AGB, (ii) GA-SVR performed well in forest AGB retrieval using the single frequency SAR observations or combinations of multi-frequency observations; the highest Acc. value for single-frequency-retrieved results is 75.13% acquired at P-band, with multi-frequency, the highest Acc. values is 77.34% acquired by combining C- and P-band. (iii) For forest AGB retrievals, the single-frequency P-band accuracy is comparable to the combined C- and P-band retrieval accuracy, indicating that the long-wavelength single-frequency P-band is sufficient for an accurate forest AGB retrieval. The findings reconfirmed potential of P-band for forest AGB retrievals, they also demonstrated that the optimal combination of multi-frequency SAR for AGB retrievals is by using a short-wavelength (X/C-) and a long-wavelength (L/P-).

A Novel Multidimensional Perspective on Dynamic Characteristics of the Peculiar Feather-Shaped Dunes in Kumtag Desert With Time-Series Optical and SAR Observations

A Novel Multidimensional Perspective on Dynamic Characteristics of the Peculiar Feather-Shaped Dunes in Kumtag Desert With Time-Series Optical and SAR Observations

Visibility Metric for Planning SAR Observations in Challenging Terrain

Visibility Metric for Planning SAR Observations in Challenging Terrain

Study on InSAR image fusion for improved visualization of active landslides in alpine valley areas: A case in the Batang Region, China

InSAR observation with a wide range and high sensitivity is one of the main techniques for active landslide identification, but due to the influence of the SAR satellite side-view imaging method and observation angle, single-track images in alpine valley areas facing geometric distortion, such as shadowing, layover, and foreshortening, which leads to ineffective observation in some areas. Combining the ascending and descending observation results is expected to compensate for this deficiency to a large extent. This work constructs a pixel-level image fusion method to achieve the fusion of InSAR ascending and descending orbit deformation results, consisting of three algorithms: the DMPI (decision matrix-guided pixel-level InSAR image fusion algorithm), the DMFS (deformable median filter based on decision feedback), and the QSDS (superpixel-based dual-space InSAR image fusion quality evaluation algorithm). To verify the effect, the Batang section of the Jinsha River with high mountain canyon terrain and frequent active landslides is selected. The results show that using the DMPI fused InSAR ascending and descending orbit images can improve spatial perception by enhancing the texture gradient value of remote sensing images by a maximum of two orders of magnitude, compared with the single-track InSAR observation images. Meanwhile, the number of landslides identified increased by 31.5%, and the total area increased by 50.9% on average, improving the accuracy of target detection and identification. In noise reduction, DMFS improves fusion image readability and feature detail by removing SPN (salt and pepper noise) while preserving target edge detail. Furthermore, the QSDS which can evaluate the quality of InSAR fusion images from the perspective of human eye vision, provides an adequate basis for determining the interpretability of InSAR images. This fusion method can partially improve the question of incomplete ground observation due to radar line of sight occlusion in the high mountain canyon area, thus better presenting the InSAR observation results and helping to comprehensively and conveniently interpret the InSAR deformers.

Two Mw ≥ 6.5 Earthquakes in Central Pamir Constrained by Satellite SAR Observations

The Pamir, situated in central Asia, is a result of the ongoing northward advance of the Indian continent, leading to compression of the Asian landmass. While geodetic and seismic data typically indicate that the most significant deformation in Pamir is along its northern boundary, an Mw 7.2 earthquake on 7 December 2015 and an Mw 6.8 earthquake on 23 February 2023 have occurred in the remote interior of Pamir. These two Mw ≥ 6.5 earthquakes, with good observations of satellite synthetic aperture radar data, provide a rare opportunity to gain insights into rupture mechanics and deformation patterns in this challenging-to-reach region. Here, we utilize spaceborne synthetic aperture radar data to determine the seismogenic faults and finite slip models for these two earthquakes. Our results reveal that the 2015 earthquake ruptured a ~88 km long, left-lateral strike-slip fault that dips to northwest. The rupture of the 2015 earthquake extended to the ground surface over a length of ~50 km with a maximum slip of ~3.5 m. In contrast, the 2023 earthquake did not rupture the ground surface, with a maximum slip of ~2.2 m estimated at a depth of ~9 km. Notably, the seismogenic fault of the 2015 earthquake does not align with the primary strand of the Sarez–Karakul fault system (SKFS), and the 2023 earthquake occurred on a previously unmapped fault. The well-determined seismogenic faults for the 2015 and 2023 earthquakes, along with the SKFS and other distributed faults in the region, suggest the existence of a wide shear zone extending from south to north within the central Pamir.

Failure process and three-dimensional motions of mining-induced Jianshanying landslide in China observed by optical, LiDAR and SAR datasets

ABSTRACT The occurrence of collapses and landslides due to underground mining has its unique failure mechanism, especially in the Karst mountainous regions of China. Spaceborne and airborne remote sensing observations provide rapid and effective tools for assessing surface changes and monitoring surface deformation of such landslides. In this study, we take the Jianshanying landslide, a typical mining-induced and fast-deformed landslide, as an example, and reveal the failure mechanism of such landslide by investigating the historical surface displacement. First, the complete evolution of the landslide surface was investigated from its original state to the overall sliding. The data include the satellite and Unmanned Aerial Vehicle (UAV) optical images, UAV three-dimensional (3-D) real scene models, high-resolution Light Detection and Ranging (LiDAR) DEM, and field survey. The results show that the head region entered the high deformation stage after 2019, the maximum deformation rate was 12.3 m/yr. The landslide morphology was formed after the overall slide occurred in September 2020. Then, the pre-event 3-D surface deformation after the landslide entered the high deformation stage was recovered using Interferometric Synthetic Aperture Radar (InSAR), differential DEM, and SAR/optical offset-tracking techniques. The vertical deformation was recovered around −30 m from 2019 to 2020. In particular, we solved the problem of unequal accuracy of SAR and optical offset-tracking observations in 3-D deformation inversion by employing the Helmert variance component estimation method. The maximum deformation was 6 m and 3 m within 4 months in the NS and EW directions, respectively. Finally, we revealed the failure mechanism of the Jianshanying landslide based on the disparity of horizontal and vertical deformation. That is, underground mining causes a significant subsidence of the rear part of the landslide body, resulting in different stress changes in the rear and front parts of the landslide body, which eventually led to sliding of the front part of the slope along the free surface. This work investigates and monitors the typical underground mining-induced Jianshanying landslide by using multi-sensor remote sensing approaches to trace the pre-event surface motions and to reveal its failure mechanism.

InSAR time series and LSTM model to support early warning detection tools of ground instabilities: mining site case studies

Early alarm systems can activate vital precautions for saving lives and the economy threatened by natural hazards and human activities. Interferometric synthetic aperture radar (InSAR) products generate valuable ground motion data with high spatial and temporal resolutions. Integrating the InSAR products and forecasting models make possible to set up early alarm systems to monitor vulnerable areas. This study proposes a technical support to early warning detection tools of ground instabilities using machine learning and InSAR time series that is capable of forecasting regions affected by potential collapses. A long short-term memory (LSTM) model is tailored to predict ground movements in three forecast ranges (i.e., SAR observations): 3, 4, and 5 multistep. A contribution of the proposed strategy is utilizing adjacent time series to decrease the possibility of falsely detecting safe regions as significant movements. The proposed tool offers ground motion-based outcomes that can be interpreted and utilized by experts to activate early alarms to reduce the consequences of possible failures in vulnerable infrastructures, such as mining areas. Three case studies in Spain, Brazil, and Australia, where fatal incidents happened, are analyzed by the proposed early alert detector to illustrate the impact of chosen temporal and spatial ranges. Since most early alarm systems are site dependent, we propose a general tool to be interpreted by experts for activating reliable alarms. The results show that the proposed tool can identify potential regions before collapse in all case studies. In addition, the tool can suggest an optimum selection of InSAR temporal (i.e., number of images) and spatial (i.e., adjacent measurement points) combinations based on the available SAR images and the characteristics of the study area.

Effect of Linker Elongation on the VGSC Affinity and Anticonvulsant Activity among 4-Alkyl-5-aryl-1,2,4-triazole-3-thione Derivatives

The main aim of the current project was to investigate the effect of the linker size in 4-alkyl-5-aryl-1,2,4-triazole-3-thione derivatives, known as a group of antiepileptic drug candidates, on their affinity towards voltage-gated sodium channels (VGSCs). The rationale of the study was based both on the SAR observations and docking simulations of the interactions between the designed ligands and the binding site of human VGSC. HYDE docking scores, which describe hydrogen bonding, desolvation, and hydrophobic effects, obtained for 5-[(3-chlorophenyl)ethyl]-4-butyl/hexyl-1,2,4-triazole-3-thiones, justified their beneficial sodium channel blocking activity. The results of docking simulations were verified using a radioligand binding assay with [3H]batrachotoxin. Unexpectedly, although the investigated triazole-based compounds acted as VGSC ligands, their affinities were lower than those of the respective analogs containing shorter alkyl linkers. Since numerous sodium channel blockers are recognized as antiepileptic agents, the obtained 1,2,4-triazole derivatives were examined for antiepileptic potential using an experimental model of tonic–clonic seizures in mice. Median effective doses (ED50) of the compounds examined in MES test reached 96.6 ± 14.8 mg/kg, while their median toxic doses (TD50), obtained in the rotarod test, were even as high as 710.5 ± 47.4 mg/kg.

Friction Velocity and Aerodynamic Drag Coefficient Retrieval from Sentinel-1 IW Cross-Polarization C-SAR Images under Hurricane Conditions

This study presents an approach for friction velocity and aerodynamic drag coefficient retrieval utilizing C-band VH SAR observations from Sentinel-1. The dataset contained 14 SAR images collected under six hurricane scenes co-analyzed with stepped frequency microwave radiometer (SFMR) measurements. The basis for creating this approach utilizes the results proposed earlier linking the parameters of the atmospheric boundary layer from GPS-dropsondes data to the ocean surface emissivity from SFMR measurements. The obtained dependencies of the ocean surface emissivity on surface friction velocity, aerodynamic drag coefficient, and surface wind speed are analyzed together with the collocated SAR data leading to the new GMF valid for the retrieval of friction velocities ranging from 0.55–1.56 m/s and drag coefficient values ranging from 0.00076–0.00232 for all sub swaths. Within the framework of the proposed approach, dependences of the normalized radar cross-section on the surface wind speed were also obtained and used for comparison with existing GMFs to show that the proposed approach is valid. A good consistency was obtained when comparing our results with H14E and MS1A. As an example the distributions of friction velocity, drag coefficient, and surface wind speed retrieved from the Hurricane Maria SAR image (23 September 2017) were considered.

Target Scattering Feature Extraction Based on Parametric Model Using Multi-Aspect SAR Data

The multi-aspect SAR observation can obtain the backscattering information of the illuminated scene target. There are targets with different structures in the scene, and their backscattering responses are also different. Using backscattering amplitude information to analyze the differences between targets is a conventional method. For point target types, one-dimensional backscattering curves can be used to analyze scattering characteristics, but it is difficult to analyze the overall structure of the target. Therefore, it is necessary to perform statistical analysis on the backscattering information combinating with the multi-aspect target area, and establish parameters to model the target area. In this paper, the algorithm uses the G0 distribution based on expectation maximization (EM) to fit the target area of the SAR scene. For different target types in the scene, the β and σ parameters obtained by the model combined with the backscattering amplitude information are used to perform the target. The results show that full-target in multi-aspect SAR image can be differentiated by two parameters. The scattering of partial-target slices can be characterized using two parameters (amplitude difference from surrounding points, scattering energy). The parametric model quantitatively characterizes the scattering feature of the target level, and the parameters changing corresponds to the change of the target image feature. C-band circular SAR data is used to validate our method. The experimental results give the parameter representation with sampling window based on the analysis of the target scattering, and give parameter estimates to characterize the partial target scattering.

InSAR-Based Early Warning Monitoring Framework to Assess Aquifer Deterioration

Aquifer surveillance is key to understanding the dynamics of groundwater reservoirs. Attention should be focused on developing strategies to monitor and mitigate the adverse consequences of overexploitation. In this context, ground surface deformation monitoring allows us to estimate the spatial and temporal distribution of groundwater levels, determine the recharge times of the aquifers, and calibrate the hydrological models. This study proposes a methodology for implementing advanced multitemporal differential interferometry (InSAR) techniques for water withdrawal surveillance and early warning assessment. For this, large open-access images were used, a total of 145 SAR images from the Sentinel 1 C-band satellite provided by the Copernicus mission of the European Space Agency. InSAR processing was carried out with an algorithm based on parallel computing technology implemented in cloud infrastructure, optimizing complex workflows and processing times. The surveillance period records 6-years of satellite observation from September 2016 to December 2021 over the city of Chillan (Chile), an area exposed to urban development and intensive agriculture, where ~80 wells are located. The groundwater flow path spans from the Andes Mountain range to the Pacific Ocean, crossing the Itata river basin in the Chilean central valley. InSAR validation measurements were carried out by comparing the results with the values of continuous GNSS stations available in the area of interest. The performance analysis is based on spatial analysis, time series, meteorological stations data, and static level measurements, as well as hydrogeological structure. The results indicate seasonal variations in winter and summer, which corresponds to the recovery and drawdown periods with velocities > −10 mm/year, and an aquifer deterioration trend of up to 60 mm registered in the satellite SAR observation period. Our results show an efficient tool to monitor aquifer conditions, including irreversible consolidation and storage capacity loss, allowing timely decision making to avoid harmful exploitation.

Irrigation Timing Retrieval at the Plot Scale Using Surface Soil Moisture Derived from Sentinel Time Series in Europe

The difficulty of calculating the daily water budget of irrigated fields is often due to the uncertainty surrounding irrigation amounts and timing. The automated detection of irrigation events has the potential to greatly simplify this process, and the combination of high-resolution SAR (Sentinel-1) and optical satellite observations (Sentinel-2) makes the detection of irrigation events feasible through the use of a surface soil moisture (SSM) product. The motivation behind this study is to utilize a large irrigation dataset (collected during the ESA Irrigation + project over five sites in three countries over three years) to analyze the performance of an established algorithm and to test potential improvements. The study’s main findings are (1) the scores decrease with SSM observation frequency; (2) scores decrease as irrigation frequency increases, which was supported by better scores in France (more sprinkler irrigation) than in Germany (more localized irrigation); (3) replacing the original SSM model with the force-restore model resulted in an improvement of about 6% in the F-score and narrowed the error on cumulative seasonal irrigation; (4) the Sentinel-1 configuration (incidence angle, trajectory) did not show a significant impact on the retrieval of irrigation, which supposes that the SSM is not affected by these changes. Other aspects did not allow a definitive conclusion on the irrigation retrieval algorithm: (1) the lower scores obtained with small NDVI compared to large NDVI were counter-intuitive but may have been due to the larger number of irrigation events during high vegetation periods; (2) merging different runs and interpolating all SSM data for one run produced comparable F-scores, but the estimated cumulative sum of irrigation was around −20% lower compared to the reference dataset in the best cases.

A distributed land cover classification of FP and CP SAR observation using MapReduce-based multi-layer perceptron algorithm over the Mumbai mangrove region of India

ABSTRACT Globally, the rapid loss of mangrove forests often creates long-lasting environmental damage alongside coastlines, mudflats, and river banks. All-weather, physical monitoring of such areas is almost impossible because of inaccessibility to swampy areas and a hostile substrate. Subsequently, conventional field surveys are relatively unavailable to monitor human encroachment in coastal areas like the Mumbai mangrove region of India. In this context, the polarimetric synthetic aperture radar (PolSAR) remote sensing tool becomes a potential candidate for mangrove conservation and management. Since the Mumbai mangrove region of India exists over the extensive land cover, the large-scale classification can articulate the continuous encroachment because of the location of human settlements or activities across this metropolitan coastal city. For this, the traditional algorithms need the essential improvisation to apply for large-scale data analysis, aiming simultaneously at getting the highest decisive and time efficiency. Here, we introduce a shallow learning model of MapReduce-based Multi-Layer Perception (MLP) algorithm to classify the hybrid compact polarimetric (CP) and fully polarimetric (FP) feature space. Even though a shallow learning model of the automated method is easily scalable, it requires a distinctive shallow learnable feature set for better land cover classification. In this effort, this paper investigates the efficacy of derived feature space compared to direct polarimetric measurements of sensors and shows that the shallow learnable feature set is more effective with both CP & FP observations. Simultaneously, the relevancy of the proposed distributed model of MLP is also justified compared to distributed extreme learning machine (DELM) algorithm and provides a practically implementable scaled-MLP algorithm of shallow learning model. Ultimately, this paper comes up with a better polarimetric signature of land types for both the CP and FP datasets, which can be used in an alternative manner as per data availability for a multi-sensor data analysis.

A Generalized Geodesic Distance-Based Approach for Analysis of SAR Observations Across Polarimetric Modes

Present and future sensors are diversifying from traditional quad polarimetric mode of synthetic aperture radar acquisition. Thus, an approach that is interpretative in nature and applicable across polarimetric modes is required. In this context, the geodesic distance (GD)-based approach within the polarimetric synthetic aperture radar (PolSAR) literature is seen as an eigenvalue-decomposition free approach to interpret and analyze quad PolSAR data. This approach is highly adaptive toward applications due to its ability to compare the SAR observation with a known scatterer/model, or with another SAR observation in general providing a means for direct interpretation. In this work, we show that the GD (originally defined for the quad polarization mode) is generalizable across any arbitrary SAR polarimetric mode while retaining its simple form for ready computation. We show that the GD-based approach provides level ground for comparison of different polarimetric modes given a fixed application. We demonstrate it using change detection as the chosen application. In addition, we show how the behavior of the three roll-invariant GD-based parameters change under different polarimetric modes (e.g., quad, dual, and compact polarization modes). We also discuss how the GD-based approach can also be adapted to ground range detected (GRD) product data, which is presently available from Sentinel-1 and widely used in many applications. However, in this case, we show that only one of the three GD-derived parameters can be defined. We believe this work will make the GD-based approach important for present and future PolSAR applications cutting across sensors and its available polarimetric modes.

Reduction of Rain-Induced Errors for Wind Speed Estimation on SAR Observations Using Convolutional Neural Networks

Reduction of Rain-Induced Errors for Wind Speed Estimation on SAR Observations Using Convolutional Neural Networks