SEARCH · Engineering Papers
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Report presents one in continuing series of studies of segmentation of polarimetric synthetic-aperture-radar, SAR, image data into regions. Studies directed toward refinement of method of automated analysis of SAR data.
SAR (Synthetic Aperture Radar) interferometry allows the measurement of high resolution topography of terrain from repeat orbit SAR data sets obtained by the ERS-1 AMI (Active Microwave Instrument). The system parameters which affect the accuracy of the topographic maps are errors in determination of the interferometric baseline, non parallel orbit tracks, decorrelation caused by baseline length, thermal noise, and surface change. Surface change can be observed in the interferograms either through decorrelation of the interferometric phase, or coherent phase shifts caused by locally uniform surface displacements. Phase gradient maps can be derived directly from the complex interferograms and can be transformed into surface slopes mapped onto a geometrically corrected grid.
Because of the needs of future projects like ENVISAT and the experiences made with the current operational ERS-1 facilities, a radical change in the synthetic aperture radar (SAR) processing scenarios can be predicted for the next years. At the German PAF several new developments were initialized which are driven mainly either by user needs or by system and operational constraints ('lessons learned'). At the end there will be a major simplification and uniformation of all used computer systems. Especially the following changes are likely to be implemented at the German PAF: transcription before archiving, processing of all standard products with high throughput directly at the receiving stations, processing of special 'high-valued' products at the PAF, usage of a single type of processor hardware, implementation of a large and fast on-line data archive, and improved and unified fast data network between the processing and archiving facilities. A short description of the current operational SAR facilities as well as the future implementations are given.
A set of considerations that are pertinent to future international cooperation in the area of synthetic aperture radar (SAR) ground data systems are presented. The considerations are as follows: (1) success of future spaceborne SAR missions will require multi-agency and/or multi-national collaboration; (2) ground processing is typically performed by each agency for their user base; (3) international standards are required to achieve a uniform data product independent of the processing center; (4) to reduce the aggregate cost of the ground data systems, collaboration is required in design and development; (5) effective utilization of the data by an international user community; (6) commercialization of data products; and (7) security of data systems.
A study is in progress comparing Airborne Synthetic Aperture Radar (AIRSAR) backscatter from coniferous forest plots containing gaps to backscatter from adjacent gap-free plots. Issues discussed are how do gaps in the range of 400 to 1600 sq m (approximately 4-14 pixels at intermediate incidence angles) affect forest backscatter statistics and what incidence angles, wavelengths, and polarizations are most sensitive to forest gaps. In order to visualize the slant-range imaging of forest and gaps, a simple conceptual model is used. This strictly qualitative model has led us to hypothesize that forest radar returns at short wavelengths (eg., C-band) and large incidence angles (e.g., 50 deg) should be most affected by the presence of gaps, whereas returns at long wavelengths and small angles should be least affected. Preliminary analysis of 1989 AIRSAR data from forest near Mt. Shasta supports the hypothesis. Current forest backscatter models such as MIMICS and Santa Barbara Discontinuous Canopy Backscatter Model have in several cases correctly predicted backscatter from forest stands based on inputs of measured or estimated forest parameters. These models do not, however, predict within-stand SAR scene texture, or 'intrinsic scene variability' as Ulaby et al. has referred to it. For instance, the Santa Barbara model, which may be the most spatially coupled of the existing models, is not truly spatial. Tree locations within a simulated pixel are distributed according to a Poisson process, as they are in many natural forests, but tree size is unrelated to location, which is not the case in nature. Furthermore, since pixels of a simulated stand are generated independently in the Santa Barbara model, spatial processes larger than one pixel are not modeled. Using a different approach, Oliver modeled scene texture based on an hypothetical forest geometry. His simulated scenes do not agree well with SAR data, perhaps due to the simple geometric model used. Insofar as texture is the expression of biological forest processes, such as succession and disease, and physical ones, such as fire and wind-throw, it contains useful information about the forest, and has value in image interpretation and classification. Forest gaps are undoubtedly important contributors to scene variance. By studying the localized effects of gaps on forest backscatter, guided by our qualitative model, we hope to understand more clearly the manner in which spatial heterogeneities in forests produce variations in backscatter, which collectively give rise to scene texture.
The Gulf of Mexico (GOM) exhibits a wealth of energetic ocean features; they include the Loop Current with velocities of about 2 m/s and strong shear fronts, mesoscale eddies, double vortices, internal waves, and the outflow of the 'Mighty Mississippi' river. These energetic features can have a strong impact on the economies of the states surrounding the Gulf. Large fisheries, oil and gas production as well as pollution transport are relevant issues. These circulation features in the Gulf are invisible to conventional IR and visible satellite imagery during the Summer months due to cloud cover and uniform surface temperatures. Synthetic Aperture Radar (SAR) imagery of the Gulf does penetrate the cloud cover and shows a rich assembly of features there year-round. Below are preliminary results from GOM SAR imagery taken by SEASAT in 1978 and by the AIRSAR program in 1991.
It is well known that under favorable meteorological and hydrodynamical conditions the bottom topography of shallow seas can be mapped with airborne or spaceborne imaging radar. This phenomenon was observed for the first time in 1969 by de Loor and co-workers in Q-band Side Looking Airborne Radar (SLAR) imagery of sandwaves in the North Sea. It is now generally accepted that the imaging mechanism consists of three steps: (1) interaction between (tidal) current and bottom topography causes spatial modulations in the surface current velocity; (2) modulations in the surface current velocity give rise to variations in the spectrum of wind-generated waves, as described by the action balance equation; and (3) variations in the wave spectrum show up as intensity modulations in radar imagery. In order to predict radar backscatter modulations caused by sandwaves, an imaging model, covering the three steps, was developed by the Dutch Sea Bottom Topography Group. This model and some model results will be shown. On 16 Aug. 1989 an experiment was performed with the polarimetric P-, L-, and C-band synthetic aperture radar (SAR) of NASA/JPL. One scene was recorded in SAR mode. On 12 Jul. 1991 another three scenes were recorded, of which one was in the ATI-mode (Along-Track Interferometer). These experiments took place in the test area of the Sea Bottom Topography Group, 30 km off the Dutch coast, where the bottom topography is dominated by sand waves. In-situ data were gathered by a ship in the test area and on 'Measuring Platform Noordwijk', 20 km from the center of the test area. The radar images made during the experiment were compared with digitized maps of the bottom. Furthermore, the profiles of radar backscatter modulation were compared with the results of the model. During the workshop some preliminary results of the ATI measurements will be shown.
A new technique for fitting a three-component scattering mechanism model to the polarimetric synthetic aperture radar (SAR) data itself, without utilizing any ground truth measurements, is presented. The three scattering mechanism components included in the model are volume scatter from randomly oriented dipoles, first-order Bragg surface scatter and a dihedral scattering mechanism for two surfaces with different dielectric constants. The model fit yields an estimate of the contribution to the total backscatter of each of the three components. The backscatter contributions can also be compared to give the relative percentage weight of each. The model fit has an equal number of input parameters (the polarimetric radar backscatter measurements) and output parameters (the backscatter parameters describing them). The model can be applied to entire images or to small areas within an image to give a first-order estimate of the relevant scattering mechanisms. The model was applied to many C-, L- and P-band Airborne SAR (AIRSAR) images of different types of terrain. Results were presented at the workshop.
Microwave remote sensing technique has a high potential for measuring soil moisture due to the large contrast in dielectric constant of dry and wet soils. Recent work by Pults et al. demonstrated the use of X/C-band data for quantitative surface soil moisture extraction from Airborne Synthetic Aperture Radar (SAR) system. Similar technique was adopted using polarimetric SAR data acquired with the JPL-AIRSAR system over the Mahantango watershed area in central Pennsylvania during July 1990. The data sets reported include C-, L-, and P-bands of 10, 13, 15, and 17 July 1990.
Previously, a scattering model was used to classify SAR data as due to three mechanisms: surface or odd-bounce scattering; double-bounce scattering; and volume scattering. A Bragg rough-surface model was used for odd-bounce scattering, and two reflecting surfaces, as in a dihedral corner reflector, represented the double-bounce scattering mechanism. A random collection of infinitely thin cylinders was used to model the volume scattering contribution. It was assumed that the cylinders are uniformly oriented in all directions. In this work, modifications are made to the above model, which include specifying a preferred orientation for the cylinders representing branches and using thin cylinders of finite thickness as opposed to infinitely thin ones. In the first case, the preferred orientation is represented by a probability density function in the form of a dirac delta function. For the second case, small argument asymptotic expansions of Bessel functions are used to express the volume scattering contributions in terms of quantities containing the average radius of the cylinders. Since in each case the number of unknowns exceeds that of equations, simplifying assumptions are made. The solutions are applied to multifrequency polarimetric SAR data to obtain the new unknowns in addition to the relative contributions of each scattering mechanism. Several results are presented. It is found that the calculated percentage of each mechanism does not change significantly compared to the original model; however, additional information is gained about the scatterers.
Noise equivalent sigma(sup 0) is a fundamental parameter in any synthetic aperture radar (SAR) measurement. Knowledge of the noise equivalent sigma(sup 0) allows the radar application scientist to band the error contributions due to noise present in his/her data. In this paper, a novel technique for determining noise equivalent sigma(sup 0) is presented. It has been observed that a pair of images obtained by a SAR over the same area in a nearly repeating orbit can demonstrate a high degree of correlation, depending on the baseline. If noise independence and signal power invariance are assumed, an estimate for the noise equivalent sigma(sup 0) for each image can be determined without any known ground targets. We examine the application of this technique to several images of different backscatter obtained by ERS-1.
Two types of C-band aircraft interferometric Synthetic Aperture Radar (SAR) are being developed at JPL to measure the ocean wave characteristics. Each type requires two identical antennas with each having a long rectangular aperture to radiate fan-shaped beam(s). One type of these radars requires each of its antennas to radiate a broadside beam that will measure the target's cross-track velocity. The other type, having each of its antennas to radiate two off-broadside pointed beams, will allow the measurement of both the cross-track and the along-track velocities of the target. Because flush mounting of the antenna on the aircraft fuselage is desirable, microstrip patch array is selected for these interferometric SAR antennas. To meet the radar system requirement, each array needs a total of 76 microstrip patches which are arranged in a 38 x 2 rectangular aperture with a physical size of 1.6m x 16.5cm. To minimize the insertion loss and physical real estate of this relatively long array, a combined series/parallel feed technique is used. Techniques to suppress cross-pol radiation and to effectively utilize the RF power are also implemented. Cross-pol level of lower than -30 dB from the co-pol peak and low insertion loss of 0.36 dB have been achieved for both types of arrays. For the type of radar that requires two off-braodside pointed beams, a simple phasing technique is used to achieve this dual-beam capability with adequate antenna gain (20 dBi) and sidelobe level (-14 dB). Both radar arrays have been flight tested on aircraft with excellent antenna performance demonstrated.
We describe a potential procedure for retrieving ice thickness from multi-frequency polarimetric SAR data for thin ice. This procedure includes first masking out the thicker ice types with a simple classifier and then deriving the thickness of the remaining pixels using a model-inversion technique. The technique used to derive ice thickness from polarimetric observations is provided by a numerical estimator or neural network. A three-layer perceptron implemented with the backpropagation algorithm is used in this investigation with several improved aspects for a faster convergence rate and a better accuracy of the neural network. These improvements include weight initialization, normalization of the output range, the selection of offset constant, and a heuristic learning algorithm. The performance of the neural network is demonstrated by using training data generated by a theoretical scattering model for sea ice matched to the database of interest. The training data are comprised of the polarimetric backscattering coefficients of thin ice and the corresponding input ice parameters to the scattering model. The retrieved ice thickness from the theoretical backscattering coefficients is compare with the input ice thickness to the scattering model to illustrate the accuracy of the inversion method. Results indicate that the network convergence rate and accuracy are higher when multi-frequency training sets are presented. In addition, the dominant backscattering coefficients in retrieving ice thickness are found by comparing the behavior of the network trained backscattering data at various incidence angels. After the neural network is trained with the theoretical backscattering data at various incidence anges, the interconnection weights between nodes are saved and applied to the experimental data to be investigated. In this paper, we illustrate the effectiveness of this technique using polarimetric SAR data collected by the JPL DC-8 radar over a sea ice scene.
An across-track interferometric (XTI) SAR has the capability of determining the three dimensional target location from the recorded radar echos thus providing high spatial resolution both horizontally and vertically. The output of an XTI SAR will typically include both a radar backscatter map and a digital elevation model (DEM). The accuracy of the derived DEM is determined by factors such as data acquisition geometry, signal-to-noise ratio, knowledge of platform position and attitude as well as the accuracy of the processing system. In the summer of 1992 the authors acquired TOPSAR data over the Ft. Irwin area in California, a desert area with significant relief (height standard deviation 150 m). Very accurate DEM's have been derived for this area by the Topographic Engineering Center (TEC) using digital correlation methods on 1:20,000 scale digitized photographs. Corner reflectors were deployed in the area, and their locations were determined to cm accuracies by the Defense Mapping Agency (DMA) using differential GPS techniques. DEM's generated from the acquired radar data have been rotated and translated to overlay the reference DEM's provided by TEC. A detailed description of the errors and their characteristics will be given. The standard deviations of the radar data sets we acquired measured over a 5.6 by 7 km area were 1.9 and 2.3 m. The corresponding numbers for flat areas were 1.1 and 2.0 m and for mountain areas 3.3 and 2.2 m.
In order to separate the effects of the different tectonic and climatic processes on the shapes of desert piedmonts, a modified conic equation was fitted to digital topographic data for individual alluvial fans in Death Valley (California, U.S.). The topographic data were obtained from a SPOT panchromatic stereo pair and from the airborne interferometric SAR (Synthetic Aperture Radar) (TOPSAR). The conic fit allows parameters for the epex position, slope, and radial curvature to be compared with unit age, uplift rate, and climatic conditions. Preliminary results indicate that slope flattens with age and radial curvature is concave up, but decreases with age. Work is continuing on correlation of fit residuals and apex position with fan unit age. This information will help in the determination of tectonic uplift rates and the climatic history of the western U.S. ERS-1 SAR images were used to study an area of western China where a large strike slip fault crosses a series of alluvial fans and stream valleys. Previous analysis of SPOT panchromatic images of the area shows that offsets fans and streams can be recognized. Measurement of the rate of motion of this fault will help in the overall model of deformation of the Asian tectonic plate in response to the collision of the Indian plate.
The Geophysical Processor System (GPS) at the Alaska (U.S.) SAR (Synthetic Aperture Radar) Facility (ASF) uses ERS-1 SAR images as input to generate three types of products: sea ice motion, sea ice type, and ocean wave spectra. The GPS, operating automatically with minimal human intervention, delivers its output to the Archive and Catalog System (ACS) where scientists can search and order the products on line. The GPS has generated more than 10,000 products since it became operational in Feb. 1992, and continues to deliver 500 new products per month to the ACS. These products cover the Beaufort and Chukchi Seas and the western portion of the central Arctic Ocean. More geophysical processing systems are needed to handle the large volumes of data from current and future satellites. Images must be routinely and consistently analyzed to yield useful information for scientists. The current GPS is a good, working prototype on the way to more sophisticated systems.
This paper is a summary of some research activities in calibrating DC-8 SAR data conducted in 1989/90 by the author and others. Firstly, by examining the process of 'symmetrization' in the production of DC-8 Stokes' matrix data, it is shown that an exact solution to the problem of calibrating this data exists, given certain assumptions. Secondly a new, more general, system model for radar polarimeters is proposed, together with a simple approach for verifying the model's validity. The importance of the orientation angle about the radar line of sight for some types of calibration devices is discussed and some results of an experiment to cross-calibrate the C-Band VV data from the DC-8 and DLR E-SAR sensors are presented.
A proof-of-concept experiment for remote sensing of precipitation by SAR is part of the SIR-C/X-SAR experiment. This thesis presents a feasibility study and recommendations for detection of precipitation using SIR-C/X-SAR. The principal limitation to rain measurement from a spaceborne SAR is the poor SCR (signal-to-clutter ratio). This is in part due to the system configuration and largely due to the large magnitude of echoes associated with the surface component. Two geometries apply: off-vertical and vertical pointing angles. Here we present calculations for both. With vertical geometry a large clutter component is associated with range sidelobes of the chirped transmitter pulse. To overcome this problem a narrow transmitted pulse (3 mu sec) processed without dechirping was used. Since the magnitude of the clutter over the ocean is high it is recommended that data in the chirped mode be obtained over the forest due to the significantly lower backscatter associated with it at nadir. With these recommendations, at nadir, it is believed that rain rates greater than 5 mm/hr may be detected. The use of a better weighting function that gives lower sidelobe levels than that used (a raised cos(exp 2)) is also recommended. At off-vertical look angles all the range cells have a large clutter component associated with them due to the geometry. The use of higher angles of incidence (theta greater than 60 deg) is recommended because of better SCR at these angles. With this recommendation, at off-vertical, it is believed that rain rates greater than 10 mm/hr may be detected. Various other techniques are described and recommended to improve the minimum detectable precipitation rate. These include trying to subtract the estimate of the clutter from the combined signal and clutter and trying to separate the Doppler of the rain echo and the surface echo. With these recommendations it is believed that it is possible to detect precipitation as low as 1 mm/hr at vertical and greater than 5 mm/hr at off-vertical look angles.