The WindSat spaceborne polarimetric microwave radiometer: sensor description and early orbit performance

The global ocean surface wind vector is a key parameter for short-term weather forecasting, the issuing of timely weather warnings, and the gathering of general climatological data. In addition, it affects a broad range of naval missions, including strategic ship movement and positioning, aircraft carrier operations, aircraft deployment, effective weapons use, underway replenishment, and littoral operations. WindSat is a satellite-based multifrequency polarimetric microwave radiometer developed by the Naval Research Laboratory for the U.S. Navy and the National Polar-orbiting Operational Environmental Satellite System Integrated Program Office. It is designed to demonstrate the capability of polarimetric microwave radiometry to measure the ocean surface wind vector from space. The sensor provides risk reduction for the development of the Conical Microwave Imager Sounder, which is planned to provide wind vector data operationally starting in 2010. WindSat is the primary payload on the Department of Defense Coriolis satellite, which was launched on January 6, 2003. It is in an 840-km circular sun-synchronous orbit. The WindSat payload is performing well and is currently undergoing rigorous calibration and validation to verify mission success.     

微波辐射计反演海表面温度和风场研究进展

海表面温度（SST）和海面风场是重要的海面气象水文参数,与海面的大部分物理过程关系密切。卫星遥感探测技术可以同时获得全球范围内的海表面温度和海面风场数据。其中,微波辐射计遥感海表面温度和风场具有不受云层遮挡的优势,本文就其发展过程作简要介绍。     

Deep-space calibration of the WindSat radiometer

The WindSat microwave polarimetric radiometer consists of 22 channels of polarized brightness temperatures operating at five frequencies: 6.8, 10.7, 18.7, 23.8, and 37.0 GHz. The 10.7-, 18.7-, and 37.0-GHz channels are fully polarimetric (vertical/horizontal, /spl plusmn/45/spl deg/ and left-hand and right-hand circularly polarized) to measure the four Stokes radiometric parameters. The principal objective of this Naval Research Laboratory experiment, which flys on the USAF Coriolis satellite, is to provide the proof of concept of the first passive measurement of ocean surface wind vector from space. This paper presents details of the on-orbit absolute radiometric calibration procedure, which was performed during of a series of satellite pitch maneuvers. During these special tests, the satellite pitch was slowly ramped to +45/spl deg/ (and -45/spl deg/), which caused the WindSat conical spinning antenna to view deep space during the forward (or aft portion) of the azimuth scan. When viewing the homogeneous and isotropic brightness of space (2.73 K) through both the main reflector and the cold-load calibration reflector, it is possible to determine the absolute calibration of the individual channels and the relative calibration bias between polarimetric channels. Results demonstrate consistent and stable channel calibrations (with very small brightness biases) that exceed the mission radiometric calibration requirements.     

全极化微波辐射计数字相关器下变频算法设计

全极化微波辐射计是一种用于海洋表面风场测量的新型被动微波遥感器,数字相关器是全极化辐射计的核心部件。针对全极化微波辐射计接受到的数字信号采样率较高的问题,提出应用多相滤波的算法以降低数字信号速率,从而满足FPGA时序约束的要求。之后为进一步节省乘法器资源和提高处理速率,滤波部分采用分布式算法,两种方法结合完成下变频模块处理,通过Matlab软件对下变频算法进行了仿真。仿真结果表明算法是可行的。     

Evaluation of WindSat wind vector performance with respect to QuikSCAT estimates

The WindSat instrument was launched on January 6, 2003 as part of a risk reduction effort to assess the potential of using spaceborne fully polarimetric radiometry to measure the marine wind vector. Microwave radiometry on the Special Sensor Microwave/Imager onboard the Defense Meteorological Satellite Program satellites has long provided wind speed measurements. Fully polarimetric radiometry offers the additional possibility of obtaining wind direction as well. By contrast, the QuikSCAT satellite uses active microwave measurements to estimate the wind vector from space. It represents the most comprehensive satellite dataset against which to compare WindSat measurements. In this paper, we systematically compare temporally and spatially coincident WindSat and QuikSCAT wind vector measurements against the design goals of the WindSat instrument, taking into consideration expected differences related to instrument precision and the spatial and temporal variability of the wind field.     

Fully polarimetric microwave radiometer for remote sensing

The design, characteristics, and operation of the Helsinki University of Technology Fully Polarimetric Radiometer (FPoR) are described. The developed 36.5-GHz radiometer can be used for airborne remote sensing; however, ground-based and laboratory measurements are also possible. A direct cross-correlation technique with analog correlators, which measures all four Stokes parameters simultaneously, is applied. This paper is the first successful demonstration of an analog direct cross-correlation technique for polarimetric remote sensing radiometry. The radiometer was subjected to a variety of laboratory tests, and considerable attention is given to analysis of the characteristics of the instrument. Owing to the effective active temperature control system of the receiver, the radiometric stability of the instrument was found to be very high; test results showing stabilities below 10 mK and of 4-40 mK on time scales of 800 and 8000 s, respectively, are presented. Furthermore, the absolute accuracy of the system is analyzed to be at a sub-Kelvin level for most measurement conditions. A maritime wind vector experiment was carried out over the Gulf of Finland. The feasibility and performance of the applied correlation technique and the whole radiometer system were verified for fully polarimetric airborne measurements. The obtained brightness temperatures of the first three Stokes parameters show typical harmonic behavior with respect to the surface wind; the results suggest, however, that the model coefficients presented earlier for oceans may not be directly applicable for different conditions.     

Soil moisture retrieval using the C-band polarimetric scanning radiometer during the Southern Great Plains 1999 Experiment

The Advanced Microwave Scanning Radiometer (AMSR) holds promise for retrieving soil moisture in regions with low levels of vegetation. Algorithms for this purpose have been proposed, but none have been rigorously evaluated due to a lack of datasets. Accordingly, the Southern Great Plains 1999 Experiment (SGP99) was designed to provide C-band datasets for AMSR algorithm development and validation. Ground observations of soil moisture and related variables were collected in conjunction with aircraft measurements using a C-band radiometer similar to the AMSR sensor (6.92 GHz), the Polarimetric Scanning Radiometer with its C-band scanhead (PSR/C). The study region has been the focus of several previous remote sensing field experiments and contains vegetation conditions compatible with the expected capabilities of C-band for soil moisture retrieval. Flights were conducted under a wide range of soil moisture conditions, thus providing a robust dataset for validation. A significant issue found in data processing was the removal of anthropogenic radio-frequency interference. Several approaches to estimating the parameters of a single-channel soil moisture retrieval algorithm were used. PSR/C soil moisture images show spatial and temporal patterns consistent with meteorological and soil conditions, and the dynamic range of the PSR/C observations indicates that the AMSR instrument can provide useful soil moisture information.     

High-resolution passive polarimetric microwave mapping of oceansurface wind vector fields

The retrieval of ocean surface wind fields in both one and two dimensions is demonstrated using passive polarimetric microwave imagery obtained from a conical-scanning airborne polarimeter. The retrieval method is based on an empirical geophysical model function (GMF) for ocean surface thermal emission and an adaptive maximum likelihood (ML) wind vector estimator. Data for the GMF were obtained using the polarimetric scanning radiometer/digital (PSR/D) on the NASA P-3 aircraft during the Labrador Sea Deep Convection Experiment in 1997. To develop the GMF, a number of buoy overflights and GPS dropsondes were used, out of which a GMF of 10.7, 18.7, and 37.0 GHz azimuthal harmonics for the first three Stokes parameters was constructed for the SSM/I incident angle of 53.1°. The data show repeatable azimuthal harmonic coefficient amplitudes of ~2-3 K peak-to-peak, with a 100% increase in harmonic amplitudes as the frequency is increased from 10.7 to 37 GHz. The GMF is consistent with and extends the results of two independent studies of SSM/I data and also provides a model for the third Stokes parameter over wind speeds up to 20 m/s. The aircraft data show that the polarimetric channels are much less susceptible to geophysical noise associated with maritime convection than the first two Stokes parameters. The polarimetric measurement technique used in the PSR/D also demonstrates the viability of digital correlation radiometry for aircraft or satellite measurements of the full Stokes vector. The ML retrieval algorithm incorporates the additional information on wind direction available from multiple looks and polarimetric channels in a straightforward manner and accommodates the reduced SNRs of the first two Stokes parameters in the presence of convection by weighting these channels by their inverse SNR     

A new wind vector algorithm for C-band SAR

Ocean wind speed and wind direction are estimated simultaneously using the normalized radar cross sections /spl sigma//sup 0/ corresponding to two neighboring (25-km) blocks, within a given synthetic aperture radar (SAR) image, having slightly different incidence angles. This method is motivated by the methodology used for scatterometer data. The wind direction ambiguity is removed by using the direction closest to that given by a buoy or some other source of information. We demonstrate this method with 11 ENVISAT Advanced SAR sensor images of the Gulf of Mexico and coastal waters of the North Atlantic. Estimated wind vectors are compared with wind measurements from buoys and scatterometer data. We show that this method can surpass other methods in some cases, even those with insufficient visible wind-induced streaks in the SAR images, to extract wind vectors.     

The accuracy of preliminary WindSat vector wind measurements: comparisons with NDBC buoys and QuikSCAT

Two preliminary, six-month long global WindSat vector wind datasets are validated using buoys and QuikSCAT measurements. Buoy comparisons yield speed and direction root mean square accuracies of 1.4 m/s and 25/spl deg/ for the "NESDIS0" product and 1.3 m/s and 23/spl deg/ for the more recently produced "B1" product from the Naval Research Laboratory. WindSat along- and across-wind random component errors of 0.7-1.0 and 2.6-2.8 m/s (respectively) are larger than those calculated for QuikSCAT in the same period. Global WindSat-QuikSCAT comparisons generally confirmed the buoy analyses. While simple rain flags based directly on WindSat brightness temperature measurements alone are shown to overflag for rain systematically, the advanced "Environmental Data Record" rain flag in the B1 product matches well with Special Sensor Microwave/Imager rain detection frequency and preserves the accuracy of the unflagged vector wind measurements.     

ERS-1 scatterometer measurements. II. An algorithm forocean-surface wind retrieval including light winds

For pt.I see ibid., vol.36, no.2, p.603-22 (1998). An algorithm for retrieving European Remote Sensing Satellite (ERS-1) scatterometer winds, denoted the Rufenach-Bates-Tosini (RBT) algorithm, is developed and used to retrieve winds collocated within ±25 km of buoy measurements in two oceanic regions, equatorial and midlatitude. An improvement in the retrieved RBT winds over the European Space Agency (ESA) winds is due mainly to a geophysical model employing the full available wind-speed range, including the lightest winds. This model, denoted BMOD5, is tuned by using the scatterometer and buoy measurements, resulting in two different models for the midlatitude and equatorial regions. The RBT retrieved winds exhibit (1) a larger number of solutions (wind vectors) and (2) smaller biases in wind speed than the ESA wind product. The increase in the number of retrieved winds is primarily due to lighter winds employed, 0.2 m/s to 18 m/s; whereas, the ESA winds are truncated near 3 m/s. The ESA winds underestimate the highest winds significantly, by about 20%, and overestimate the lightest winds. The RBT wind bias is less than a few percent at the highest winds and a few tenths of a m/s at the lowest winds. Both algorithms retrieve 180° ambiguous directions almost as often as the true direction. Regression fits to the winds using the RBT algorithm produce standard deviations of 1 m/s and 25° near the equator for winds varying from 0.2-10 m/s and 1.2 m/s and 250 at midlatitudes for winds varying from 0.2-18 m/s, provided that the ambiguities are removed     

An algorithm for uniform vector quantizer design

A vector quantizer maps ak-dimensional vector into one of a finite set of output vectors or "points". Although certain lattices have been shown to have desirable properties for vector quantization applications, there are as yet no algorithms available in the quantization literature for building quantizers based on these lattices. An algorithm for designing vector quantizers based on the root latticesA_{n}, D_{n}, andE_{n}and their duals is presented. Also, a coding scheme that has general applicability to all vector quantizers is presented. A four-dimensional uniform vector quantizer is used to encode Laplacian and gamma-distributed sources at entropy rates of one and two bits/sample and is demonstrated to achieve performance that compares favorably with the rate distortion bound and other scalar and vector quantizers. Finally, an application using uniform four- and eight-dimensional vector quantizers for encoding the discrete cosine transform coefficients of an image at0.5bit/pel is presented, which visibly illustrates the performance advantage of vector quantization over scalar quantization.     

An algorithm for determining the radiance reflected from the roughsea surface using MODIS-N satellite radiometer data

A new algorithm is suggested for determining the radiance B_r (&thetas;, φ) reflected from the rough sea surface. The knowledge of B_r(&thetas;, φ) is needed for excluding the reflected solar radiation from the signal at the space receiver. Information on the reflected radiance is also required when estimating the state of the sea surface from space. The information (on reflected radiance) is important for problems such as estimating the chlorophyll concentration in sea water. The suggested algorithm was constructed based on direct measurements of the radiance coefficient ρ published in V. G. Akimov ei al. (1993) and J. A Shaw et al. (1997). The experimental results show that the slope distribution of sea surface elements depends not only on the wind speed near the sea surface but also on the stability of the lower part of the marine atmospheric boundary layer. It is shown that the failure to account for this effect causes an error of ±30% when estimating B_r(&thetas;, φ)     

An accelerated decomposition algorithm for robust support vector Machines

This paper proposes an accelerated decomposition algorithm for the robust support vector machine (SVM). Robust SVM aims at solving the overfitting problem when there is outlier in the training data set, which makes the decision surface less contoured and results in sparse support vectors. Training of the robust SVM leads to a quadratic optimization problem with bound and linear constraint. Osuna provides a theorem which proves that the Standard SVM's quadratic programming (QP) problem can be broken down into a series of smaller QP subproblems. This paper derives the Kuhn-Tucker condition and decomposition algorithm for the robust SVM. Furthermore, a pre-selection technique is incorporated into the algorithm to speed up the calculation. The experiment using standard data sets shows that the accelerated decomposition algorithm makes the training process more efficient.     

An L-band geophysical model function for SAR wind retrieval using JERS-1 SAR

An L-band geophysical model function is developed using Japanese Earth Resources Satellite-1 (JERS-1) synthetic aperture radar (SAR) data. First, we estimate the SAR system noise, which has been a serious problem peculiar to the JERS-1 SAR. It is found that the system noise has a feature common in all the SAR images and that the azimuth-averaged profile of noise can be expressed as a parabolic function of range. By subtracting the estimated noise from the SAR images, we can extract the relatively calibrated ocean signals. Second, using the noise-removed SAR data and wind vector data from the NASA Scatterometer and buoys operated by the Japan Meteorological Agency, we generate a match-up dataset, which consists of the SAR sigma-0, the incidence angle, the surface wind speed, and wind direction. Third, we investigate the sigma-0 dependence on incidence angle, wind speed, and wind direction. While the incidence angle dependence is negligible in the present results, we can derive distinct sigma-0 dependence on wind speed and direction. For wind speeds below 8 m/s, the wind direction dependence is not significant. However, for higher wind speeds, the upwind-downwind asymmetry becomes very large. Finally, taking into account these characteristics, a new L-band-HH geophysical model function is produced for the SAR wind retrieval using a third-order harmonics formula. Resultant estimates of SAR-derived wind speed have an rms error of 2.09 m/s with a negligible bias against the truth wind speed. This result enables us to convert JERS-1 SAR images into the reliable wind-speed maps.     

Wind direction over the ocean determined by an airborne, imaging,polarimetric radiometer system

The speed and direction of winds over the ocean can be determined by polarimetric radiometers. This has been established by theoretical work and demonstrated experimentally using airborne radiometers carrying out circle flights and thus measuring the full 360° azimuthal response from the sea surface. An airborne experiment, with the aim of measuring wind direction over the ocean using an imaging polarimetric radiometer, is described. A polarimetric radiometer system of the correlation type, measuring all four Stokes brightness parameters, is used. Imaging is achieved using a 1-m aperture conically scanning antenna. The polarimetric azimuthal signature of the ocean is known from modeling and circle flight experiments. Combining the signature with the measured brightness data from just a single flight track enables the wind direction to be determined on a pixel-by-pixel basis in the radiometer imagery     

一种减小语音矢量编码码率的算法

2维PV编码算法是一种将线性预测编码、SOM神经网络2维矢量编码以及Huffman编码相结合的语音信号编码算法。为了在保证译码恢复的语音质量良好的前提下,进一步减小编码的压缩率,以减小语音信号的存储空间,提出一种增加2维PV编码中矢量量化维数的高维PV编码算法。并利用Matlab软件编程进行2维、4维和8维PV算法的语音信号编解码实验。实验结果表明,在保证译码恢复声音质量良好的条件下,增加2维PV编码算法的量化矢量维数,可以减小码率,其中8维PV编码算法的码率最小为5.94 Kb/s,小于采用ADPCM编码算法的波形编码标准G.721的码率32 Kb/s(波形编码的最小码率),甚至小于采用LD-CELP编码算法的混合编码G.728的码率16 Kb/s。文中提出的编码算法在语言压缩编码方面将具有较高的研究价值和很好的应用前景。     

An efficient encoding algorithm for vector quantization based on subvector technique

In this paper, a new and fast encoding algorithm for vector quantization is presented. This algorithm makes full use of two characteristics of a vector: the sum and the variance. A vector is separated into two subvectors: one is composed of the first half of vector components and the other consists of the remaining vector components. Three inequalities based on the sums and variances of a vector and its two subvectors components are introduced to reject those codewords that are impossible to be the nearest codeword, thereby saving a great deal of computational time, while introducing no extra distortion compared to the conventional full search algorithm. The simulation results show that the proposed algorithm is faster than the equal-average nearest neighbor search (ENNS), the improved ENNS, the equal-average equal-variance nearest neighbor search (EENNS) and the improved EENNS algorithms. Comparing with the improved EENNS algorithm, the proposed algorithm reduces the computational time and the number of distortion calculations by 2.4% to 6% and 20.5% to 26.8%, respectively. The average improvements of the computational time and the number of distortion calculations are 4% and 24.6% for the codebook sizes of 128 to 1024, respectively.     

An analytical calibration approach for microwave polarimetric radiometers

We present an analytical calibration approach for passive microwave polarimeters that is applicable where the instrument can be partitioned into distinct, functional radio-frequency blocks. The methodology is focused on polarimetric system characterization, not polarimetric measurements. It requires characterization of each major internal functional subsystem with a vector network analyzer to obtain a closed-form transfer function. The goals of this approach are to provide a transfer function describing the system in its entirety and to isolate the contribution of each subsystem to the uncertainty in the final modified Stokes parameters. Notably, the approach does not assume ideal polarization isolation in the radiometer system. A significant benefit of this approach is that the cascaded transfer functions serve as a realistic instrument simulator revealing where improvements in component performance would have greatest benefit for system performance over the dynamic range of the instrument. This systems-focused approach is applied to the National Aeronautics and Space Administration Goddard Space Flight Center polarimetric Airborne C-band Microwave Radiometer (ACMR), whose architecture allows the necessary subsystem partitioning. The characteristics of each subsystem were extensively measured, converted to a transfer function, and imported into the overall closed-form system model. Inversion of the system model and error analysis inherent to this calibration approach are illustrated by a full Stokes parameter retrieval for a senescent cornfield.     

Multitraining support vector machine for image retrieval.

Relevance feedback (RF) schemes based on support vector machines (SVMs) have been widely used in content-based image retrieval (CBIR). However, the performance of SVM-based RF approaches is often poor when the number of labeled feedback samples is small. This is mainly due to 1) the SVM classifier being unstable for small-size training sets because its optimal hyper plane is too sensitive to the training examples; and 2) the kernel method being ineffective because the feature dimension is much greater than the size of the training samples. In this paper, we develop a new machine learning technique, multitraining SVM (MTSVM), which combines the merits of the cotraining technique and a random sampling method in the feature space. Based on the proposed MTSVM algorithm, the above two problems can be mitigated. Experiments are carried out on a large image set of some 20,000 images, and the preliminary results demonstrate that the developed method consistently improves the performance over conventional SVM-based RFs in terms of precision and standard deviation, which are used to evaluate the effectiveness and robustness of a RF algorithm, respectively.