No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Radar Polarimetry for Geoscience Applications
Geocarto International
Abstract
The present volume on radar polarimetry for geoscience applications discusses wave properties and polarization, scattering matrix representation for simple targets, scattering models for point and distributed targets, polarimetric scatterometer systems and measurements, polarimetric radar system design, and polarimetric SAR applications. Attention is given to plane waves in a lossless homogeneous medium-wave polarization, polarization synthesis and response, and coordinate system transformations. Topics addressed include high- and low-frequency scattering, rough-surface scattering models, radiative transfer theory and deficiencies thereof, solutions for the radiative transfer equation, and a radiative transfer model for a forest canopy. Also discussed are network analyzer-based polarimetric scatterometers, calibration of polarimetric scatterometers, synthesized polarization response of distributed targets, and measurement of the propagation parameters of a forest canopy.
... where r is a position vector,ŝ denotes the direction of propagation, I is the Stokes parameters of specific intensity, = κ ε is the extinction matrix, κ ag is the power absorption coefficient of the background material, = P is the phase matrix, and J e is the emission vector. The radiative transfer equation is formulated in terms of three constitutive functions: (1) the extinction matrix, which describes the attenuation of the specific intensity due to absorption and scattering; (2) the phase matrix, which characterizes the coupling between the incident and scattered intensities at every point in the medium; and (3) the emission function, which accounts for the black body emission by the medium [24]. However, since the background material of a chaff cloud is air and the emission term is also negligible in radar remote sensing, we can neglect the loss term and the emission term in the VRT equation, then the VRT equation for a chaff cloud can be rewritten as: ...
... The zeroth-and first-order transformation matrices are given by [24]: ...
... The zeroth-and first-order transformation matrices are given by [24]: ; , sec , , sec , , , , , , , se c , sec , , , , , sec , sec , , s ec ...
Chaff is a passive jammer widely used to disrupt radar or radio-frequency sensors. A mass of chaff fibers dispersed in the air is commonly referred to as a chaff cloud. It is nearly impossible to numerically simulate in real-time the enormous amount of chaff fibers composing the chaff cloud. In this paper, we model the behavior of numerically estimated chaff clouds as probability density functions (PDFs) and apply approximation techniques to estimate the radar cross-section (RCS) of the chaff cloud in real time. To model the aerodynamics of the chaff cloud, we represented the combination of PDFs as functions of time and wind speed. The applied approximation techniques—vector radiative transfer and generalized equivalent conductor method—showed a computation time that cannot be achieved by low-frequency methods such as the method of moments or finite-difference time-domain. Moreover, the dynamic RCS results of the approximation techniques showed a similar trend to those of other studies simulating similar situations. The proposed scheme is effective for real-time chaff cloud simulation, and the modeled dynamics and estimated dynamic RCSs can be a standard baseline for developing new analysis methods for chaff clouds. In the future, the proposed scheme will extend to more chaff fibers and more diverse environmental parameters.
... This paper aims to investigate this ability. Using an approximate scattering model ( [23,24] already used in [17], we attempt to build a full-inversion process for a low complexity problem which is intended to be increased step by step. We have chosen to develop an inversion process able to first solve the source localization problem (3-D imaging), then perform the electromagnetic inversion (scattering coefficients), and finally operate the biophysical inversion (cylinders parameters). ...
... First, at the scatterer level. The dielectric cylinder has to be long enough against the wavelength (infinite cylinder approximation [24]): Its scattered electric and magnetic fields are obtained in the approximation of the infinite length cylinder and are used to calculate fictitious surface magnetic and electric currents. These fictitious surface currents are then the fictitious source of a finite height cylinder when the scattered (E,H) fields are calculated on a finite surface [23,24]. ...
... The dielectric cylinder has to be long enough against the wavelength (infinite cylinder approximation [24]): Its scattered electric and magnetic fields are obtained in the approximation of the infinite length cylinder and are used to calculate fictitious surface magnetic and electric currents. These fictitious surface currents are then the fictitious source of a finite height cylinder when the scattered (E,H) fields are calculated on a finite surface [23,24]. This approximation allows an analytical solution for the scattering of a vertical lossy dielectric cylinder in free space, which can be translated and tilted through translation and Euler angles based rotation operator matrices. ...
... Previous studies using SMAP-R data have focused in preliminary assessments [19]- [22] and preliminary calibrations [26]. Furthermore, in [27] the first polarimetric study using the cross-correlation between H and V signals is conducted, showing the Sinclair [28], [29] coefficients collected by the H/V antennas. ...
... A bistatic HCP radar scenario is sketched in Fig. 1 for a dual linear polarization receiver (horizontal and vertical) when a circular polarization is transmitted [29]. The reflected GNSS signal polarization changes depending on the surface scattering properties. ...
... where ERH and ERV are the RHCP signals scattered over ground and collected by a pair of H and V linearly polarized antennas, SXX are the scattering coefficients for a Forward-Scattering Alignment (FSA) in the Jones basis [29] ...
Global Navigation Satellite System-Reflectometry (GNSS-R) is a promising field with a diverse set of applications. Polarimetric GNSS-R has been shown to be sensitive to different land parameters, such as freeze/thaw, crop growth, or soil moisture. In this article, the polarimetric scenario of a GNSS reflection is defined, and the theoretical basis and the algorithm to retrieve the Stokes parameters of a GNSS-R signal are proposed. The presented algorithms are validated using data collected by the two linear orthogonal polarization channels of the Soil Moisture Active Passive (SMAP) radar antenna working as GNSS-R, known as SMAP-Reflectometry (SMAP-
$R$
). The Stokes parameter retrieval is presented for the SMAP-
$R$
case, and the receiver phase offset and intensity are calibrated. The same calibration procedure proposed by the CYclone GNSS (CYGNSS) mission is applied to the SMAP-
$R$
first Stokes parameter, allowing the retrieval of the normalized bistatic radar cross Section (NBRCS) of such measurement. Finally, the SMAP-
$R$
data are compared to the CYGNSS NBRCS over the ocean, showing an unbiased root-mean-square error of 3.3 dB and a small bias.
... The phase difference between two complex scattering signals is an important parameter in the study of polarimetric data and it is related to the physical properties of the scattering medium under study [1][2][3][4][5][6][7]. To reduce statistical variations, it is often necessary to average data. ...
... Within the validity region of the first-order SPM, the coscattering and cross-scattering amplitudes A (ba) (1) (θ, φ) within the upper medium take the following form [22]: ...
... (1) (θ, φ) depends on rough interface realizations and for a given direction (θ, φ), it is a complex random variable. Let R (ba) and J (ba) be the real and imaginary parts of the scattering amplitude A (ba) (1) (θ, φ) and let R (b'a') and J (b'a') be the real and imaginary parts of ...
Abstract The authors derive the distribution of the phase difference between two multilook scattering signals for a multilayer stack with randomly rough surfaces under a plane wave excitation. First, for infinite slightly rough surfaces described by Gaussian centred stochastic processes, the authors show that the underlying complex scattering signals follow a Gaussian joint distribution. Also, it is demonstrated that this property is within the scope of the first‐order perturbation theory. Secondly, the authors use this joint probability law to derive the closed‐form expression for the probability density function of the phase difference. The theoretical formula is verified by comparison with Monte‐Carlo simulations.
... Extracting valuable feature information for neural network classification involves decomposing PolSAR images into target polarimetric components using these matrices. Researchers have employed Sinclair scattering matrices [11], texture features [12,13], and spatial segmentation features [14] for PolSAR image classification. Pseudo-color synthesis, using decomposed target components, yields color characteristics of the targets, providing diverse information for PolSAR deep learning classification [15]. ...
... (www.preprints.org) | NOT PEER-REVIEWED | Posted: 28 April 2024 doi:10.20944/preprints202404.1726.v111 ...
This study employs the reflection symmetry decomposition (RSD) method to extract polarization scattering features from ground object images, aiming to determine the optimal data input scheme for deep learning networks in polarimetric synthetic aperture radar classification. Eight distinct polarizing feature combinations are designed, and the classification accuracy of various approaches is evaluated using the classic convolutional neural networks(CNNs) AlexNet and VGG16. The findings reveal that the commonly employed 6-parameter input scheme, favored by many researchers, lacks the comprehensive utilization of polarization information and warrants attention. Intriguingly, leveraging the complete 9-parameter input scheme based on the polarization coherence matrix results in improved classification accuracy. Furthermore, the input scheme incorporating all 21 parameters from the RSD and polarization coherence matrix notably enhances overall accuracy and the Kappa coefficient compared to the other 7 schemes. This comprehensive approach maximizes the utilization of polarization scattering information from ground objects, emerging as the most effective CNN input data scheme in this study. Additionally, the classification performance using the second and third component total power values (P2 and P3) from the RSD surpasses the approach utilizing surface scattering power value (PS) and secondary scattering power value (PD) from the same decomposition.
... When analyzing PolSAR images, traditional machine learning algorithms usually rely on shallow features of PolSAR images obtained through feature extraction methods. These shallow features include statistical features such as the linear and circular intensities, linear and circular coefficient of variation, and span [13], as well as target decomposition features such as the Pauli decomposition [15], Freeman decomposition [16], and Huynen decomposition [17]. However, this approach has several drawbacks. ...
... Remote Sens. 2023,15, 4801 ...
Polarimetric synthetic aperture radar (PolSAR) image classification has been an important area of research due to its wide range of applications. Traditional machine learning methods were insufficient in achieving satisfactory results before the advent of deep learning. Results have significantly improved with the widespread use of deep learning in PolSAR image classification. However, the challenge of reconciling the complex-valued inputs of PolSAR images with the real-valued models of deep learning remains unsolved. Current complex-valued deep learning models treat complex numbers as two distinct real numbers, providing limited assistance in PolSAR image classification results. This paper proposes a novel, complex-valued deep learning approach for PolSAR image classification to address this issue. The approach includes amplitude-based max pooling, complex-valued nonlinear activation, and a cross-entropy loss function based on complex-valued probability. Amplitude-based max pooling reduces computational effort while preserving the most valuable complex-valued features. Complex-valued nonlinear activation maps feature into a high-dimensional complex-domain space, producing the most discriminative features. The complex-valued cross-entropy loss function computes the classification loss using the complex-valued model output and dataset labels, resulting in more accurate and robust classification results. The proposed method was applied to a shallow CNN, deep CNN, FCN, and SegNet, and its effectiveness was verified on three public datasets. The results showed that the method achieved optimal classification results on any model and dataset.
... For bare soil and vegetation, the random surface scattering model and first-order radiative transfer (RT) model are employed to calculate geophysical parameters [23,24]. On 5 July 2021, the Chinese meteorological satellite FY-3E was successfully launched. ...
... The discrete vegetation model considers that the vegetation layer can be treated as a single scatter with different size, shape, and spatial distribution probability. In this study, when analyzing vegetation, a double layer noncoherent scattering model for high crown layers was employed, which divides the vegetation layers above the surface into the crown and the trunk layer [24]. The model is based on the Stokes matrix of a single scatterer. ...
Global navigation satellite system (GNSS) reflectometry (GNSS-R) developed into a promising remote sensing technique. However, few previous related studies considered the potential of its polarization. Owing to lack of sufficient in situ measurement data to support comprehensive investigation of GNSS-R polarization, this study used theoretical models and reference to our previous work to explore this topic. The commonly used microwave scattering models are employed to get the bare soil or vegetation scattering properties of GNSS-R configurations, i.e., the random surface scattering model and the first-order radiative transfer equation were improved and then employed to obtain the scattering properties of both bare soil and vegetation. Since the final output of the space-borne GNSS-R missions is a delay Doppler map (DDM), a spaceborne (DDM) simulator, oriented for the Chinese FengYun-3E (FY-3E) GNSS-R payload, was utilized to obtain the final output at different polarizations. Using the developed models (such as the bare soil and vegetation scattering models), corresponding polarization simulations were performed. That is to say, not only the commonly used LR (left hand circular polarizations (LHCP) received and the right hand circular polarizations (RHCP) received) can be presented, but also the scattering properties at RR, VR, and HR (the transmitted signals are RHCP, while the received polarizations are RHCP, vertical (V) and horizontal (H) polarizations, respectively) can be predicted by our developed models. Results reveal obvious polarization differences for the bistatic scattering and DDM. Therefore, the use of GNSS-R polarization information has potential to provide competitive and fruitful results in the future detection of land surface geophysical parameters.
... The RTM may need to be examined further in terms of the effect of the curvature of a leaf on the radar backscatter to be more practical for applying this model to the determination of the backscattering coefficients of vegetation canopies. The effect of curvature on the radar backscatter is quite large for backscattering from a curved dielectric sheet at C-and X-band frequencies [5,6]. Therefore, we may need to examine in detail the effect of the leaf curvature on the RTM. ...
... where p or q denotes v-or h-polarization, m or n is 1 or 2, and vv-, vh-, hv-, and hh-polarizations correspond to 11, 12, 21, and 22 elements of the 4 × 4 transformation matrix that can be computed with the following matrix multiplications with the canopy scattering matrix ̿ , the eigen matrix , the diagonal extinction matrix , the reflectivity matrix , the Stokes scattering operator matrix , and the phase matrix [1,5]. ...
The first-order vector radiative transfer model (FVRTM) is modified mainly by examining the effects of leaf curvature of vegetation canopies, the higher-order multiple scattering among vegetation scattering particles, and the underlying-surface roughness for forward reflection on radar backscattering from farming fields at C- and X-bands. At first, we collected the backscattering coefficients measured by scatterometers and space-borne synthetic aperture radar (SAR), field-measured ground-truth data sets, and theoretical scattering models for radar backscattering from vegetation fields at microwaves. Then, these effects on the RTM were examined using the database at the C- and X-bands. Finally, an improved RTM was obtained by adjusting its parameters, mainly related with the leaf curvature, the higher-order multiple scattering, and the underlying-surface small-roughness characteristics, and its accuracy was verified by comparisons between the improved RTM and measurement data sets.
... Extracting valuable feature information for neural network classification involves decomposing PolSAR images into target polarimetric components using these matrices. Researchers have employed Sinclair scattering matrices [12], texture features [13][14][15], and spatial segmentation features [16] for PolSAR image classification. Pseudo-color synthesis using decomposed target components yields color characteristics of the targets, providing diverse information for PolSAR deep learning classification [17][18][19]. ...
This study employs the reflection symmetry decomposition (RSD) method to extract polarization scattering features from ground object images, aiming to determine the optimal data input scheme for deep learning networks in polarimetric synthetic aperture radar classification. Eight distinct polarizing feature combinations were designed, and the classification accuracy of various approaches was evaluated using the classic convolutional neural networks (CNNs) AlexNet and VGG16. The findings reveal that the commonly employed six-parameter input scheme, favored by many researchers, lacks the comprehensive utilization of polarization information and warrants attention. Intriguingly, leveraging the complete nine-parameter input scheme based on the polarization coherence matrix results in improved classification accuracy. Furthermore, the input scheme incorporating all 21 parameters from the RSD and polarization coherence matrix notably enhances overall accuracy and the Kappa coefficient compared to the other seven schemes. This comprehensive approach maximizes the utilization of polarization scattering information from ground objects, emerging as the most effective CNN input data scheme in this study. Additionally, the classification performance using the second and third component total power values (P2 and P3) from the RSD surpasses the approach utilizing surface scattering power value (PS) and secondary scattering power value (PD) from the same decomposition.
... In order to solve the problems, we will provide three models for evaluation in this section. Since we want to calculate the specular reflectivity at RR polarization, we will employ the wave synthesis technique to get the polarization properties [29]. ...
Polarization in GNSS-R (Global Navigation Satellite System-Reflectometry) or SoOP-R (Signal of Opportunity-Reflectometry) is commonly used for retrieving geophysical parameters. However, the attention towards other polarizations of reflected signals has increased with developments in this field. The widely used equation for RR polarization suggests that it decreases as soil moisture content increases, which contradicts experimental data.The accurate forward calculation of RR polarization is essential for the subsequent retrieval algorithm in polarization GNSS-R/SoOP-R. To address this issue, three new models have been developed: Spec4PolR (Specular reflectivity model for polarization GNSS-R), SPM4Pol (small perturbation model for polarization GNSS-R), and Umich4Pol (Umich model for polarization GNSS-R). The Mueller matrix of these three models has been presented, and the wave synthesis technique has been employed to calculate the reflectivity at RR polarization. Spec4polR uses only three elements in the Mueller matrix for final reflectivity, while five elements are used in Umich4polR. In SPM4Pol, all elements construct the Mueller matrix, and only nine elements are employed for calculation. The effects of each element on soil moisture content are presented, and the final reflectivity at RR polarization is illustrated. However, due to the simple formulation of Spec4Pol, its reflectivity at RR polarization still decreases as soil moisture content increases. On the other hand, the results of SPM4Pol and Umich4Pol are consistent with measured data, and the reflectivity at RR polarization increases as soil moisture content increases. The formula developed in this paper for calculating RR polarization will contribute to subsequent polarization studies and geophysical parameter retrieval based on RR polarization.
... ship target detection), it is important to analyze the characteristic of azimuth ambiguities of targets. As polarimetric SAR (PolSAR) system provides four channel capabilities to measure the four scattering factors of a target [9], one can use more available information to analyze the characteristic of azimuth ambiguities of targets. This paper studies polarimetric characteristics of ship target and its azimuth ambiguities by using Eigenvalue-Eigenvector decomposition method. ...
Azimuth ambiguities in SAR image are very common phenomena. The ambiguities are visually identifiable due to their high intensities in low radar backscatter background of sea environment. The ambiguities can be mistaken as ships and cause false alarms in ship detection. Therefore, for maritime applications (i.e. ship target detection), it is important to analyze the characteristic of azimuth ambiguities of targets. By using Eigenvalue-Eigenvector decomposition method, we analyze the polarimetric characteristics of ships and ambiguities. We conclude that the eigenvalues can be used to differentiate ship target and its azimuth ambiguities. One C-band JPL AIRSAR polarimetric data has been chosen to validate the conclusions.
... In addition to all these polarimetric parameters cited above and to the conventional multipolarization (HH, HV, and VV) channels, the HH-VV phase difference (φ HH -φ VV ) was also investigated. The HH-VV phase difference, which used to be among the standard polarimetric parameters investigated for a natural target characterization [55][56][57], has been widely promoted for the detection of sand subsurface wet structures in arid regions [58,59]. ...
Recently, it has been shown that the long penetrating polarimetric L-band ALOS is very promising for boreal and subarctic peatland mapping and monitoring. The unique information provided by the Touzi decomposition, and the dominant-scattering-type phase in particular, on peatland subsurface water flow permits an enhanced discrimination of bogs from fens, two peatland classes that can hardly be discriminated using conventional optical remote sensing sensors and C-band polarimetric SAR. In this study, the dominant and medium-scattering phases generated by the Touzi decomposition are investigated for discontinuous permafrost mapping in peatland regions. Polarimetric ALOS2, LiDAR, and field data were collected in the middle of August 2014, at the maximum permafrost thaw conditions, over discontinuous permafrost distributed within wooded palsa bogs and peat plateaus near the Namur Lake (Northern Alberta). The ALOS2 image, which was miscellaneously calibrated with antenna cross talk (−33 dB) much higher than the actual ones, was recalibrated. This led to a reduction of the residual calibration error (down to −43 dB) and permitted a significant improvement of the dominant and medium-scattering-type phase (20∘ to −30∘) over peatlands underlain by discontinuous permafrost. The Touzi decomposition, Cloude–Pottier α-H incoherent target scattering decomposition, and the HH-VV phase difference were investigated, in addition to the conventional multipolarization (HH, HV, and VV) channels, for discontinuous permafrost mapping using the recalibrated ALOS2 image. A LiDAR-based permafrost classification developed by the Alberta Geological Survey (AGS) was used in conjunction with the field data collected during the ALOS2 image acquisition for the validation of the results. It is shown that the dominant- and scattering-type phases are the only polarimetric parameters which can detect peatland subsurface discontinuous permafrost. The medium-scattering-type phase, ϕs2, performs better than the dominant-scattering-type phase, ϕs1, and permits a better detection of subsurface discontinuous permafrost in peatland regions. ϕs2 also allows for a better discrimination of areas underlain by permafrost from the nonpermafrost areas. The medium Huynen maximum polarization return (m2) and the minimum degree of polarization (DoP), pmin, can be used to remove the scattering-type phase ambiguities that might occur in areas with deep permafrost (more than 50 cm in depth). The excellent performance of polarimetric PALSAR2 in term of NESZ (−37 dB) permits the demonstration of the very promising L-band long-penetration SAR capabilities for enhanced detection and mapping of relatively deep (up to 50 cm) discontinuous permafrost in peatland regions.
... Therefore, the wave synthesis technique is employed in model development. The polarization state of the wave may also be expressed in terms of the Stokes vector, in which the orientationψand ellipticity anglesχare sufficient for the complete specification of the polarization (Ulaby and Elachi 1990). ...
During the past three decades, Global Navigation Satellite System (GNSS)-Reflectometry (GNSS-R) has become a promising remote sensing technique. In developing GNSS-R, several spaceborne missions have been launched, e.g., the UK’s TDS-1, NASA’s CYGNSS, and China’s BuFeng-1A/B and FY-3E GNOS-R satellites.The application of GNSS-R to vegetation monitoring has attracted attention because GNSS-R signals have less saturation effect than those of a traditional monostatic radar. However, most previous related work has concentrated on the analysis of experimental data, with little attention given to the physical models or polarization. We present a spaceborne vegetation simulator for full polarization GNSS-R, i.e., the LAGRS-Veg (Land surface GNSS-R simulator-Vegetation) model. This model is based on the vegetation radiative transfer equation model and integrated with the GNSS scattering model. Thus, it provides a comprehensive end-to-end simulator for spaceborne GNSS-R study of vegetation. Using this simulator, the impact of vegetation water content, biomass, soil moisture, and surface roughness on delay Doppler map measurements can be thoroughly analyzed because they are all based on physical scattering mechanisms. The effects of observation geometry and different polarizations can all be included and analyzed using LAGRS-Veg. The presented work will benefit spaceborne data simulation, vegetation parameter retrievals, and future spaceborne sensor design.
... However, the S measured in a particular polarization basis might not provide optimum scattering response from targets possessing complex geometry. Hence, one can synthesize the received power for any combination of transmit-receive antenna polarization basis using the technique known as polarization synthesis [2]. ...
This work extends the polarimetric signature (PS) concept from full polarimetric Synthetic Aperture Radar (SAR) data to compact polarimetric SAR data, following the wide application of this mode for land cover characterization. We first demonstrate the variation of PS for several elementary targets and then illustrate its utility for improved characterization of various natural & human-made targets. We found that the compact-pol signature (CPS) allows enhanced discrimination among the land cover targets, viz., waterbody, different types of the urban area, and vegetation. The proposed method holds significant interest for the Radar Constellation Mission (RCM) datasets and future compact polarimetric SAR missions for enhanced target characterization .
... Two bistatic radar scenarios over land are sketched in Fig. 1 when a circular polarization is transmitted 92 (Ulaby and Elachi, 1990). Over land, the polarimetric properties of the received GNSS signal change 93 depending on the reflection surface. ...
Polarimetric Global Navigation Satellite System – Reflectometry (GNSS-R) has proven to be sensitive to different land parameters, such as vegetation or soil moisture. In this manuscript, the correlation of the Stokes parameters of the GNSS reflections collected by the Soil Moisture Active Passive - Reflectometry (SMAP-R) with SMAP soil moisture (SM), SMAP vegetation optical depth (VOD), and the Global Forest Canopy Height (CH) are analyzed. SMAP-R is the first hybrid compact polarimetric reflectometer in space, and it provides a valuable dataset useful to design future GNSS-R mission that target polarimetric retrievals. The SMAP-R total-intensity reflectivity is presented, estimated from the bistatic radar cross section assuming coherent reflection. The calibrated intensity and the normalized Stokes parameters of the received wave are compared for a 1-year period to the SMAP VOD, the SMAP SM, the Global Forest CH, and the standard deviation of the digital elevation model. To provide a first assessment of the performance of full-Polarimetric GNSS-R to estimate SM and VOD, a linear regression approach is implemented using the four Stokes parameters and the child Stokes parameters. The two approaches show a correlation coefficient of R = 0.73 and R = 0.55 to estimate SM and VOD, respectively, enabling the estimation of both variables using polarimetric GNSS-R data without the need for ancillary data. Finally, the differences between the polarimetric signature at different incidence angles and polarizations are discussed.
... Attempts to describe the backscattering from vegetation covered areas have been made since the late 1970s. They have evolved from the simple 'cloud' model of Attema and Ulaby (1978) to multilayered, multi-constituent models like the MIchigan MIcrowave Canopy Scattering Model (MIMICS) proposed by Ulaby and Elachi (1990) or the radiative transfer model of Karam et al. (1992). More complex radiative transfer (RT) models have been developed to take into account the 3-dimensional canopy structure (e.g. ...
Joint retrieval of vegetation status from synthetic aperture radar (SAR) and optical data holds much promise due to the complimentary of the information in the two wavelength domains. SAR penetrates the canopy and includes information about the water status of the soil and vegetation, whereas optical data contains information about the amount and health of leaves. However, due to inherent complexities of combining these data sources there has been relatively little progress in joint retrieval of information over vegetation canopies. In this study, data from Sentinel-1 and Sentinel-2 were used to invert coupled radiative transfer models to provide synergistic retrievals of leaf area index and soil moisture. Results for leaf area are excellent and enhanced by the use of both data sources (RSME is always less than 0:5 and has a correlation of better than 0:95 when using both together), but results for soil moisture are mixed with joint retrievals generally showing the lowest RMSE but underestimating the variability of the field data. Examples of such synergistic retrieval of plant properties from optical and SAR data using physically based radiative transfer models are uncommon in the literature, but these results highlight the potential for this approach.
... , , , and . The complex scattering matrix [48] is used to describe the electromagnetic scattering characteristic of the ground objects. The complex scattering matrix is given by: ...
Polarimetric synthetic aperture radar (PolSAR) image classification is a pixel-wise issue, which has become increasingly prevalent in recent years. As a variant of the Convolutional Neural Network (CNN), the Fully Convolutional Network (FCN), which is designed for pixel-to-pixel tasks, has obtained enormous success in semantic segmentation. Therefore, effectively using the FCN model combined with polarimetric characteristics for PolSAR image classification is quite promising. This paper proposes a novel FCN model by adopting complex-valued domain stacked-dilated convolution (CV-SDFCN). Firstly, a stacked-dilated convolution layer with different dilation rates is constructed to capture multi-scale features of PolSAR image; meanwhile, the sharing weight is employed to reduce the calculation burden. Unfortunately, the labeled training samples of PolSAR image are usually limited. Then, the encoder–decoder structure of the original FCN is reconstructed with a U-net model. Finally, in view of the significance of the phase information for PolSAR images, the proposed model is trained in the complex-valued domain rather than the real-valued domain. The experiment results show that the classification performance of the proposed method is better than several state-of-the-art PolSAR image classification methods.
... where x and y are the elements of the reflection matrix known as complex amplitude reflection coefficients (Ackley and Keliher, 1979;Ulaby and Elachi, 1990;Fujita et al., 2000Fujita et al., , 2006. Here we only use the real part of x and y . ...
Ice crystals are mechanically and dielectrically anisotropic. They progressively align under cumulative deformation,
forming an ice-crystal-orientation fabric that, in turn, impacts ice deformation. However, almost all the observations
of ice fabric are from ice core analysis, and its influence on the ice flow is unclear. Here, we present a non-linear
inverse approach to process co- and cross-polarized phase-sensitive radar data. We estimate the continuous depth profile
of georeferenced ice fabric orientation along with the reflection ratio and horizontal anisotropy of the ice column. Our
method approximates the complete second-order orientation tensor and all the ice fabric eigenvalues. As a result, we
infer the vertical ice fabric anisotropy, which is an essential factor to better understand ice deformation using
anisotropic ice flow models. The approach is validated at two Antarctic ice core sites (EPICA (European Project for Ice Coring in Antarctica) Dome C and EPICA Dronning
Maud Land) in contrasting flow regimes. Spatial variability in ice fabric characteristics in the dome-to-flank
transition near Dome C is quantified with 20 more sites located along with a 36 km long cross-section. Local
horizontal anisotropy increases under the dome summit and decreases away from the dome summit. We suggest that this is a
consequence of the non-linear rheology of ice, also known as the Raymond effect. On larger spatial scales, horizontal
anisotropy increases with increasing distance from the dome. At most of the sites, the main driver of ice fabric
evolution is vertical compression, yet our data show that the horizontal distribution of the ice fabric is consistent
with the present horizontal flow. This method uses polarimetric-radar data, which are suitable for profiling radar
applications and are able to constrain ice fabric distribution on a spatial scale comparable to ice flow observations
and models.
... The continuous and strong absorption of water in the near-infrared and short-wave infrared regions in remote-sensing images has led to the development of different water-detection indexes, such as the normalized differential water index (NDWI) [41], the modified normalized difference water index (MDNWI) [42], the automated water extraction index (AWEI) [6], and the water index (WI) [43]. Although the spatial and temporal resolutions of optical remote-sensing images are constantly improving, their data quality is easily affected by climate conditions, especially clouds [44]. Synthetic aperture radar (SAR) is a remote-sensing microwave sensor that is capable of continuous operation and is sensitive to water [45,46]. ...
Urban resilience to natural disasters (e.g., flooding), in the context of climate change, has
been becoming increasingly important for the sustainable development of cities. This paper presents a
method to assess the urban resilience to flooding in terms of the recovery rate of different subdistricts
in a city using all-weather synthetic aperture radar imagery (i.e., Sentinel-1A imagery). The factors
that influence resilience, and their relative importance, are then determined through principal component
analysis. Jakarta, a flood-prone city in Indonesia, is selected as a case study. The resilience of
42 subdistricts in Jakarta, with their gross domestic product data super-resolved using nighttime-light
satellite images, was assessed. The association between resilience levels and influencing factors, such as
topology, mixtures of religion, and points-of-interest density, were subsequently derived. Topographic
factors, such as elevation (coefficient = 0.3784) and slope (coefficient = 0.1079), were found to have the
strongest positive influence on flood recovery, whereas population density (coefficient = 0.1774) a
negative effect. These findings provide evidence for policymakers to make more pertinent strategies to
improve flood resilience, especially in subdistricts with lower resilience levels.
... Therefore, for the monitoring and retrieval of vegetation biomass, it is urgent to carry out research on the polarization characteristics of GNSS-R [51,53]. It is necessary to point out the difference in polarization from the theoretical mechanism level, and carry out the parameterized models based on polarization characteristics [58]. ...
Vegetation is an important part of the terrestrial ecosystem and plays a vital role in the global carbon cycle. Traditional remote sensing methods have certain limitations in vegetation monitoring, and the development of GNSS-R (Global Navigation Satellite System-Reflectometry) technology provides a new and complimentary method. With the CYGNSS (Cyclone Global Navigation Satellite System) launch and the increased data acquisition, the use of spaceborne GNSS-R for vegetation monitoring has become a research hotspot. However, due to the complex characteristics of vegetation, its application in this field is still in the exploratory research stage. On the basis of reviewing the current research status, this paper points out the weak links of this technology in terms of polarization and observation geometry. Combined with the microwave vegetation scattering model, this paper analyzes the full polarization bistatic scattering characteristics of vegetation and points out the influence of vegetation parameters (density, water content, and vegetation diameters). The potential feasibility of polarization GNSS-R and future development trends of GNSS-R technology in quantitative retrieval (such as vegetation water content and biomass) are also discussed.
... For SM studies, observations at L-band frequencies of 1.2-1.4 GHz are desirable due to larger penetration depths (Ulaby and Elachi, 1990;Liu et al., 2013Liu et al., , 2016aSteele-Dunne et al., 2017). Currently, the European Space Agency (ESA) -Soil Moisture and Ocean Salinity (SMOS) and National Aeronautics and Space Administration (NASA) Soil Moisture Active Passive (SMAP) missions (Kerr et al., 2001;Entekhabi et al., 2010) provide brightness temperatures (T B ) at L-band and SM products every 2-3 days at 25 km and 36 km (or 9 km enhanced), respectively. ...
Soil moisture (SM) retrieval in agricultural regions during the growing seasons is particularly challenging due to high spatial variability and dynamic vegetation conditions. The retrievals have been problematic even when the passive signatures at different spatial scales match well since they depend upon the accuracy of vegetation information such as the vegetation water content (VWC). The VWC used in the Soil Moisture Active Passive (SMAP) single channel retrieval algorithm (SCA) is derived from remotely sensed, climatologically-based Normalized Difference Vegetation Index (NDVI), which does not respond to real-time vegetation dynamics and is prone to saturation. This study explored the differences and seasonal trend in passive signatures and SM at satellite-and field-scales and investigated uncertainties in retrievals arising from different approaches used to estimate VWC from optical and radar indices. It used high temporal resolution, ground-based data collected during the SMAP Validation-Microwave Water, Energy Balance Experiment in 2016 (SMAPVEX16-MicroWEX) during a growing season of corn and soybean. Overall, the brightness temperatures (T B) from SMAP matched well with the upscaled, ground-based T B , with a root mean square differences (RMSDs) of about 5 K. In contrast, the SMAP SM retrievals were worse during rapid vegetation growth in the mid-season, with higher RMSDs compared to the upscaled in situ SM, than those in the late-season. In addition, the ground-based T B from corn and soybean were similar in the early and the late seasons, while their emission differences were > 40 K in the mid season. This indicates the importance of accurate VWC information, particularly during the early and late growing seasons, to account for sub-pixel heterogeneities in agricultural regions. VWC obtained from five optical and radar indices were used in the SMAP SCA for soil moisture retrieval for the entire growing season of corn. The NDVI-based VWC provided SM retrievals that were consistently lower compared to those using in situ VWC, with a higher RMSD of 0.030 m 3 /m 3 and a negative bias of 0.020 m 3 /m 3 for VWC > 4 kg/m 2. The Normalized Difference Water Index (NDWI)-derived VWC resulted in lower SM retrieval RMSD of 0.022 m 3 /m 3 when compared with in situ SM. Among the three radar indices, vertically polarized cross-pol ratio (CR vv)-derived VWC provided similar RMSDs in retrieved SM as the NDWI-derived VWC during the growing season. The radar vegetation index (RVI)-derived VWC improved in the late season compared to the in situ VWC and resulted in SM retrievals with RMSDs 2 similar to the CR vv-derived retrievals. Results presented here suggest that SMAP SCA SM retrievals could be improved through the use of near-real time NDWI and CR vv-derived vegetation information. Microwave data are available regardless of cloud cover, so the guaranteed availability of CR vv to capture seasonal and interannual variability is advantageous.
... Therefore, we have employed the polarization synthesis to get the corresponding scattering properties. In order to obtain the circular polarization scattering coefficients, this paper uses the method of polarization synthesis [25]. ...
The transition of the freeze–thaw state of the land surface soil occurs every year with the season and is closely related to the human living environment. The freezing and thawing changes of the ground surface have important effects on hydrological activities, meteorological conditions, and ecological gas dynamics. Traditional monitoring methods have their limitations. In the past two decades, the emerging GNSS-R/IR (Global Navigation Satellite System-Reflectometry/Interference Reflectometry) technology has provided a new method for monitoring the surface f state; however, fewer works have paid attention to the scattering mechanism models in the current study. In this paper, a forward GNSS multipath model suitable for a complex cold surface is developed. The dielectric constant model with different surface parameters is added. The calculation of snow layer attenuation is employed to take the snow cover into consideration. Based on the first-order radiation transfer equation model, a polarization synthesis method is used to obtain the circularly and linearly polarized vegetation specular scattering characteristics. The surface characteristics and antenna model are coupled. A more detailed forward GNSS multipath model of frozen and thawed soil under complex surface conditions is established. The model is used to simulate and analyze the forward GNSS multipath (Signal to Noise Ratio (SNR), phase and pseudorange) responses of frozen and thawed soil under complex surface conditions (soil salinity, snow and vegetation coverage). Studies have shown that when the soil changes from freezing to thawing due to the change in the phase of the water in the soil, the dielectric constant and BRCS (bi-static radar cross-section) increase, causing the increase in the amplitude of the multipath observation. The higher the salinity content, the larger the amplitude of the multipath observation. The attenuation of the snow cover and the vegetation layer will lead to the reduction of the multipath observation amplitude. For the first time, the model developed by this paper reveals the GNSS multipath observation response of frozen and thawed soil under complex surface conditions in detail, which can provide some theoretical support for subsequent experimental design and data analysis.
... In the microwave spectrum, the total backscatter comprises contributions from the air-surface, snowpack volume and contributions from the underlying ground layer (Ulaby and Elachi, 1990). In the passive microwave spectrum, the brightness temperature corresponding to the emissivity from the snowpack volume and the underlying ground layer are used. ...
In this chapter, we first discuss the significance of snow studies in the alpine regions. We then give an overview of the snowpack properties in the alpine regions that are essential for snowpack characterization. We then introduce the role of remote sensing in the context of retrieving the snowpack's physical properties, followed by discussions on the primary techniques used for the same. Finally, we provide the implication of these techniques for monitoring the alpine snowpack and their potential application for hydrological modeling and avalanche forecasting. This chapter is useful for beginners and experienced readers interested in the studies of the alpine cryosphere.
... In forestry, PolSAR data are adopted to monitor the fire and excessive logging as well as estimate the biomass in forest [6]. In geology, PolSAR data are employed to analyze information such as geological structure, mineral distribution, surface roughness, ground coverage, and soil moisture [7]. Polarimetric SAR data classification is the key for data interpretation and one of the important research for PolSAR data processing. ...
Classification of polarimetric synthetic aperture radar (PolSAR) images has achieved good results due to the excellent fitting ability of neural networks with a large number of training samples. However, the performance of most convolutional neural networks (CNNs) degrades dramatically when only a few labeled training samples are available. As one well-known class of semi-supervised learning methods, graph convolutional networks (GCNs) have gained much attention recently to address the classification problem with only a few labeled samples. As the number of layers grows in the network, the parameters dramatically increase. It is challenging to determine an optimal architecture manually. In this paper, we propose a neural architecture search method based GCN (ASGCN) for the classification of PolSAR images. We construct a novel graph whose nodes combines both the physical features and spatial relations between pixels or samples to represent the image. Then we build a new searching space whose components are empirically selected from some graph neural networks for architecture search and develop the differentiable architecture search method to construction our ASGCN. Moreover, to address the training of large-scale images, we present a new weighted mini-batch algorithm to reduce the computing memory consumption and ensure the balance of sample distribution, and also analyze and compare with other similar training strategies. Experiments on several real-world PolSAR datasets show that our method has improved the overall accuracy as much as 3.76% than state-of-the-art methods.
... To get the circular polarization properties, here we employ the wave synthesis technique to make the scattering models can calculate various polarization combination (Ulaby and Elachi 1990). ...
Signals of Opportunity Reflectometry (SoOp-R) employs the communication system, GNSS (Global Navigation Satellite System) constellation and other potential Signals of Opportunity (SoOp) as the transmitters. In recent years, it has gained increased interests. Several experiments have been carried out, however it is still in the initial development stage. Theoretical predictions of SoOp Reflectometry for land surface parameters detection, such as soil moisture and vegetation biomass, should be carried out simultaneously. Meanwhile, at present less works are paid attention to the polarization study of the polarizations. The first-order radiative transfer equation models are employed here and they are developed according to the wave synthesis technique to get the various polarization combinations. Using the two models as analysis tools, we simulate the bistatic scattering at all potential SoOp Reflectometry bands, i.e., P-, L-, C- and X-band for circular polarizations and linear polarizations. While the original commonly used microwave scattering models are linear polarizations, here we compare the difference. Although the models can simulate bistatic scattering at any incident angles and scattering angles. Four special observation geometry are taken into considerations during the analysis. Using the developed models as tools, the developed models establish the relationship between the land surface parameters (such as soil moisture, soil roughness and vegetation water content, diameters et al.) and bistatic radar cross section. The forward scattering models developed here enables the understanding of the effects of different geophysical parameters and transmitter–receiver observation scenarios on the bisatic scattering at any polarization combinations for any potential SoOP reflectometry bands. Robust retrieval methods for soil moisture and vegetation biomass can benefit from the forward scattering models.
Estimating above-ground biomass (AGB) using machine learning (ML) algorithms and multi-sensor satellite data is a promising approach for monitoring and managing forest resources. This research integrated synthetic aperture radar (SAR) and multispectral imagery alongside in-field observations to accurately estimate above-ground biomass (AGB) in the Purna regional landscape of northern Western Ghats, India. The satellite data employed in the study included dual-polarization (VV + VH) imagery from Sentinel-1 and multi-spectral bands from Sentinel-2, processed and analysed using advanced ML algorithms. The ML algorithms, namely Random Forest (RF), Extreme Gradient Boosting (XGB), and Boosted Regression Trees (BRT), were strategically applied across different model scenarios to determine their effectiveness in AGB prediction. The XGB model displayed the highest accuracy with an R2 value of 0.61 and the lowest RMSE of 37.85 t/ha. The spatial distribution of AGB was successfully mapped, showing varied biomass concentrations throughout the study area. The study’s findings demonstrate the potential of integrating SAR and multispectral data for enhanced AGB estimation and suggest that ML models, specifically algorithms like RF, XGB, and BRT can address the complex relationships between AGB and satellite-derived variables more effectively than traditional methods.
This manuscript provides the first Hybrid Compact Polarimetric (HCP) Global Navigation Satellite System – Reflectometry (GNSS-R) analysis of the cryosphere. By reconstructing the full-Stokes parameters and the child Stokes parameters from the reflectometry signals collected by the SMAP radar receiver, the sensitivity to the geophysical changes of the Earth’s landscapes of these opportunistic signals are analyzed. This analysis is focused over the Artic Sea ice formation zone, the northern latitudes presenting freeze and thawed transitions, and the Greenland ice sheet. We conduct a sensitivity analysis for each landscape. The signals analyzed over the Arctic Sea show correlation with sea ice concentration (SIC) of 0.77 for the second Stokes parameter
S
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub>
and 0.7 for the child Stokes parameter χ, which describes the ellipticity of the wave. The signals analyzed over the northern latitudes present a sensitivity to the two states (frozen and thawed). The analysis over Alaskan landscape shows a total reflectivity Γ
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub>
difference between freeze and thaw states of ~8 dB with a standard deviation of ~1 dB. Greenland presents a challenge since available datasets are not from the same timeframe, despite of this our initial evaluation shows that there is a ~0.5 correlation between ice thickness and the third Stokes parameter
S
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub>
and the relative phase δ between the two linear electromagnetic components, h-pol and v-pol. This manuscript provides an extensive view of the HCP GNSS-R signals over the cryosphere and finds initial sensitivities to the seasonal geophysical changes of the different areas.
The Soil Moisture Active Passive (SMAP) mission has dramatically benefited our knowledge of the Earth’s surface processes. The SMAP mission was initially designed to provide complementary L-band measurements from a radiometer and a radar, producing geophysical measurements at a finer spatial resolution than the radiometer alone. Both instruments, sensitive to the geophysical parameters in the swath, provided independent measurements at different spatial resolutions. A few months after SMAP’s launch, the radar transmitter’s high-power amplifier suffered an anomaly, and the instrument could no longer return data. During recovery activities, the SMAP mission switched the radar receiver frequency facilitating the reception of Global Positioning System (GPS) signals scattered off the Earth’s surface, and enabling the radar to become the first spaceborne polarimetric Global Navigation Satellite System – Reflectometry (GNSS-R) instrument. With more than 7 years of continued measurements, SMAP GNSS-R data are the most extensive existing GNSS-R dataset and the only one providing GNSS-R polarimetric measurements. We demonstrate that the SMAP polarimetric GNSS-R reflectivity, derived from Stokes parameters mathematical formulation, can enhance the radiometer data over dense vegetation areas, recovering some of the original SMAP radar capability to contribute to the science products and pioneering the first polarimetric GNSS-R mission.
O presente trabalho teve como objetivo avaliar o uso potencial de imagens ALOS/PALSAR, no modo dual de imageamento FDB (Fine Beam Double Polarization) nível 1.1 de correção (banda L), na caracterização e no mapeamento do uso e cobertura da terra no semi-árido brasileiro. Neste estudo foram utilizados algoritmos de classificação MAXVER-ICM e de Wishart, envolvendo pares e/ou o conjunto de componentes derivadas da matriz de covariância [AHH, AHV e AHH*HV].. Obtidas as classificações por esses métodos acima mencionados, foram feitas avaliações do grau de exatidão de mapeamento através da estatística Kappa. Levantamentos fisionômico-estruturais das fácies de caatinga e também pontos de observações das tipologias de uso e cobertura da terra, devidamente georreferenciados, serviram como amostras de treinamento e calibração temática, ficando definidas as seguintes classes de uso e cobertura da terra para área em estudo: caatinga arbórea, caatinga arbórea-arbustiva, caatinga arbustiva, agricultura e corpos d'água. O procedimento metodológico aplicado mostrou que o melhor resultado obtido foi no uso do conjunto de componentes [AHV, AHV e AHH*HV], por meio da classificação MAXVER-ICM, apresentando uma exatidão global de 66% e índice Kappa 0.58.
Dado o crescente interesse no monitoramento de florestas tropicais, este trabalho apresenta uma análise quantitativa de classes de uso e cobertura da terra em dados SAR complexos, com ênfase nas respostas espectrais e análise de composição colorida. Para isto, foram utilizadas as imagens do sistema SAR orbital RADARSAT-2, operante na banda C, na área da Floresta Nacional do Tapajós no Estado do Pará. Por meio de pontos GPS (Global Positioning System), obtidos em trabalho de campo, realizou-se a seleção de amostras das classes para geração de respostas polarimétricas (RP). As amostras forneceram pontos máximos e mínimos empregados nos cálculos da fração de polarização e coeficiente de variação para cada RP. Foram obtidos ainda os ângulos de elipticidade e orientação, para posterior simulação de imagens e análise de mudança em uma composição RGB aditiva. Os resultados para fração de polarização e coeficiente de variação demonstraram a proporção de radiação despolarizada e sua influência na intensidade do sinal ao mudar a polarização da onda. Em termos práticos, observou-se que a classe floresta apresentou pouca mudança ao variar a polarização da onda incidente, e as classes de agricultura e solo apresentaram uma grande variação. A análise de fase para cada classe não apresentou tendência nas distribuições. As composições coloridas permitiram realçar os pontos de mudança e visualizar as variações na intensidade do sinal, com base nos atributos da elipse de polarização obtidos das respostas polarimétricas. Sugere-se aplicação de algoritmos de classificação para verificar os efeitos da polarização na detecção de objetos.
This paper presents a new computational approach that allows rapid analysis of the Electro-Magnetic Scattering characteristics (EMS) of static or moving complex radar targets. The scattering features of the object are represented through a generalized scattering matrix H, whose elements can be measured or computed using conventional numerical techniques, e.g., CST software in the proposed case, and considering prescribed sampled directions. A cardinal series then is adopted to reconstruct the complete scattering pattern by suitably extending the approach to calculate the target’s scattering matrix for any incidence wave and any observation point. The number of finite samples, that is, the dimension of the scattering matrix depends only on the maximum dimension of the target. A PEC sphere and a PEC three-dimensional (3D) complex object have been analyzed in detail. Precise, fast, and stable sampling algorithms have been applied to these targets in the static case and in motion. In particular, the field scattered by the moving objects is then used to carry out the Micro-Doppler analysis of the object radar signature.
At present, the neural network is often based on the real field of operation, research shows that, compared with the real field, the complex has incomparable advantages in the field of image processing, such as the complex represents more information, such as the phase information and modulus value, which play a great role in some fields. To take full advantage of complex data, This paper mainly studies CNN network, and through complex value processing, and get Complex Convolutional Neural Networks(CCN), complete the construction of complex convolution neural network. In order to study complex neural network, we start from two aspects, one is convolution operation, the other is network construction. In this paper, we use ENet as the basic structure of the model, replace the convolutional structure, pooling structure, and BatchNorm structure with the complex form, use it in the Flevoland dataset, and get a good test results.
Deep learning works normally in PolSAR image classification because the complex terrain scattering characteristic results in large intraclass differences and high interclass similarity. Deep metric learning (DML) aims to make the features keep a closer intraclass and a farther interclass distance. Therefore, we introduce DML and then propose an N-cluster generative adversarial net (N-cluster GAN) framework for PolSAR image classification. However, existing DML losses mainly focus on the relationship between individual samples in feature space. Hence, we propose N-cluster loss that pays more attention to the overall structure of all samples. Meanwhile, traditional hard negative sample mining methods occupy lots of computational resources. In addition, the hard level of the negative samples will affect the model’s performance. Therefore, we explore a new method based on a GAN framework to replace the sample mining. Positive N-cluster loss is added to the discriminator (
$D$
), and a negative one is added to the generator (
$G$
). In this way,
$D$
will possess better classification ability, and
$G$
can produce hard negative samples for
$D$
. Then, the hard level of the generated negative samples will change with the discrimination of
$D$
, which is appropriate for the proposed model. N-cluster loss can be directly calculated through the extracted features rather than redundant data preparation. The proposed model is verified on four PolSAR datasets from two aspects of the loss function and negative samples mining. Then, it achieves competitive performance compared with state-of-the-art algorithms.
The real and imaginary components of the scattering matrix elements measured for each pixel in single look fully polarimetric SAR(POLSAR) are coherently combined from a large number of scatterers in each resolution bin. These components for the three independent Sinclair matrix scattering elements have normal distributions when collected over homogeneous regions. This is observed from the ocean surface, homogeneous desert regions, agricultural areas and more. These distributions are studied in this paper. Other distributions such as those for the amplitude ratios and the phase differences of the Sinclair matrix elements are addressed. This is done for single look fully polarimetric SIR-C data at L and C band and TerraSAR-X data. Distributions of observables from 4 look SIR-C data at L and C band are also addressed and compared to the single look distributions. All terrestial surfaces are considered where data is available.
Recently, many studies exploit deep neural networks to promote terrain segmentation in polarimetric synthetic aperture radar (PolSAR) images. However, these works usually inherit the nature-scene approaches directly and may not be robust for PolSAR image segmentation task. The main limitations include single-type feature construction, weak feature consistency, and geometry-agnostic collection of scattering information. In this paper, we present the DENet, a double encoder network with feature refinement and region adaption for the terrain segmentation in PolSAR images. First, a double encoder architecture is proposed to leverage the multi-type information of PolSAR images, which can provide more discriminative features than the previous methods using the single-type feature. Second, considering that the polarization information has strong consistency over the category-identical regions, a polarization-guided refinement module is proposed to maintain the feature consistency in the PolSAR segmentation model. This design alleviates the phenomenon of incomplete and fragmented segmentation results. Third, in view of the rich targets characteristics in the scattering information, a region-adaptive convolution module is developed to facilitate the scattering information collection over the geometry-irregular regions. This design can improve the segmentation accuracy on the geometry-irregular regions. Extensive experiments are conducted on six PolSAR images to verify the effectiveness of the DENet. Compared with the previous works, our method achieves competitive performance.
In this article, we propose a novel method of estimation of crack orientation on a metallic surface using linearly polarized circular synthetic aperture radar (SAR) imaging. We modeled a crack using a thin linear scatterer and considered the rotation of polarization basis vectors that were designed to match a predefined orientation vector corresponding to a temporary estimate of the orientation of the linear scatterer. The radar image intensity was expected to be a maximum when the specified orientation vector exactly matched the true scatterer orientation angle; therefore, to estimate the orientation of the scatterer, we aimed to obtain the maximum intensity of the image by varying the orientation vector. We validated the proposed scheme of estimation via a numerical electromagnetic simulation using the method of moments (MoM) and laboratory measurements carried out in an anechoic chamber. In this experiment, a metallic plate with slits was used as a simple surface crack model and the root mean square error and maximum error of the angular estimation were 1.9-degree and 5.1-degree, respectively.
Interferometric Synthetic Aperture Radar (InSAR) uses the pairwise phase difference of SAR images to estimate the amount of deformation and displacement of the earth's surface. InSAR technique is used to estimate the displacement caused by earthquakes, volcanoes, subsidence, and settlement. InSAR has good spatial accuracy and density compared to other methods of estimating surface changes, such as the global positioning system (GPS). It is also economically acceptable due to its time and cost savings.
Interpreting the results of InSAR requires the processing of SAR images by SAR image processing software. GMTSAR software is written the C Programming Language by Sandwell, D. et al. This software has high speed in processing SAR images and can process SAR products such as Envisat, ALOS-1, TerraSAR-X, COSMOS-SkyMed, Sentinel-1, and ALOS-2. This software is open-source (GNU General Public License) that can be used to process Small Baseline Subset (SBAS) time-series
This book is written in two major general chapters. In the first chapter, we will show step by step how to install GMTSAR software and other software on Ubuntu 20.04 is taught. Also, in this chapter, download the necessary data such as Sentinel-1 images, and DEM is taught. In the second part, which is the main chapter of the book, the processing of the Sentinel-1 images with GMTSAR software is explained. First of all, the process of processing two earthquake-related images using the DInSAR technique is explained so that the user can obtain the amount of displacement caused by the earthquake with this technique. The SBAS time-series is then processed using GMTSAR software to estimate time series ground deformation determine the rate of subsidence of the earth's crust.
Polarimetric synthetic aperture radar (PolSAR) image classification is one of the hottest issues in remote sensing, where studies on pixel-level information and relationship are of great significance. In this article, graph convolutional network (GCN) is employed to accomplish this pixel-level task benefiting from its excellent capability in structure exploration and information propagation between different pixels. To reduce the communication burden between various PolSAR pixels and high computational cost for the whole PolSAR image, an adaptive GCN (AdapGCN) consisting of pixel-centered subgraphs is proposed in this article. In the AdapGCN, a data-adaptive kernel and a spatial-adaptive kernel are introduced to, respectively, model data structure and spatial structure for PolSAR image. Moreover, a multiscale learning structure is integrated to further explore complicated relations between pixels. Extensive comparative evaluations validate the superiority of our new AdapGCN model for PolSAR image classification over a wide range of state-of-the-art methods on three challenging benchmarks.
Based on experimental results, this chapter describes applications of SAR polarimetry to extract relevant information on agriculture and wetland scenarios by exploiting differences in the polarimetric signature of different scatterers, crop types and their development stage depending on their physical properties. Concerning agriculture, crop type mapping, soil moisture estimation and phenology estimation are reviewed, as they are ones with a clear benefit of full polarimetry over dual or single polarimetry. For crop type mapping, supervised or partially unsupervised classification schemes are used. Phenology estimation is treated as a classification problem as well, by regarding the different stages as different classes. Soil moisture estimation makes intensive use of scattering models, in order to separate soil and vegetation scattering and to invert for soil moisture from the isolated ground component. Then, applications of SAR polarimetry to wetland monitoring are considered that include the delineation of their extent and their characterisation by means of polarimetric decompositions. In the last section of the chapter, the use of a SAR polarimetric decomposition is shown for the assessment of the damages consequential to earthquakes and tsunamis.
This chapter critically summarizes the main theoretical aspects necessary for a correct processing and interpretation of the polarimetric information towards the development of applications of synthetic aperture radar (SAR) polarimetry. First of all, the basic principles of wave polarimetry (which deals with the representation and the understanding of the polarization state of an electromagnetic wave) and scattering polarimetry (which concerns inferring the properties of a target given the incident and the scattered polarized electromagnetic waves) are given. Then, concepts regarding the description of polarimetric data are reviewed, covering statistical and scattering aspects, the latter in terms of coherent and incoherent decomposition techniques. Finally, polarimetric SAR interferometry and tomography, two acquisition modes that enable the extraction of the 3-D scatterer position and separation, respectively, and their polarimetric characterization, are described.
Fire is one of the main factors leading to forest degradation in the Amazon by changing
species composition, biomass and structure. Considering the (potential) geographical
scale of forest fires, remote sensing provides essential data for mapping, monitoring and
even modeling. Despite radar sensors gather information on forest structure, no study
based on such data mainly focused on fire affected areas in the Brazilian Amazon. The
study area was situated in the “Arc of Fire”, in the State of Roraima (Northern
Amazonia), where 50 plots (0.25 ha each) were set out. We aimed to model the aboveground biomass (AGB) as a function of polarimetric attributes and to characterize the
forest degradation from these attributes. The impacts of understory fires on species
composition, stand structure and AGB were also assessed. Field and polarimetric data
analyses provided concordant results, revealing that frequent fires promoted a
‘secondarization’ of primary forests. Tree species diversity was significantly reduced
after recurrent fires. Even after a 12 years post-fire of a single fire, some units were still
dominated by Cecropia spp. Significant differences were found on stand structure
(density, height, basal area) between fire degradation levels, particularly in thrice
burned forests. The AGB stocks were reduced by 60% and similar to values previously
reported for secondary forests in the same region. Forests that experimented frequent
fire showed lower entropy and the presence of surface scattering was emphasized in the
H/ plane. The polarimetric responses indicated the dominance of the VV polarization
scattering in primary and lightly burned forests, whereas a dominance of the scattering
in HH polarization was noted in heavily and frequent burned forests. The results
suggests that polarimetric coherent and incoherent attributes (ψ2, A, Pd, VSI) are both
important in modeling AGB in forests characterized by a fire history (R²=0.76;
RMSE=32.1 Mg.ha-1 or 27.8% of the mean). No saturation point was detected for the
estimates as the AGB values were predicted up to 300 Mg.ha-1. This was attributed to
the inclusion of coherent parameters in the model. The model cross-validation (leaveone-out) showed a RMSE equal to 36.6 Mg.ha-1 (31% of the mean). We found an
adjusted R² of 0.7 and a RMSE of 32.45 Mg.ha-1 (23% of the mean) in the model
validation from independent set of samples (hold-out). We thus confirmed that the fullpolarimetric data used in biomass studies are sensitive to the fire degradation level. The
high vulnerability of Northern Amazonia tropical forests to fire was also verified. Finally, we underlined the importance of field data collection and analyze as they
provide valuable insights to understand the target itself and how it interacts with the
radar microwaves.
People have realized and been studying electromagnetic phenomena for thousands of years. However, electricity and magnetism have been treated independently for a long time, and knowledge has never been beyond the scope of qualitative understanding. Until the middle of the nineteenth century, the British physicist J. C. Maxwell inherited the field method of Faraday’s electromagnetism, and creatively proposed the concept of displacement current. Therefore, the macroscopic electromagnetic field equations—Maxwell’s equations and the electromagnetic wave theory of light are established. His achievements laid the foundation for the study of electromagnetic field theory and opened up a new era of knowing electromagnetic wave and making benefit for human society.
ResearchGate has not been able to resolve any references for this publication.