The success of current spaceborne SyntheticApertureRadar (SAR) is boosting the performance requirement of next gener- ation systems. In order to cope with the evolution of SAR the design of the new systems will need to meet higher requirements for spatial and radiometric resolution together with an increased availability. This tendency is recognized nearly independently of the application area and manifests itself through several study programs initiated by space agencies aiming at the design of fu- ture SAR systems. In this context the use of large reflectors combined with digital feed arrays for SAR is considered a pos- sible alternative to planar array antennas. This paper suggests an X-band spaceborne SAR system utilizing a deployable reflector together with a digital feed array, analyzes its performance and highlights its advantages compared to other systems based on direct radiating arrays.
In  it was shown that losses of several dB could occur when no information about topographic height is conveyed in the steering mechanism. This observation suggested the option to compute adaptively the steering direction of the receive beam, by processing the signals available from the vertical sub-apertures of the multi-channel receive antenna. In  a novel algorithm for receive beam steering, the Adaptive Digital Beam-Forming (ADBF), based on the estimation of the actual DOA of the received echo was proposed. The ADBF performance was analyzed by Monte Carlo (MC) simulation versus the imaged scene parameters, proving the capability of ADBF to overcome SCORE limitations [6, 7]. In this paper, the analysis is further developed by focusing on the dependence of the ADBF performance on the SAR systems parameters. In particular, the ADBF performance is computed versus those parameters, such as dimension of the antenna and number of sub- apertures, whose value is bound by physical and economical constraints and strongly affects the complexity of ADBF itself.
We investigate the occlusions between different objects in high resolution SAR imagery in urban areas. For instance, we estimate that buildings occlude trees and shadow, trees occlude lawn, etc. We adopt an iterative strategy [ 99 ] exploring boundary strength and region characteristics coherently to solve this difficult problem. We integrate the occlusion boundary estimation with segmentation problem. An initial oversegmentation is obtained by applying watershed algorithm to a PolSAR span image. The boundaries of the generated segments are potential occlusion boundaries. Weak boundaries that are less likely to be occlusions can be removed and the small regions can be grouped if they have the same surface type. Many effective features are adopted in our experiments, which help to characterize boundaries and regions efficiently. The boundary and region likelihoods are integrated into a CRF framework, which models the interaction of boundaries, junctions and regions. The CRF inference outputs the occlusion boundary map. Our goal is to find the occlusion relationships and boundaries. The recovered occlusion boundary map shows major occlusions in a SAR image. Therefore, it would be helpful for 3D scene understanding of a single high resolution SAR image. An accurate occlusion boundary map also defines ahigh-quality segmentation. The segmentation formed by the boundaries also gives an efficient figure/ground segmentation for further object analysis.
To fulfill the highest quality standards in DBF-SAR and MIMO-SAR, and to avoid the previously mentioned mis-pointing effect, it is necessary to derive an effective application orientated algorithm. It has been found, that the Matrix Pen- cil (MP) method, which was originally intended to solve a linear equation system, serves as a good starting point for that . In the meantime, the MP also finds application in some basic angular estimation problems [110, 111] and even for back projection processing of inverse SAR . This parametric approach does not need a calculation of a covariance matrix. With reference to the run-time, this is a significant advantage in contrast to other established methods, such as MU- SIC or ESPRIT (a detailed comparison of the run-time between the estimators is given in section 5.4.1). Because the MP method is applied on-board, aconceptfora possible practical implementation is given, which enables a precise and stable topography-adaptive elevation DBF in real-time.
The relativistic phase and time offsets from this paper are not only of high importance for DEM generation with a formation flying SAR cross-track interferometer. Formations with multiple satellites have also been suggested fora wide range of further remote sensing applications, ranging from along-track interferometry for moving object and ocean current measurements over sparse aperture ambiguity suppression and super resolution for enhanced high-resolution wide-swath SAR imaging up to single-pass SAR tomography for vertical struc- ture measurements –. Due consideration of relativistic effects from varying along-track baselines is again of essential importance for these advanced bistatic and multistatic SAR systems to avoid mutual range and phase offsets between the received SAR signals. The phase accuracy requirements for the combination of the different receiver signals are typically in the order of 1° or a few degrees. For comparison, an along-track baseline of 100 m causes in an X-band system a relativistic phase shift in the order of several tens of degrees. Future multistatic SAR satellite missions should therefore take into account relativistic effects in the design of the radar synchronization system and/or the SAR processor to avoid a possible performance loss.
In , partially buried landmines are detected optically using a down-looking on-board camera on a low-cost UAS. A UAS equipped with a metal detector and a thermographic cam- era is reported in . UAS-based approaches using a down- looking ground penetrating radar (GPR) for mine detection are presented in , , . The down-looking approach has the advantage that the radar can directly measure the depth of a target. However, the area throughput is low. The area under investigation has to be scanned in parallel straight lines (B- scans) so that ahigh-resolution 3D image (C-scan) can be generated. With regard to the detection of tripwires, depth resolution plays a subordinate role. Compared to the down- looking approach, the advantage of a side-looking radar is the increased area throughput and the lower ground reflections. In , different UASs designed for military applications in combination with a side-looking GPR are investigated for mine detection. However, due to the low altitude required to detect landmines with a GPR and the high material and financial risk associated with this type of UASs, the approach was considered impractical.
A spaceborne SAR system is used in the wide spectrum of the remote sensing owing to its wide-area observation capability and high resolution. A modern SAR system based on the phased-array antennas realizes the multiple SAR operation such as the stripmap, spotlight and scan mode operation. However, this system does not fulfill the users demands for SAR data fora wide area with the fine resolution since there is a fundamental trade-off between the ambiguity and the swath width. Additionally, the lifetime of a satellite depends on the fuel budget to maintain its orbit and motion control. A SAR sensor observing the wide area can reduce the revisit time and thereby extend the operation period. Therefore, the simultaneous achievement of the wide coverage and fine spatial resolution is quite attractive not only for the SAR users, but also in the system point of view. Such a
With regard to SAR images, the influence of pixel size for distributed targets has been partially examined by Nesti et al. (1996). They conclude that the same target appears different depending on the spatial resolution of the SAR processor. This has an immediate impact on the estimation of the backscattering coefficient σ 0 . Their results show that with spatial resolutions smaller than 2 AC length of the target, the statistics of the backscattered signal do not follow those of the speckle model presented in Section 2.1.3. For larger pixel sizes, the experimental data become instead consistent with the Rayleigh model. This finding substantially confirms the results presented in Sarabandi and Oh (1995) which, based on numerical simulations, show that a correct estimate for the σ 0 can be obtained with pixel sizes above 2 AC length. Nesti et al. (1996) do not specifically discuss the spatial structure variation. They simply notice that “in the high resolution image, there are many bright spots sparsely distributed over the surface, whereas in the low resolution image, there are fewer and larger spots almost uniformly covering the entire surface”. In other words, an increasing resolution progressively reveals a different structure of the target, theoretically reaching the point at which each individual scatterer becomes observable and dominates the return in the pixel. This does not mean that the wave-target scattering mechanism which is directly linked to the frequency and polarisation of the SAR system actually changes. For instance, the 5 multilook averaging process in azimuth described in Section 5.2 suitably reduces the speckle effect of SAR imagery but it may also hide texture features which are unrelated to the speckle models of homogeneous targets. Furthermore, this analysis can also be linked to the Ulaby et al. (1982) criteria used in Section 3.1.3 such that horizontal spacing distance in the AC length estimation must be ≤ 0 . 1 λ .
to a considerable improvement in system performance when compared to the state-of-the-art SAR, and its benefits can be maximized when combined with digital beamforming (DBF) capability . So, SAR systems with DBF functionality on receive are spotlighted as a promising way toward the next generation of SAR instrument –. The high-resolution wide-swath (HRWS) SAR is a representative example of the DBF SAR concept. The HRWS SAR system enables one to circumvent the fundamental limitation in the SAR design and produces high-resolution images over a wide area with the aid of DBF , . The multiple channels of the HRWS system provide increased degrees of freedom that are fully exploited for the suppression of azimuth ambiguities caused by lower pulse repetition frequency (PRF) than the Nyquist sampling requirement. The DBF techniques make the ambiguity suppres- sion flexible and adaptive , . In this paper, we attempt to build a reconfigurable SAR system concept in which the HRWS SAR imaging operation can be effectively combined with other operation modes, avoiding performance degradation. For this purpose, one needs to increase the spatial degrees of freedom. We achieve it by adding transmit antennas to the multiple receive antennas, the so-called multiple–input multiple–output (MIMO) SAR.
The HRWS system concept assumes a wide area illu- mination by a separate transmit antenna. This enables an independent electrical design and optimization of the transmit and receive paths, but it requires also the accom- modation of an additional antenna on the spacecraft and reduces the flexibility to operate the radar system in differ- ent SAR imaging modes like ultra-wide-swath ScanSAR, high SNR spotlight, or new hybrid modes to be discussed later. It is hence worth to consider also the application of digital beamforming techniques in radar systems that use the same antenna array for both the transmission and re- ception of radar pulses, thereby taking advantage of al- ready existing space-qualified T/R module technology. Since the high-resolution wide-swath SAR imaging capa- bility is essentially based on a large antenna array, this poses in turn the question of how to distribute the signal energy on the ground. The trivial solution would be ampli- tude tapering, or as an extreme case, the use of only a part of the antenna for signal transmission, but this causes a significant loss of efficiency. Another possible solution is phase tapering, but the derivation of appropriate phase coefficients is an intricate task which requires in general complicated numerical optimization techniques.
SAR system application scenarios impose severe constraints: on-board compression tech- niques need to be relatively simple from the complexity and computational point of view, on the other hand the highest compression performance should be achieved. High perfor- mance and complex techniques for data reduction field are not applicable for spaceborne SAR, where on-board electric power is not available in large quantity. If considering a discrete time signal, the encoding is normally done sample by sample (or on block of sam- ples, as implemented by the BAQ, see Section 4.1.2), associating to each amplitude value a digital number according to discretization levels. This operation is the simplest encod- ing system and is known in communication systems as Pulse Code Modulation (PCM). Increasing the complexity, it is possible to consider the Differential Pulse Code Modula- tion (DPCM), which introduces the concept of differential coding. This means that the to-be-quantized signal fora sample s[n] will be then the difference between s[n] and the previous sample s[n − 1].
A spaceborne SAR system is used in the wide spectrum of the remote sensing owing to its wide-area observation ca- pability and high resolution. A modern SAR system based on the phased-array antennas realizes the multiple SAR operation such as the stripmap, spotlight and scan mode operation. However, this system does not fulfill the users demands for SAR data fora wide area with the fine re- soultion since there is a foundamental trade-off between the ambiguity and the swath width. Additionally, the life- time of a satellite depends on the fuel budget to maintain its orbit and motion control. A SAR sensor observing the wide area can reduce the revisit time and thereby extend the operation period. Therefore, the simultaneous achieve- ment of the wide coverage and fine spatial resoultion is quite attractive not only for the SAR users, but also in the system point of view. Such a simultaneous observa- tion of a wide area with fine resolution provides the use- ful information, especially for the dynamic target surveil- lance of oceans, ice and artificial moving targets. On a bi- or multi-static configuration this fundamental restric- tion can be resolved by introducing an appropriate DBF technique. The HRWS SAR concept exploits the smart an- tenna technique on 2 dimensional array antenna constella- tion in order to compensate the azimuth ambiguity result- ing from reduced PRF. In this paper we focus on the DBF SAR performance in azimuth directio. Several DBF algo- rithms were proposed with respect to a spectral estimation in spatial frequency domain,,. In this paper, we introduce the experimental results through a simplified 2- dimensional measurement. Numerous array configurations
Whereas the predictions fora correction with the translation approach have already been validated fora PS-InSAR time series (B¨ ahr et al., 2012), the reference frame effect itself has not been explicitly observed yet. Nevertheless, its existence is evident without explicit verification, and a correction is advisable whenever large-scale deformation phenomena are measured with high accuracy requirements. Based on the comparison in figure 7.8, a dedicated Euler rotation of orbital state vectors can be considered sufficiently accurate in most cases. Also a homogeneous translation derived from one representative ITRF velocity performs well but may involve minor phase artefacts. It should be preferred nevertheless in regions where the approximation quality of plate kinematic models is bad. The general transformation approach may be worth considering whenever InSAR measurements are used to densify an existing (GNSS) velocity field. If the benefit of the complex transformation outweighs the involved effort in this case still needs to be investigated.
An alternative to the sequential transmission from multiple azimuth apertures is the formation of multiple narrow azimuth beams in the transmitter. A sequence of full-bandwidth chirp signals is then transmitted while switching between different azimuth beams from sub-pulse to sub-pulse, as illustrated in Figure 10 on the left. This specific illumination sequence results for each point on the ground in multiple and mutually delayed chirp signal returns. If we consider now a scatterer at a given range, one will at each instance of time only receive the scattered signal from one sub-pulse while the other sub-pulses lead to a superposition of the received signal with range ambiguous echoes from scatterers located at different ranges. These different ranges are in turn associated with different look angles in elevation. It is hence possible to suppress the ambiguous returns from different ranges by digital beamforming on receive in elevation which enables a clear and unambiguous separation of the received echoes from the different azimuth beams (cf. Figure 10, right). The echoes from multiple azimuth beams are finally combined coherently to recover the full Doppler spectrum forhigh azimuth resolution. This combination is equivalent to a signal reconstruction from a multi-channel bandpass decomposition, where the individual bandpass signals correspond to narrow band azimuth spectra with different Doppler centroids. A detailed description and the corresponding processing algorithms can be found in .
raw data signal (c.f. section 2.4 ) after range compression (c.f. section 2.6). This, for the ﬁrst time, includes a comprehensive mathematical treatment for real aperture imaging radar using multiple transmit antennas. Depending on the antenna conﬁguration and the imaged area an additional near ﬁeld cor- rection could become necessary, which is investigated in detail in section 5.3. In addition, the antenna placement should be such as to prevent angular am- biguities; section 5.4 gives a novel investigation of this issue for non-uniform element placement. The operation and advantage from utilizing pulse coded chirp waveforms is described in section 5.5. In order to understand the basic parameter dependencies, the special case of a uniform linear array is treated in section 5.6, which results in compact expressions for the compressed signal, where the inﬂuence of system parameters can directly be related to the perfor- mance. Finally a MATLAB c based simulator is presented in section 5.7 which implements the described reconstruction algorithm. Example scenarios are in- put to the simulation tool in order to verify the analytical results obtained in this chapter.
18.104.22.168 System complexity
Fora geo-synchronous SAR, a significant part of the cost has to be ascribed to the launch. Taking into account this aspect, the mass budget plays a major role in the system design. Therefore the key parameters to measure the complexity of a given system design are the size of the antenna (in both monostatic and bi-static systems) and the amount of radiated power (monostatic systems only) that is proportional to the size of the solar panel required. This choice is based also on the assumption that RF components working at high frequencies up to K band will become COTS in the near future. This is clearly a simplification that does not include consideration on hardware complexity. This is acceptable in this preliminary stage of the trade-off.
A number of algorithms for focusing the raw SAR data have been developed since its debut in 1950s. Most of them were originally developed for remote sensing app- lications. Therefore some approximations and assumptions due to the restrictions of the hardware, e.g. antenna beamwidth, squint angle, the computation ability of signal processor, have been made when focusing the raw SAR data. The most accurate SAR algorithm is the time-domain correlation algorithm , which can be used to pro- cess the SAR data acquired with arbitrary beam-width and space sampling trajectory. Neither restriction on the antenna’s beamwidth nor motion compensation need to be applied. However, it requires a very high computational effort, which makes it im- practical for any true real-time SAR application given the performance of a realistic hardware environment [43, 44]. By restricting the space sampling interval to uniform, frequency-domain algorithms, which perform the focusing in frequency domain with the power of FFT, can substantially reduce the computational burden.
The HRWS system concept assumes a wide area illumination by a separate transmit antenna. This enables an independent electrical design and optimization of the transmit and receive paths, but it requires also the accommodation of an additional antenna on the spacecraft and reduces the flexibility to operate the radar system in different SAR imaging modes like ultra- wide-swath ScanSAR, high SNR spotlight, or new hybrid modes to be discussed later. It is hence worth to consider also the application of digital beamforming techniques in radar sys- tems that use the same antenna array for both the transmission and reception of radar pulses, thereby taking advantage of al- ready existing space-qualified T/R module technology. Since the high-resolution wide-swath SAR imaging capability is es- sentially based on a large antenna array, this poses in turn the question of how to distribute the signal energy on the ground. The trivial solution would be amplitude tapering, or as an ex- treme case, the use of only a part of the antenna for signal transmission, but this causes a significant loss of efficiency. Another possible solution is phase tapering, but the derivation of appropriate phase coefficients is an intricate task which re- quires in general complicated numerical optimization tech- niques.
The Sea Sate Processor (SSP) was developed for fully automatic processing of high resolution SyntheticApertureRadar (SAR) data from TerraSAR-X (TS-X) satellites and implemented into the processing chain for Near Real Time (NRT) services in the DLR Ground Station “Neustrelitz”. The NRT chain was organised and tested to provide the processed data to the German Weather Service (DWD) in order to validate the new coastal forecast model CWAM (Coastal WAve Model) in the German Bight of the North Sea with 900m horizontal resolution. The NRT test-runs, wherein the processed TS-X data were transferred to DWD and then incorporated into forecast products reach the best performance about 10min for delivery of processed TS-X data to DWD server after scene acquisition.
In Baraldi & Parmiggiani [ 1995 ], the use of GLCM’s texture parameters for the texture characterization in optical satellite images (NOAA AVHRR) have been investigated. It has also been commented that in certain cases gray level diﬀerence histogram (GLDH) outperforms the GLCM when considering the overall texture measurement accuracy, along with computer storage and computation time. The texture features generated using the GLCM have been successfully applied to the texture analysis for mapping sea ice patterns in the low-resolution ERS-1 SAR images in Soh & Tsatsoulis [ 1999 ]. The combined use of the GLCM texture parameters with Markov random ﬁelds (MRFs) texture features have been suggested to improve the classiﬁcation of SAR sea-ice imagery in Clausi [ 2000 ], later extensive comparative study of GLCM and MRFs based texture features have been presented in Clausi & Yue [ 2004 ]. The use of GLCM in eﬃcient texture analysis in ENVISAT-ASAR images by introducing the notion of patch re-occurrences has been proposed and compared with the texture features obtained using Gabor expansion in Kandaswamy et al. [ 2005 ]. The texture features based on GLCM generally outperforms other classical methods of classiﬁcation, at least in the case of low-resolution SAR and optical satellite images. For SAR images, the use of GLCM is largely limited to the sea-ice mapping and clustering application only. We will investigate the usability of GLCM to the land-cover classiﬁcation as well as to the patch-oriented image categorization to the very-high-resolution SAR images only.