Shuttle Imaging Radar-B Research Articles

Application of the semivariogram textural classifier (STC) for vegetation discrimination using SIR‐B data of the guiana shield, northwestern brazil

Abstract In 1984, the Shuttle Imaging Radar‐B (SIR‐B) acquired digital data over an inaccessible, heavily‐vegetated and generally cloud‐covered region of the Guiana Shield in northwestern Brazil. Available natural resource information in this area is limited to a reconnaissance survey conducted by the RADAMBRASIL Project at a 1:1,000,000 scale. Classification of this SIR‐B data set is performed using image texture. The algorithm used is the semivariogram textural classifier (STC). This is a deterministic, supervised parallelepiped type classifier which provides the option of combining textural and radiometric information. Textural information is expressed by the semivariogram function. Radiometric information is conveyed by the mean digital number (DN) value. Results have shown that the SIR‐B image representation of each cover type in the study area has a distinctive semivariogram, making feasible the discrimination of vegetation types and water bodies by means of STC.

Synergistic effects of combined Landsat-TM and SIR-B data for forest resources assessment

Abstract Three different incidence angle data sets, obtained by the Shuttle Imaging Radar-B in October 1984 over a forested area in northern Florida, were combined with Landsat-5 Thematic Mapper data, to create a digitally registered 10 channel optical/microwave data set. The work discussed in this paper involved the analysis of the data obtained by the two sensors separately and in combination, to determine if there are synergistic effects obtained through the simultaneous use of data obtained from both the optical and microwave portion of the spectrum. The radar data were filtered with a low-pass filter to eliminate the speckle noise. Classifications of the TM, SIR-B and combined TM +SIR-B data sets were performed with both per-point and contextual classifiers. The results showed that filtered radar data can be used to classify accurately major cover types (i.e., pine forest, swamplands and radar smooth targets) and that the contextual classifier provided better classification performance. The combined ...

DIGITAL TERRAIN MODELING FROM RADAR IMAGE STEREOPAIRS

A methodology for completing a digital terrain model of the Earth's surface is described, which utilizes overlapping radar images from different orbits of the Almazl spacecraft. Algorithmic digital image processing procedures are described for compiling a three-dimensional image of surface terrain from two images forming a stereopair. Using imagery of North America's Great Basin desert that is roughly analogous to that of the Shuttle Imaging Radar-B of the U. S., the paper demonstrates how orthographic and topographic images can be derived via the processing of two overlapping images. Translated by Edward Torrey, Alexandria, VA 22308 from: Issledovaniye Zemli iz kosmosa, No. 6, 1992, pp. 53–61.

Application of the semivariogram textural classifier (STC) for vegetation discrimination using SIR-B data of Borneo

Abstract Classification of Shuttle Imaging Radar-B (SIR-B) data from a rainforest-covered portion of Borneo is performed using image texture. The algorithm used is the semivariogram textural classifier (STC). This is a deterministic, supervised parallelepiped type classifier which provides the option of combining textural and radiometric information. Textural information is expressed by the semivariogram function. Radiometric information is conveyed by the mean digital number (DN) value. Results of the classificaiion cmulale a previously published map obtained by visual interpretation of the same SIR-B data set.

A pyramidal scheme for stereo matching SIR-B imagery

Abstract A procedure is presented in the stereo matching of digital synthetic aperture radar (SAR) imagery obtained from NASA's shuttle imaging radar B (SIR-B) experiment. The coherent nature of SAR means that speckle is an inevitable by-product. The presence of speckle in SAR imagery makes the conventional use of region growing stereo-matching algorithms untenable. To overcome the problem of speckle, a technique was developed whereby stereo matching was performed in a course-to-fine pyramidal fashion. It is shown that the stereo-matched extent obtained with this technique can give an approximate three-fold increase in stereo coverage, if a suitable speckle-reduction filter is also employed.

The consequences of sampling variability on the estimation of wave number and propagation direction from spaceborne SAR image spectra

Even if it were true that SAR (synthetic aperture radar) ocean surface imagery provides a perfect noiseless representation of ocean surface wave height, each spectrum computed from such imagery would be but a single realization of the ensemble-mean ocean spectrum. There would be sampling variability associated with the parameters of dominant wave number and propagation direction extracted from such a spectrum. The present study addresses two questions: (i) what statistical distribution is applicable to the spectra of SAR ocean images? and (ii) what are the consequences of such statistics on the precision with which wave number and propagation direction can be extracted? An examination is made of spectra computed from ocean imagery acquired during the SIR-B (Shuttle Imaging Radar-B) mission. >

Estimation of surface topography from SAR imagery using shape from shading techniques

In this paper a practical method is demonstrated for estimating topography of natural terrain from the radiometric, or shading, information in a synthetic aperture radar (SAR) image. While this problem has been considered before for radar, methods available in the computer vision literature have not previously been utilized in its solution. We treat this as a computer vision problem, viz. shape from shading (SFS). A review of the relevant characteristics of SAR imagery is presented followed by a formulation of the SFS problem for SAR. Because of the noise inherent in SAR imagery a cost minimization approach is used which allows for noise and incorporates a regularization term in the cost function. Previously developed numerical methods are adapted to SAR imagery by incorporating radar reflectance models and by solving for surface slopes in a rotated system of coordinates—one which represents surface height relative to a plane parallel to the line-of-sight. Two difficulties are recognized. First, unknown reflectance model parameters must be estimated from the image data. Second, shading provides reliable information about the high frequency components of the surface but not the low frequency components. Both difficulties are reduced if auxiliary low resolution surface height information is available. This is demonstrated by combining Shuttle Imaging Radar-B (SIR-B) SAR images with much lower resolution terrain elevation data to construct high resolution terrain elevation estimates. The estimation of Venusian surface topography from Magellan SAR imagery is discussed and methods are suggested for combining shading information with geometric stereo.

Ocean wave spectra derived from Shuttle Imaging Radar‐B imagery and surface measurements

During the Shuttle Imaging Radar‐B (SIR‐B) mission in October 1984, synthetic aperture radar (SAR) images were taken off the coast of Japan. Wave spectra derived from SIR‐B imagery were compared with simultaneous wave buoy observations. Waves at the time of flight were nearly perpendicular to the SAR flight direction and were composed of two major systems: a swell of about 200‐m wavelength and wind waves with 70‐m wavelength. Both the swell and wind wave peaks were identifiable in the SAR image spectrum after stationary response correction. Peak frequencies derived from SIR‐B images and from the wave gauge agreed satisfactorily. However, the relative magnitudes of the peaks are not in accordance with the surface measurements. Possible reasons for the disagreement are discussed.

Validation of a synthetic aperture radar ocean wave imaging theory by the Shuttle Imaging Radar‐B Experiment over the North Sea

Synthetic aperture radar (SAR) image intensity spectra measured during the Shuttle Imaging Radar‐B (SIR‐B) mission on October 6 and 8, 1984, over the North Sea are compared with ocean wave spectra obtained from a wave prediction model. The relationship between both spectra is discussed in the light of a previously proposed SAR imaging theory of ocean waves. Most of the measured SAR image spectra of October 6 show double peaks, while the hindcast ocean wave spectra have only single peaks. It is shown that the double peaks are generated by the SAR imaging mechanism. They occur only when velocity bunching is sufficiently strong near the range direction and when enough spectral energy density is present in this spectral region. Double peaks are generated when the SAR modulation transfer function (MTF) consisting of the sum of the complex real aperture radar (RAR) MTF and the velocity bunching MTF has a strong minimum near the range direction by which the wave spectrum is cut into two, in general unequal, pieces. The azimuth angle at which this minimum occurs depends strongly on the phase of the RAR MTF. In general, the two peaks are not located symmetrically with respect to the range direction, nor do they have equal heights. We interpret the reasonably good agreement between simulated and measured SAR image spectra during the SIR‐B mission over the North Sea as further experimental evidence in support of the validity of the previously proposed SAR imaging theory of ocean waves.

Combined Radar and Landsat data analysis for land use/cover studies over parts of the Punjab plains

L-band (HH) synthetic aperture radar imagery from Shuttle Imaging Radar-B (SIR-B) and Landsat multispectral scanner (MSS) images over parts of the Punjab plains were combined in order to utilize the complementary information contained in multispectral data sets. Among the various combination of Landsat MSS with SIR-B, the combination of Landsat MSS band 5 (0.6–0.7 μm) and band 7 (0.8–1.1 μm) with SIR-B data was found to be optimum in delineating landcover units. The integrated data was found to be superior in providing landcover information in comparison to SIR-B alone or a combination of landsat MSS band 4,5 and 7.

Flood boundary delineation through clouds and vegetation using l-band space-borne radar: A potential new tool for disease vector control programs

In October 1984, space-borne multiple incidence angle synthetic aperture radar (SAR) data acquired by the Shuttle Imaging Radar-B Program were used to delineate pools of standing water beneath a 12.5 m tall 100% closed forest canopy and to map flood boundaries for the assessment of flood damage in the Peoples Republic of Bangladesh. The projects, as described here, demonstrate the potential use of SAR systems as input to disease and disease vector control programs in monsoon countries.

Specular point scattering contribution to the mean synthetic aperture radar image of the ocean surface

In general, the return signal scattered from the ocean surface used to form synthetic aperture radar (SAR) images contains contributions from at least two scattering mechanisms. In addition to resonant Bragg‐type scattering, specular point scattering becomes important as the angle of incidence becomes small ( ≲ 20°). In this paper we include the specular point rough surface scattering mechanism in a model for the mean SAR image of the ocean surface and examine the effects of this scattering mechanism theoretically. We find that the complete mean SAR intensity image consists of a sum of images due to specular point scattering and Bragg‐type resonant scattering. Because surface specular points have a short coherence time and move with considerable velocities, the contribution to the mean image due to these scatterers is of low azimuthal resolution and is displaced from the actual sea surface, typically by several SAR resolution cells. The bandwidth of this image can easily exceed the bandwidth of a typical SAR processor, leading to a loss of mean image intensity. The local backscatter cross‐section modulation is strong and nonlinear in the slope of the longwave field in the SAR range direction. At small incidence angles, this causes the specular point return from wave slopes tipped toward the SAR to become much brighter than the Bragg‐scattering return. Taken together, these effects are capable of producing azimuthally oriented streaks in SAR images, such as have been observed by Seasat. We present numerical estimates of coherence time, azimuthal displacement, cross‐section modulation, etc., computed using the parameters of the Seasat and shuttle imaging radar‐B SARs as well as typical parameters for an airborne X band SAR.

Cartographic feature extraction with integrated SIR-B and Landsat TM images

Abstract A digital cartographic multisensor image database of excellent geometry and improved resolution was created by registering Shuttle Imaging Radar-B (SIR-B) images to a rectified Landsat Thematic Mapper (TM) reference image and applying intensity-hue-saturation enhancement techniques. When evaluated against geodetic control, RMSEXY values of approximately ±20m were noted for the composite S1R-B/TM images. The completeness of cartographic features extracted from the composite images exceeded those obtained from separate SIR-B and TM image data sets by approximately 10 and 25 per cent, respectively, indicating that the composite images may prove suitable for plani-metric mapping at a scale of 1:100000 or smaller. At present, the most effective method for extracting cartographic information involves digitizing features directly from the image processing display screen.

A demonstration of stereophotogrammetry with combined SIR-B and Landsat TM images

Shuttle Imaging Radar-B (SIR-B) and Landsat Thematic Mapper (TM) images can be viewed stereoscopically if the illumination geometries are compatible. To create the stereoscopic effect points must be coregistered. Simplified stereophotogrammetric equations permit the height of an object above a reference plane to be crudely calculated from the target offset toward the SIR-B radar antenna with reference to its position on a TM image. Precision is limited by pixel resolution and target correlation. Future spaceborne imaging radar missions will offer the potential for topographic mapping in many areas where TM coverage is available.

An effect of coherent scattering in spaceborne and airborne SAR images

Abstract Radar imagery obtained by the Shuttle Imaging Radar-B (SIR-B) is compared to high resolution aircraft imagery of the same urban and agricultural areas close to the city of Montreal, Canada. It is clear that the SIR-B radar is more sensitive than the aircraft radar to reflections from extended, along-track radar targets. The effect is evident for both urban and agricultural areas in which the street or field orientation is near parallel to the radar azimuth direction. It is likely that such reflectivity enhancement is due to coherent combination of the scattered field from appropriate scattering centres. The paper considers the observed phenomena and probes potential scattering models to explain the results.

Determination of the vertical pattern of the SIR-B antenna

Abstract Determination of the antenna pattern is important for a spaceborne Synthetic Aperture Radar such as Shuttle Imaging Radar-B (SIR-B). For SI R-B the antenna was so large that apart from one section, no complete pattern could be measured on the ground. Attempts were made to measure the pattern while the shuttle was in space by using ground receivers and active radar calibrators. The method used and described here is a supplement to these measurements. The vertical pattern of an antenna can be extracted from radar signals returned from regions whose scattering coefficients versus incidence angle characteristics are suitably flat and uniform. The method used here shows that the main vertical lobe of the SIR-B antenna is slightly wider than previously reported (6·9° at 3dB points versus 6·2° used in radiometric corrections).

Structural geology and regional tectonics of the Mineral County area, Nevada, using Shuttle Imaging Radar-B and digital aeromagnetic data

SIR-B and aeromagnetic lineaments show a correlation with known faults within the Walker Lane of the western Basin and Range Province. Walker Lane faults can be extended based on SIR-B data and unmapped Walker Lane features are identified. South of the Pancake Range lineament, Walker Lane faults are seldom recognized but SIR-B data have the potential for delineating structural features in this area. Earthquake hypocenter distribution in profiles across a SIR-B lineament is consistent with that expected from a vertical fault. Field examination indicates that the lineament is an unmapped neotectonic Walker Lane feature.

Shuttle radar mapping with diverse incidence angles in the rainforest of Borneo

Abstract Shuttle Imaging Radar-B (SIR-B) images, obtained at two different incidence angles were analysed for discrimination and mapping of vegetation in the rainforest of Borneo. In an area of coastal lowland three units of forest canopy and two units of open surface cover were distinguished and mapped. The backscatter characteristics of the units were compared and analysed by ratioing digital image data of sample sites against digital data of reference sites. The ratios show that the backscatter for each surface unit is clustered in separate, mostly non-overlapping, domains at both incidence angles. Different rates of change with incidence angle indicate a strong angular dependence in the dominant backscatter mechanism of swamp that is not apparent for the other units. A corresponding digitally coregistered Landsat multispectral scanner (MSS) image helped to verify the spatial distribution of the mapped units in the coastal lowland. In the interior upland three major forest species associations that hav...

The Effects of Soil Moisture, Surface Roughness, and Vegetation on L-Band Emission and Backscatter

Measurements with Shuttle Imaging Radar B (SIR-B) at 1.28 GHz and an airborne multiple-beam push-broom radiometer at 1.4 GHz were made over a number of agricultural fields near Fresno, California during October 7-10, 1984. These measurements provided a unique data set for studies of microwave emission and backscatter from surfaces of various characteristics. The effects of surface roughness and vegetation (alfalfa and lettuce) were analyzed with respect to the responses of microwave emission and backscatter to soil-moisture variations. A theoretical model (Kirchhoff approximation) was employed to assess these effects. It was found that for microwave emission, the effect of surface roughness is less significant compared to that of vegetation. On the other hand, the surface roughness was shown to play a dominant role compared to the vegetation cover in the microwave backscatterve backscatter. The two roughness parameters in the theoretical model calculations were the surface correlation length and the standard deviation of surface height. These parameters were found to be affected strongly by the soil-texture effect in the emissivity calculations. A disagreement was found between the calculated and the observed scattering coefficients if the measured surface correlation length and standard deviation of surface height were input to the model. Either one of these two parameters had to be modified appreciably to bring a comparability between the measured and calculated scattering coefficients.

Shuttle Imaging Radar B (SIR‐B) Weddell Sea ice observations: A comparison of SIR‐B and scanning multichannel microwave radiometer ice concentrations

The October 1984 Shuttle Imaging Radar B (SIR‐B) flight made three radar passes over the Weddell Sea ice, providing the first high‐resolution look at the Weddell marginal ice zone properties. Using these data, this paper discusses the effect of ocean waves on the radar return at the ice edge and compares ice concentrations derived from the SIR‐B with coincident concentrations from the Nimbus 7 scanning multichannel microwave radiometer (SMMR). The comparison of the SIR and SMMR concentrations is possible because SIR cross‐track width and the diameter of the SMMR 37‐GHz integrated field‐of‐view are both about 30 km. The SIR ice concentrations are computed in two ways: first, using a training area classification scheme at the Jet Propulsion Laboratory (JPL); second, using a manual classification method at the Scott Polar Research Institute. The SMMR ice concentrations are calculated using the Goddard Space Flight Center algorithm. At the ice edge, where there were no coincident SMMR data and where ice bands predominated to yield an ice concentration of the order of 10% or less, comparison of the two different analysis techniques on the same images showed that, for the JPL technique to avoid classifying some of the open water as ice, two classes of open water must be defined. These two classes accounted for the rougher ocean surface upwind of the bands and the smoother downwind surface. In the ice interior, comparison of the coincident SIR and SMMR ice concentrations shows that for concentrations greater than 40%, which was the smallest concentration jointly observed, the mean difference between the two data sets for 12 points is 2% and the standard deviation is 7%.