Characterizing L-Band Scattering of Paddy Rice in Southeast China With Radiative Transfer Model and Multitemporal ALOS/PALSAR Imagery

Rice is a major food supply in southeast China. With increased population and urbanization, reliable rice mapping is critical in this region. Because of frequent cloud cover and precipitation during the rice-growing season, it is difficult to conduct large-area rice monitoring with optical remote sensing techniques. L-band synthetic aperture radar (SAR), with its all-weather day and night imaging and canopy penetration capabilities, provides a unique alternative. In this study, a first-order radiative transfer model was developed to simulate L-band scattering properties of paddy rice. Three Advanced Land Observing Satellite (ALOS)/Phased-Array-Type L-band Synthetic Aperture Radar (PALSAR) images in dual-polarization mode (HH and HV) acquired in early tillering (June 28, 2007), tillering (August 13, 2007), and heading (September 28, 2007) stages were processed to test the temporal variation of rice backscatter. It was found that plant height and leaf mass amount were the two major structural parameters that contributed to rice backscatter in PALSAR images. The variation of the simulated HH backscatter matched with PALSAR observations in sample fields, although the simulated backscatter coefficients were around 3 dB lower than image-extracted values. Leaf volume scattering and leaf–ground double bounce were found as the two major scattering components in L-band HH polarization and increased with leaf layer height and density. This paper demonstrated that L-band HH backscatter was more sensitive to rice's structural variation than the VV backscatter and may therefore be more useful in rice mapping and modeling studies.      

Universal multifractal scaling of synthetic aperture radar imagesof sea-ice

Multifrequency, multipolarization imaging radar scattering coefficient data sets, acquired by synthetic aperture radar (SAR) over sea-ice, were studied in order to reveal their scale-invariant properties. Two distinct scenes were acquired at C-band (5.6 cm) and L-band (25 cm) wavelengths for three different linear polarizations (HH, VV, and HV). These sea-ice radar scattering coefficient fields were investigated by applying both Fourier and multifractal analysis techniques. The (multi) scaling of the data is clearly exhibited in both scenes for all three polarizations at L-band and for the HV polarization at C-band. The fields presenting this symmetry were found to be well described by universal multifractals. The corresponding parameters α, C₁, and H were determined for all these fields and were found to vary little with only the parameter H (characterizing the degree of nonconservation) displaying some systematic sensitivity to polarization. The values found for the universal multifractal parameters are α≈1.85±0.05, C₁≈0.0086±0.0041, and H≈-0.15±0.05     

A test statistic in the complex Wishart distribution and its application to change detection in polarimetric SAR data

When working with multilook fully polarimetric synthetic aperture radar (SAR) data, an appropriate way of representing the backscattered signal consists of the so-called covariance matrix. For each pixel, this is a 3/spl times/3 Hermitian positive definite matrix that follows a complex Wishart distribution. Based on this distribution, a test statistic for equality of two such matrices and an associated asymptotic probability for obtaining a smaller value of the test statistic are derived and applied successfully to change detection in polarimetric SAR data. In a case study, EMISAR L-band data from April 17, 1998 and May 20, 1998 covering agricultural fields near Foulum, Denmark are used. Multilook full covariance matrix data, azimuthal symmetric data, covariance matrix diagonal-only data, and horizontal-horizontal (HH), vertical-vertical (VV), or horizontal-vertical (HV) data alone can be used. If applied to HH, VV, or HV data alone, the derived test statistic reduces to the well-known gamma likelihood-ratio test statistic. The derived test statistic and the associated significance value can be applied as a line or edge detector in fully polarimetric SAR data also.     

Modeling L-band radar backscatter of Alaskan boreal forest

Synthetic aperture radar (SAR) data were acquired over Bonanza Creek Experimental Forest (Alaska) in March 1988 under thawed and frozen conditions. For five stands analyzed, L-band backscatter at 42°-45° incidence angle was 2.7-6.9 dB smaller under frozen than under thawed conditions for white spruce and balsam poplar, with the largest difference at HV and the smallest at HH polarization. The differences were smaller for a stand of small black spruce. The VV-HH phase differences observed by SAR were ≈0° for all the stands. Ground data were used to parameterize the Santa Barbara canopy backscatter model. For the white spruce and balsam poplar stands under thawed conditions, simulations agreed with the SAR data within the calibration uncertainty. The model underestimated the HH, HV, and VV backscatter for all five stands under frozen conditions, and for the black spruce stand under thawed conditions. The modeled VV-HH phase differences were close to 0° for all the stands except the black spruce stand. The discrepancies in model predictions of backscatter and phase difference were attributed to inadequate surface backscatter modeling. Model results supported the hypothesis that the weaker backscatter from frozen stands was because of the smaller dielectric constant of the frozen trees     

Radiative transfer modeling of cross-polarized backscatter from a pine forest using the discrete ordinate and eigenvalue method

Radiative transfer models have been widely used to interpret the radar backscatter from forested areas. Most of these models are based on an iterative solution of the radiative transfer equation, usually solved up to first or second order, thus taking into account single and double scattering. Although this method leads to results agreeing well with copolarized backscatter measurements, it produces less accurate estimates for horizontal-vertical (HV) polarization. This paper presents a radiative transfer backscatter model that accounts for multiple scattering by using the discrete ordinate and eigenvalue method applied to a layered medium. Using parameters derived from an architectural tree model, calculations at C- and L-band are compared with HV data acquired for a maritime pine forest in the southwest of France during the Spaceborne Imaging Radar-C missions. Good agreement is found at C-band for all values of forest biomass, and reasonable agreement at L-band for high biomass, when the soil backscatter plays a minor role. For low biomass, the L-band modeling is inadequate because of difficulties in estimating the soil backscatter. Comparison with calculations from a first-order radiative transfer model shows that multiple scattering is significant, especially at C-band.     

Radar backscattering model for multilayer mixed-species forests

A multilayer canopy scattering model is developed for mixed-species forests. The multilayer model provides a significantly enhanced representation of actual complex forest structures compared to the conventional canopy-trunk layer models. Multilayer Michigan Microwave Canopy Scattering model (Multi-MIMICS) allows overlapping layer configuration and a tapered trunk model applicable to forests of mixed species and/or mixed growth stages. The model is the first-order solution to a set of radiative transfer equations and includes layer interactions between overlapping layers. It simulates SAR backscattering coefficients based on input dimensional, geometrical, and dielectric variables of forest canopies. The Multi-MIMICS is an efficient realization of actual forest structures and can be shaped for specific interest of forest parameters. We present the model's application and validation in the paper. The model is parameterized using data collected from a 220,000-ha area of forests in central Queensland, Australia. Fifteen 50/spl times/50 m test sites representing the general forest diversity and growth stages are chosen as ground truth. Polarimetric backscattering airborne SAR (AIRSAR) data of the same area are acquired to validate the model simulations. The model predicts SAR backscattering coefficients of the test areas. Simulation results show a good agreement with AIRSAR data at most frequencies and polarizations. The simulated backscattering coefficient from the multilayer model and the standard MIMICS are also compared and significant improvements are observed.     

Stand age retrieval in production forest stands in New Zealand using C- and L-band polarimetric Radar

Regressions of single-, dual-, quad-, and full-polarization L- and C-band synthetic aperture radar (SAR) against stand age from 403 radiata pine stands in Kaingaroa Forest, New Zealand have been carried out, using the National Aeronautics and Space Administration's Airborne SAR instrument. The regressions attempted to find products suitable for the separation of young (two years or less) from old stands (25 years or older), and for the estimation of stand age. Local incidence angle had no significant effect for C-band, but was always significant for L-band, giving a standard error reduction of 13% to 32% for log stand age. Stand density was highly significant for both bands, giving standard error reductions of 7% to 47%. Single- and dual-polarization products were severely biased, and it was impossible to separate young and old stands, except L-band horizontal-(HH)-plus-horizontal-vertical (HV). C-band quad-polarization gave less bias and lower error than for that L-band, when local incidence angle and stand density were excluded. C-band full-polarization using covariance magnitudes gave no improvement over C-band quad-polarization, but L-band did give a significant improvement. The C- and L-band full-polarization products with six polarimetric indices gave significant improvements in the standard error. The results show that regressions of SAR data with stand age are possible with full- and quad-polarization L- and C-band datasets, although the prediction limits increase rapidly with stand age. The smallest error in estimated stand age, with an RMS of 3.22 years, was for L-band full-polarization with six polarimetric indices, calculated from a validation dataset. Separation of young and old forest stands was only possible for full-, quad-polarization, and the L-band HH-plus-HV products.     

Contrast Enhancement Method of POLSAR Images Based on Target Decomposition Theory

Improving the target-clutter ratio(TCR) of moving targets in synthetic aperture radar(SAR),imagery is very important for target detection and identification.In this paper,using the Cloude’s decomposition theory,an average covariance matrix can be decomposed into a summation of matrices representing three different scattering processes:the single bounce scattering,double bounce scattering,and diffuse scattering.A new idea of using the combination of the three components to enhance the contrast of an image is proposed.In order to compare with the polarimetric contrast enhancement method based on HH,HV,and VV data,ship areas of two combinatorial intensity images are detected by image binarization.Experimental results show that the method proposed in this paper provides better contrast.     

A three-component scattering model for polarimetric SAR data

An approach has been developed that involves the fit of a combination of three simple scattering mechanisms to polarimetric SAR observations. The mechanisms are canopy scatter from a cloud of randomly oriented dipoles, evenor double-bounce scatter from a pair of orthogonal surfaces with different dielectric constants and Bragg scatter from a moderately rough surface. This composite scattering model is used to describe the polarimetric backscatter from naturally occurring scatterers. The model is shown to describe the behavior of polarimetric backscatter from tropical rain forests quite well by applying it to data from NASA/Jet Propulsion Laboratory's (JPLs) airborne polarimetric synthetic aperture radar (AIRSAR) system. The model fit allows clear discrimination between flooded and nonflooded forest and between forested and deforested areas, for example. The model is also shown to be usable as a predictive tool to estimate the effects of forest inundation and disturbance on the fully polarimetric radar signature. An advantage of this model fit approach is that the scattering contributions from the three basic scattering mechanisms can be estimated for clusters of pixels in polarimetric SAR images. Furthermore, it is shown that the contributions of the three scattering mechanisms to the HH, HV, and VV backscatter can be calculated from the model fit. Finally, this model fit approach is justified as a simplification of more complicated scattering models, which require many inputs to solve the forward scattering problem     

On the detection of Faraday rotation in linearly polarized L-band SAR backscatter signatures

The potentially measurable effects of Faraday rotation on linearly polarized backscatter measurements from space are addressed. Single-polarized, dual-polarized, and quad-polarized backscatter measurements subject to Faraday rotation are first modeled. Then, the impacts are assessed using L-band polarimetric synthetic aperture radar (SAR) data. Due to Faraday rotation, the received signal will include other polarization characteristics of the surface, which may be detectable under certain conditions. Model results are used to suggest data characteristics that will reveal the presence of Faraday rotation in a given single-polarized, dual-polarized, or quad-polarized L-band SAR dataset, provided the user can identify scatterers within the scene whose general behavior is known or can compare the data to another, similar dataset with zero Faraday rotation. The data characteristics found to be most sensitive to a small amount of Faraday rotation (i.e., a one-way rotation <20/spl deg/) are the cross-pol backscatter [/spl sigma//spl deg/(HV)] and the like-to-cross-pol correlation [e.g., /spl rho/(HHHV/sup */)]. For a diverse, but representative, set of natural terrain, the level of distortion across a range of backscatter measures is shown to be acceptable (i.e., minimal) for one-way Faraday rotations of less than 5/spl deg/, and 3/spl deg/ if the radiometric uncertainty in the HV backscatter is specified to be less than 0.5 dB.     

Dependence of radar backscatter on coniferous forest biomass

Two independent experimental efforts have examined the dependence of radar backscatter on above-ground biomass of monospecie conifer forests using polarimetric airborne SAR data at P-, L- and C-bands. Plantations of maritime pines near Landes, France, range in age from 8 to 46 years with above-ground biomass between 5 and 105 tons/ha. Loblolly pine stands established on abandoned agricultural fields near Duke, NC, range in age from 4 to 90 years and extend the range of above-ground biomass to 560 tons/ha for the older stands. These two experimental forests are largely complementary with respect to biomass. Radar backscatter is found to increase approximately linearly with increasing biomass until it saturates at a biomass level that depends on the radar frequency. The biomass saturation level is about 200 tons/ha at P-band and 100 tons/ha at L-band, and the C-band backscattering coefficient shows much less sensitivity to total above-ground biomass     

The relationship between the backscattering coefficient and thebiomass of narrow and broad leaf crops

The influence of the shape and dimensions of plant constituents on the backscattering of agricultural vegetation is investigated. Multifrequency multitemporal polarimetric data, collected at C- and L-bands by means of airborne and satellite synthetic aperture radar (SAR), showed that the relations between the backscattering of crops and the vegetation biomass depend on plant type, and that there are different trends for “narrow” and “broad” leaf crops. In the latter crops, backscattering increases with an increase in the biomass, especially at L-band. This behavior is typical of media in which scattering is dominant, whereas on “narrow leaf” plants, the trend is flat or decreasing, denoting a major contribution of absorption. Theoretical simulations obtained with a discrete element radiative transfer model have confirmed that a different backscattering of crops with the same biomass may be due to plant geometry     

一种基于集成学习和特征融合的遥感影像分类新方法

针对多源遥感数据分类的需要,提出了一种基于全极化SAR影像、极化相干矩阵特征、光学遥感影像光谱和纹理的多种特征融合和多分类器集成的遥感影像分类新方法.对全极化PALSAR数据进行预处理和极化相干矩阵特征提取,利用灰度共生矩阵计算光学和SAR影像的对比度、逆差距、二阶距、差异性等纹理特征参数,并与光谱特征结合,形成6种组合策略.利用集成学习方法对随机森林分类器、子空间分类器、最小距离分类器、支持向量机分类器、反向传播神经网络分类器等分类器进行组合,对不同组合策略的遥感影像特征集进行分类.结果表明提出的基于多种特征和多分类器集成的新方法很好地利用了主被动遥感数据在不同地表景观类型提取上的潜力,综合了多种算法的优势,能够有效地提高总体精度和各类别的分类精度.     

A three-dimensional radar backscatter model of forest canopies

A three-dimensional forest backscatter model, which takes full account of spatial position of trees in a forest stand is described. A forest stand was divided into cells according to arbitrary spatial resolution. The cells may include “crown”, “trunk”, and “gap” components, determined by the shape, size and position of the trees. The forest floor is represented by a layer of “ground” cells. A ray tracing method was used to calculate backscattering components of 1) direct crown backscatter, 2) direct backscattering from ground, 3) direct backscattering from trunk, 4) crown-ground scattering, and 5) trunk-ground scattering. Both the attenuation and time-delay of microwave signals within cells other than “gap” were also calculated from ray tracing. The backscattering Mueller matrices of these components within the same range intervals were incoherently added to yield the total backscattering of an image pixel. By assuming a zero-mean, multiplicative Gaussian noise for image speckle, the high-resolution images were aggregated to simulate a SAR image with a given spatial resolution and number of independent samples (looks). A well-characterized 150 m×200 m forest stand in Maine, USA, was used to parameterize the model. The simulated radar backscatter coefficients were compared with actual JPL SAR data. The model gives reasonable prediction of backscattering coefficients averaged over the entire stand with agreement between model and data within 1.35 dB for all channels. The correlations between simulated images and SAR data (10 by 15 pixels) were positive and significant at the 0.001 level for all frequencies (P, L, and C bands) and polarizations (HH, HV, and VV)     

Potential Application of Multipolarization SAR for Pine-Plantation Biomass Estimation

This paper presents the techniques and the potential utility of multipolarization Synthetic Aperture Radar (SAR) data for pineplantation biomass estimation. Three channels of SAR data, one from the Shuttle Imaging Radar SIR-A and the other two from the aircraft SAR, were acquired over the Baldwin County, Alabama, study area. The SIR-A data were acquired with HH polarization and the aircraft SAR data with VV and VH polarizations. Linear regression techniques are used to estimate the pine-plantation biomass, tree height, and age using 21 test plots. The results indicate that the multipolarization data are highly related to the plantation biomass. The results suggest a potential application of multipolarization SAR for pine-plantation biomass estimation.     

A phase signature for detecting wet subsurface structures using polarimetric L-band SAR

In this paper, we investigate the ability of L-band synthetic aperture radar (SAR) systems to penetrate soils to retrieve information about subsurface wet structures. Our experiment site, the Pyla dune, is a bare sandy area allowing high radar penetration and known to have large wet subsurface structures (paleosoils) at varying depths. Buried paleosoils, which act as moisture tanks, are detectable with radar, since they present a high permittivity due to their water content. By analyzing airborne polarimetric SAR data, we established that a phase signature is correlated to the buried wet palesoils: a phase difference of 23/spl deg/ between the horizontal (HH) and vertical (VV) channels was clearly observed. It allows detection of the paleosoil down to a larger depth (5.2 m) than when only considering HH and HV amplitude signals (3.5 m). In order to confirm this result, field measurements were performed that led to the same observed phase difference. We could fit our observations to the semiempirical model proposed by Oh and Sarabandi, and we reproduced the observed phenomenon using a two-layer integral equation method (IEM) model of the Pyla dune, which was completed by finite-difference time-domain (FDTD) numerical simulations. We show that the soil moisture significantly influences the radar response in terms of phase difference between the copolarized modes. Our study also shows that the single-scattering IEM model reproduces the observed phase difference fairly well for a natural outdoor site when combined to FDTD simulation results. This phase signature could be used as a new tool to map subsurface moisture in arid regions.     

Low VHF-band backscatter from coniferous forests on sloping terrain

Low-frequency synthetic aperture radar (SAR) is a promising technique for stem volume retrieval, particularly for dense forests, due to the good penetration of forest canopies. However, it is well known that the dominant scattering mechanism, the trunk-ground dihedral interaction, decreases rapidly on sloping terrain. In this paper, we use low VHF-band SAR data, collected with CARABAS over dense coniferous forests in Sweden, to examine the effect of topography. Using flight passes with different headings, the effect of slope and aspect angle on backscatter is characterized. For tall trees (/spl sim/30 m), on the steepest slopes in the test-site (up to /spl sim/12/spl deg/), differences of up to 8 dB are observed between images acquired with different look directions relative to the slope. A physical model is developed to investigate the different scattering mechanisms and their sensitivity to terrain slopes. The model shows that the trunk-ground scattering still dominates the response for large trees on moderate slopes, and a semiempirical model for the effect of topography on backscatter is proposed. The model shows good agreement with measurements, indicating the possibility of using it to compensate for the effects of sloping terrain when retrieving stem volume in coniferous forest.     

Retrieval of ocean wave spectra and RAR MTF's fromdual-polarization SAR data

A method for the retrieval of the real aperture radar (RAR) modulation transfer function (MTF) and ocean wave spectra from dual-polarization (i.e., simultaneously acquired HH and VV polarizations) synthetic aperture radar (SAR) image data is described. The RAR MTF is estimated by applying empirical MTF estimation methodologies to inter-look cross spectra between various combinations of individual looks and available polarizations for a given radar frequency. The concept behind the nonlinear inversion is that any combination of like- and cross-polarization image spectra should return the same wave spectrum, in agreement with in situ and model wave spectra. This permits estimation of the RAR MTF on a case-by-case basis. The results are compared with theoretical treatments of the RAR MTF, which are shown to be inadequate for the range of conditions encountered in their data set. However, the theory and measurements fit well in describing the polarization dependence of the RAR MTF. The data set consists of SIR-C/X-SAR L-band and CCRS CV-580 C-band SAR data, in situ buoy measurements, and model data from field programs in Canadian waters in October and December 1994     

Four-component scattering model for polarimetric SAR image decomposition

A four-component scattering model is proposed to decompose polarimetric synthetic aperture radar (SAR) images. The covariance matrix approach is used to deal with the nonreflection symmetric scattering case. This scheme includes and extends the three-component decomposition method introduced by Freeman and Durden dealing with the reflection symmetry condition that the co-pol and the cross-pol correlations are close to zero. Helix scattering power is added as the fourth component to the three-component scattering model which describes surface, double bounce, and volume scattering. This helix scattering term is added to take account of the co-pol and the cross-pol correlations which generally appear in complex urban area scattering and disappear for a natural distributed scatterer. This term is relevant for describing man-made targets in urban area scattering. In addition, asymmetric volume scattering covariance matrices are introduced in dependence of the relative backscattering magnitude between HH and VV. A modification of probability density function for a cloud of dipole scatterers yields asymmetric covariance matrices. An appropriate choice among the symmetric or asymmetric volume scattering covariance matrices allows us to make a best fit to the measured data. A four-component decomposition algorithm is developed to deal with a general scattering case. The result of this decomposition is demonstrated with L-band Pi-SAR images taken over the city of Niigata, Japan.