SHARAD P/L (SHAllow RADar PayLoad) is a subsurface sounding radar provided by the Italian Space Agency (ASI) as a facility instrument for the NASA mission to Mars called MRO (Mars Recoinnassance Orbiter).
The objective of the SHARAD is to detect the liquid water and to profile the ice layers within the first few hundreds of meters of the subsurface of Mars. Even if Mars surface is not uniformly apt for radar sounding, it will however be possible to find favourable conditions which may allow the identification from orbit of aqueous layers.
SHARAD will also provide new scientific data about Martian soil, ground morphologies and overall geology.
To summarize, the primary objective of the SHARAD investigation is to map, in selected sites, dielectric interfaces to depths of up to one kilometer in the Martian subsurface and to interpret these interfaces in terms of the occurrence and distribution of expected materials, including rock, regolith, water, and ice.
Key elements for the radar design are represented by the identified center frequency, 20 MHz, the bandwidth of the radar pulse equal to 10 MHz, and the requested spatial resolution which should be better than 1000 m in the along track direction and 7000 m in the cross track direction.
In the paper SHARAD design, its technological implications and tradeoffs are presented. In addition the architectural scheme of the instrument is described and an overview of the expected performances is given.
Besides spatial resolution, the peak sidelobe ratio (PSLR) is another important parameter to assess the quality of a SAR system. For the verification of SAR performance parameters, usually the point target responses of a number of ground-fixed reference targets such as corner reflectors or active transponders are evaluated in the processed SAR image. The area around these reference targets should consist of a (natural) cover with a radar backscattering coefficient σ0 as small as possible in order to limit the backscattered clutter energy. For SAR systems with low PSLR requirements the effect of this clutter is mostly neglected and the PSLR is determined in a classical manner by the estimation of mainlobe and sidelobe amplitudes from the range and azimuth section of the two-dimenensional point target response. The verification of high performance SAR systems, where challenging performance specifications are to be fulfilled, requires a more accurate PSLR estimation. Hence, the ground clutter of the surrounding target area has to be taken into account. Due to the ground clutter's statistical nature the superposition of the clutter with mainlobe/sidelobe amplitude is stochastic, therefore these amplitudes and the PSLR itself can be regarded as random variables. In our paper we suggest a combined deterministic-statistical approach as a tradeoff to the fully statistical modelling of the PSLR. This approach exploits the statistical properties of mainlobe and sidelobe with consideration of point target and clutter energy. Error bounds of the estimated PSLR are derived using established parameters such as signal-to-clutter ratio (SCR) and the classically defined PSLR. Furthermore some simulational results are presented which enable an evaluation of the calculated error bounds.
NASA's Mars Reconnaissance Orbiter mission, MRO, scheduled for launch in 2005, will be equipped with a sounder to find subsurface water and ice: Shallow Radar, SHARAD. This radar has been developed by Alenia Spazio and funded by Italian Space Agency, ASI.
An integral part of such kind missions is the development of an EGSE (Electrical Ground Support Equipment) capable to test all the radar functionalities.
CORISTA has been responsible to define the EGSE technical requirements and to design, build and test the Mars Echoes Generation System (MEGS).
This paper describes the activities developed and the results obtained during the test campaigns of SHARAD. An architectural description of the MEGS will be given with emphasis on the technical aspects related to the signal generation of Mars Echoes and possible operating modes.
TerraSAR-X is a new earth observing satellite which will be launched in spring 2006. It carries a high resolution X-band SAR sensor. For high image data quality, accurate ground calibration targets are necessary. This paper describes a novel system concept for an active and highly integrated, digitally controlled SAR system calibrator. A total of 16 active transponder and receiver systems and 17 receiver only systems will be fabricated for a calibration campaign. The calibration units serve for absolute radiometric calibration of the SAR image data. Additionally, they are equipped with an extra receiver path for two dimensional satellite antenna pattern recognition. The calibrator is controlled by a dedicated digital Electronic Control Unit (ECU). The different voltages needed by the calibrator and the ECU are provided by the third main unit called Power Management Unit (PMU).
Synthetic aperture radar (SAR) provides high resolution images of static ground scenes, but processing of data containing moving objects results in varying phase and amplitude effects. The work at hand illustrates via theoretical considerations and concrete simulations what happens to SAR imagery when parts of a scene are not static. We differentiate between four types of motion. Objects moving with a constant velocity cause position errors in azimuth as
well as target defocusing and smearing in azimuth and range. Accelerating objects are responsible for even stronger shift and defocusing effects since the position errors are now a function of time. Closely related are vibrations of an object. They may be interpreted as a regular and continuous de- and acceleration whose range component results in so-called paired echoes on each side of an object in azimuth. Finally, rotation as an extreme example of constant radial acceleration may disturb a SAR image over a wide area. Through a thorough motion analysis, we developed a
flexible SAR raw data simulator. Our simulations of point scatterers in raw data are based on the radar radiation pattern as a function of the system carrier frequency and the relative positions between the radar and each scatterer. All four types of movement described above may be expressed as varying relative positions and Doppler frequency shifts due to instantaneous phase variations. The standard SAR processing steps of range and azimuth compression for the
simulated data provide impressive results for freely adaptable system parameters of the movement and of the SAR system.
The German radar satellite TerraSAR-X is a high resolution, dual receive antenna SAR satellite, which will be launched in spring 2006. Since it will have the capability to measure the velocity of moving targets, the acquired interferometric data can be useful for traffic monitoring applications on a global scale. DLR has started already the development of an automatic and operational processing system which will detect cars, measure their speed and assign them to a road. Statistical approaches are used to derive the vehicle detection algorithm, which require the knowledge of the radar signatures of vehicles, especially under consideration of the geometry of the radar look direction and the vehicle orientation. Simulation of radar signatures is a very difficult task due to the lack of realistic models of vehicles. In this paper the radar signatures of the parking cars are presented. They are estimated experimentally from airborne E-SAR X-band data, which have been collected during flight campaigns in 2003-2005. Several test cars of the same type placed in carefully selected orientation angles and several over-flights with different heading angles made it possible to cover the whole range of aspect angles from 0° to 180°. The large synthetic aperture length or beam width angle of 7° can be divided into several looks. Thus processing of each look separately allows to increase the angle resolution. Such a radar signature profile of one type of vehicle over the whole range of aspect angles in fine resolution can be used further for the verification of simulation studies and for the performance prediction for traffic monitoring with TerraSAR-X.
Processing of SAR images of rugged terrain deserves special care
because the topography affects the focused image in a number of ways.
In order to obtain geometrically and radiometrically corrected SAR images of mountainous areas additional knowledge about the topography and the sensor's trajectory and attitude has to be included in the processing or post-processing steps. Various well-known focusing techniques are available to transform SAR raw data into a single look complex image such as the range-Doppler, the chirp scaling or the omega-k algorithm. While these algorithms perform the azimuth focusing step in the frequency domain the time-domain back-projection processing technique focuses the data geometrically, i.e., in the time domain. In contrast to the frequency-domain techniques, time-domain back-projection maintains the entire geometric relationship between the sensor and the illuminated area. This implies a couple of advantages: a stringent, terrain-based correction for the elevation antenna gain pattern may be implemented and topography-induced variation of radar brightness can be eliminated in a single step. Further, the SAR image is focused directly onto an arbitrary reconstruction grid and in the desired geodetic reference frame without requiring any additional processing steps. We discuss the influence of rugged terrain on the radiometric properties of focused SAR data and demonstrate how the time-domain back-projection approach accounts for these effects within one integrated processing framework by incorporating both a correction for terrain slope induced variation of radar brightness and a stringent correction for the elevation antenna gain pattern. The algorithm is evaluated for ENVISAT/ASAR image mode data of a mountainous area.
This paper proposes an inversion algorithm for extracting soil and vegetation parameters from multi-frequency, multi-polarimetric SAR radar data. The algorithm, based on the determination of probability density functions (pdfs) through a Bayesian methodology, has been initially developed for bare soils and tested on numerous data sets. The pdfs are obtained from the comparison between theoretical backscattering values, derived from the Integral Equation Model (IEM), and the measured ones. The aim of this work is to apply this inversion algorithm to fields that have different levels of vegetation cover. As first approach, the pdfs have been empirically calibrated according to vegetation water content values obtained from a multispectral image (Landsat image).
This allows subtracting the contribution of vegetation backscatter from radar signal and hence isolating the contribution from bare soil. In a second approach, instead of using the IEM, a simple vegetation model, the water-cloud model, has been used. This model relates the radar responses to both soil and vegetation characteristics.
Thus, the comparison between the theoretical and the measured sensor responses leads to the calculation of new pdfs which contain information about both vegetation and soil parameters. The algorithm has been tested on data sets acquired during the SMEX'02 experiment covering a variety of soil moisture and vegetation conditions. The algorithm exploits the use of L and C band at different polarisations. The results have been also compared with those obtained from the algorithm developed only for bare soils.
Speckle - appearing in SAR Images as random noise - hampers image processing techniques like segmentation and classification. Several algorithms have been developed to suppress the speckle effect. One disadvantage, even with optimized speckle reduction algorithms, is a blurring of the image. This effect, which appears especially along the edges of structures, is leading to further problems in subsequent image interpretation. To prevent a loss of information, the knowledge of structures in the image could be an advantage. Therefore the proposed methodology combines common filtering techniques with results from a segmentation of optical images for an object-based speckle filtering. The performance of the adapted algorithm is compared to those of common speckle filters. The accuracy assessment is based on statistical criteria and visual interpretation of the images. The results show that the efficiency of the speckle filter algorithm can be increased while a loss of information can be reduced using the boundary during the filtering process.
This paper presents a novel multisensor image fusion algorithm, which extends pan-sharpening of multispectral (MS) data through intensity modulation to the integration of MS and SAR imagery. The method relies on SAR texture, extracted by ratioing a map of a SAR feature to its lowpass approximation. SAR texture is used to modulate the generalized intensity (GI) of the MS image, which is given by a linear transform extending Intensity-Hue-Saturation (IHS) transform to an arbitrary number of bands. Before modulation, the GI is enhanced by injection of highpass details extracted from the available Pan image by means of the "à-trous" wavelet decomposition. The texture-modulated pan-sharpened GI replaces the GI calculated from the resampled original MS data; then the inverse transform is applied to obtain the fusion product. Experimental results are presented on Landsat-7/ETM+ and ERS-2 images of an urban area. The results demonstrate accurate spectral preservation on vegetated regions, bare soil, and also on textured areas (buildings and road network) where SAR texture information enhances the fusion product, which can be usefully applied for both visual analysis and classification purposes.
Full polarimetric data can define the scattering behavior of land use/cover through several approaches. Several classification methods have been proposed based on analysis methods. These classification methods are based on the backscattering mechanisms which are extracted using a single decomposition method. The objectives of this work are a) the investigation of the different polarimetric analysis methods; b) the interpretation of the images resulting from polarimetric analysis; c) the development of an object-oriented classification method based on polarimetric analysis imagery and the comparison of this method with the H/a and the H/a Wishart classification methods respectively.
The Shuttle Radar Topography Mission (SRTM), used an Interferometric Synthetic Aperture Radar (IFSAR) instrument to produce a near-global digital elevation map of the earth's land surface with 16 m absolute vertical height accuracy at 30 meter postings. SRTM achieved the required interferometric baseline by extending a receive-only radar antenna on a 60 meter deployable mast from the shuttle payload bay. Continuous measurement of the interferometric baseline length, attitude, and position was required at the 2 mm, 9 arcsec, and 1 m (1.6 sigma) levels, respectively, in order to obtain the desired height accuracy. The collected data were used to generate a digital topographic map of 80 percent of Earth's land surface (everything between 60 degrees north and 56 degrees south latitude), with data points spaced every 1 arcsecond of latitude and longitude (approximately 30 meters). An SRTM 3-arc-second product (90m resolution) is available for the entire world. In this paper we compare a DTM created from SRTM data to a DTM created from 1/50.000 topographic maps. The area of study is Kos Island in the Aegean Sea.
Since the SRTM elevation data are unedited, they contain occasional voids, or gaps, where the terrain lay in the radar beam's shadow or in areas of extremely low radar backscatter, such as sea, dams, lakes and virtually any water covered surface that are flat but they don't look so flat on SRTM tiles. We used different filters and masks in order to ameliorate the quality of the DEM. The first filter detected and removed the voids; a second one interpolated the missing values and then a mask was used in order to separate sea from land. We also created a DTM from digitized contours of 1/50.000 scale topographic maps and we used more than 1800 extra points in order to ameliorate the quality of this DTM around the coastline. We compared the two DTMs. All the results demonstrated that the SRTM DTM presents a very good accuracy.
Many are the examples of application of SAR and differential SAR interferometry for topographic mapping and ground deformation monitoring. However, on repeat pass geometry, the performances of these two techniques are limited by the loss of correlation (coherence) between the two radar acquisitions. The lack of coherence causes an additional noise thus a poor estimate of the interferometric phase. The disturbances can be due either to surface changes because of long period cover (temporal decorrelation) or to a too large baseline (spatial decorrelation). In this paper, we propose an empirical model for the estimate of coherence considering separately these two sources of disturbances. Starting from the observations of experimental data, we study the behaviour of coherence according to baseline and period cover in order to express the two terms of correlation. A number of 170 multi temporal and multi baseline differential interferograms covering the same region is used to validate the proposed model.
A new method, based on an Infomax Learning Algorithm and neural networks, for the pattern extraction in Synthetic Aperture Radar (SAR) images was developed. SAR Images can be obtained from the radar backscatter that is fully dependent on the surface conditions of the target. Then, it can distinguish the target from non-target by extracting the difference in patterns generated by the backscatters.
The difficulty is that the difference of geometric pattern's characteristics of these targets is unknown beforehand, and then it is not easy to adapt simple filtering to divide these patterns.
We tried to develop the computer algorithm simulating the pattern extraction system in reference to the excellence of human vision's recognition ability. We propose the classification method based on image patterns using Infomax Learning Algorithm which simulates human visual cortical neurons as the possibility to automatically extract the specific area from an image. This algorithm considers a neural network model of visual area for modeling human visual recognition process, learns neural networks based on the idea of Infomax (Information Maximization), and learns local image pattern to obtain the weight patterns required for classification, which are approximated by Gabor function. We attempted to detect airplanes and aprons in SAR image taken at Kansai International Airport by the proposed method. As the result of adopting Infomax for the SAR image based on the improved learning method, we successfully demonstrated to detect airplanes from other various, complex artificial constructions, which are specific to the airport, in the image. Hence, it was demonstrated that target was automatically detected by learning of its patterns without any mathematical definition of equations for the pattern.
SAR spaceborne capability to detect marine oil spills through damping of wind-generated short gravity-capillary waves has been extensively demonstrated during past years. In contrast, it has not yet been found the optimal use of optical/NIR imaging sensors for detection and monitoring of polluted areas. We propose the use of Modis images acquired in sun glint conditions to reveal smoothed regions such as those affected by oil pollution. The underlying physical mechanism is based on the modification of the surface slopes distribution composing the roughened sea due to the action of mineral oils. The methodology is demonstrated for selected case studies in the Mediterranean Sea and North Atlantic where spills were detected by ERS SAR imaging. The corresponding Modis images acquired within a few hours were under sun glint conditions according to satellite imaging geometry and wind field distribution over the selected areas. Results of a detailed study about the effective applicability of the method is discussed. The importance of these results are based on the possible extensive exploitation of combined Modis and SAR data in view of the high repetitive coverage (about two times a day).
The possibility to retrieve ocean wave parameters from SAR images of waves in the nearshore region is explored. Contrarily to the open ocean case, nearshore waters render more complex wave patterns due to interactions with the bottom and obstacles. The basic idea was to use the MPI inversion algorithm in combination with first guess spectra generated by the SWAN model. In order to assess the performance of the algorithm, several experiments were carried out to analyze the inverted spectrum, as well as the degree of influence of the first guess spectrum in the retrieved spectrum. Spatial variations of the wave field in the northwestern coast of Baja California, Mexico, were analyzed by using subimages extracted from ERS-2 image mode products. Wave spectra were retrieved from subimage spectra by inverting Hasselmann's spectral transformation relation, which describes the nonlinear mapping of an ocean wave spectrum onto a SAR image spectrum. The retrieved wave spectra showed a significant improvement with regard to the initial SWAN spectra, displaying more accurate spectral peaks and wave modes not present in the initial spectra. Although a low dependence of the retrieved wave spectra from the first guess was observed, the former one can influence the distribution of the secondary wave systems. It is concluded that information not accounted by the model can be inferred from the SAR image in accordance to the wave imaging theory. These results are encouraging for researchers conducting studies on wave modeling in shallow waters, as well as in data assimilation programs.
This paper describes the effectiveness of the fusion of SAR images and Optical images based on the use of wavelet transform. Through the tests for differences between wavelet transform technique and other fusion technique, we explain the strength and weakness about each technique. We have studied on the improvement of classification accuracy using the fusion of SAR images and Optical images. In this paper, we focus on the classification of urban area which results from combining Spot Pan, Spot MS and SAR images.
Radar satellite images could be used to produce digital elevation model (DEM) of certain areas by processing a couple of images, covering the same area, obtained at two different angles. In this study, the DEM generated from the Canadian RADARSAT stereoscopic data for a north western area of the Gulf of Suez, Egypt, is compared to the DEM generated from the topographic contour maps, scale 1:50,000. An evaluation and assessment of the results were conducted. The study shows that the DEM derived from RADARSAT data has a high precision as compared to the one generated from the topographic maps. It is also accurate enough to provide information where other sources of digital elevation are not available.
As one of important applications in Synthetic Aperture Radar (SAR) images, the recognition of urban area has received considerable attentions in remote sensing. The extraction of line segment is very critical technology to recognize the urban area because many objects such as streets and buildings are line segment. The common method to extract line segment is Hough transform, but most of the previous methods are based on binary images. So we have to select a threshold to binarizate the image, but at most time we can not determine the threshold properly, resulting in the lost of useful information. To solve the problem, an improved Hough transform algorithm on gray level, which can make the extraction of line segment independent of the noise and the length of line segment, is proposed. The approach is validated by the analysis of SAR images.