Masafumi Iwamoto

A Two-Dimensional Bandwidth Extrapolation Technique for Polarimetric Synthetic Aperture Radar Images

The resolution of a synthetic aperture radar (SAR) image, in range and azimuth, is determined by the transmitted bandwidth and the synthetic aperture length, respectively. Various superresolution techniques for improving resolution have been proposed, and we have proposed an algorithm that we call polarimetric bandwidth extrapolation (PBWE). To apply PBWE to a radar image, one needs to first apply PBWE in the range direction and then in the azimuth direction, or vice versa . In this paper, PBWE is further extended to the 2-D case. This extended case (2D-PBWE) utilizes a 2-D polarimetric linear prediction model and expands the spatial frequency bandwidth in range and azimuth directions simultaneously. The performance of the 2D-PBWE is shown through a simulated radar image and a real polarimetric SAR image

Three-Dimensional Target Geometry and Target Motion Estimation Method Using Multistatic ISAR Movies and Its Performance

Inverse synthetic aperture radar (ISAR) is one of the radar techniques used to observe 2-D images of a remotely based target using radio waves. If we keep observing the target and consecutively generate multiple ISAR images, which we call an ISAR movie, the target image varies considerably due to the motion of the target. The authors have proposed an algorithm for reconstructing a 3-D target shape from an ISAR movie; however, the algorithm requires a priori knowledge of the relative motion of the target. In this paper, we propose a novel method that estimates the relative motion and the 3-D shape of the target using a multistatic ISAR movie.

A novel algorithm for reconstructing three-dimensional target shapes using sequential radar images

Inverse synthetic aperture radar (ISAR) is one of the radar techniques used to observe two dimensional images of a remotely based target using radio waves. Since the cross range vector is ambiguous, estimating the three-dimensional shape of the target is difficult. However, it is possible to estimate the three-dimensional shape using moving pictures taken by a video camera, if the relative movement between the camera and the target is known. In this paper we propose an algorithm for reconstructing three-dimensional shape of the target by applying a similar technique to the ISAR images. The theory and the required condition of the proposed method are described and some results obtained by the computer simulations are shown.

A bandwidth extrapolation technique of polarimetric radar data and a recursive method of polarimetric linear prediction coefficient estimation

Resolution of a radar is limited in range by its bandwidth. One of the various super-resolution techniques for improving its resolution is bandwidth extrapolation (BWE). In the technique, a linear prediction model is fitted to the data, and the model is used to extrapolate the bandwidth. In this paper, the concept of BWE is extended, and a new algorithm called polarimetric bandwidth extrapolation (PBWE) applicable to polarimetric radar data is proposed. It is shown through numerical simulations that utilization of full polarization information allows PBWE to improve the resolution beyond the conventional BWE method. Some results of physical simulation experiment using a W-band polarimetric FMCW radar and corner reflectors are shown to confirm the advantage of PBWE.

A bandwidth extrapolation technique for improved range resolution of polarimetric radar data

Radar resolution is limited in range by its bandwidth. One of the various super-resolution techniques for improving resolution is bandwidth extrapolation (BWE). In this paper, the concept of BWE is extended, and a new algorithm called polarimetric bandwidth extrapolation (PBWE), applicable to polarimetric radar data, is proposed. It is shown through simulations that utilization of full polarization information allows PBWE to improve the resolution beyond the conventional BWE method.

Image based approach for target detection and robust target velocity estimation method for multi-channel SAR-GMTI

This paper presents new algorithms for moving target detection and velocity estimation for multi-channel SAR (Synthetic Aperture Radar) system. The algorithms are derived as the extensions of conventional DPCA (Displaced Phase Center Antenna) and ATI (Along Track Interferometry) to the multichannel case. The proposed multi-channel DPCA based on orthogonal projection completely suppresses the static clutter even with the presence of the azimuth ambiguity. The multi-channel ATI utilizes the phase differences of all the pairs of the channels to estimate the target radial velocity. The combination of the two, which we call multi-channel DPCA-ATI, is also proposed as the more robust velocity estimation method.

A simulator of synthetic aperture radar images; land, ocean surface, and man-made targets

Synthetic aperture radar (SAR) images are very useful for many applications. However, SAR images require a great deal of experience for the interpretation, since the appearances are very different from optical images. Additionally, the appearances change significantly by the observational condition, such as frequency and incident angle. We propose a SAR image simulating algorithm developed for the aid of the SAR image analysis and training. The simulator is also intended to support the SAR system configuration design by providing the expected images. The main features of the algorithm are the capability of wide area simulation and the low computational costs. The simulated image of 10 km square area is shown and compared with the real image

Analysis on the Resolution of Polarimetric Radar and Performance Evaluation of the Polarimetric Bandwidth Extrapolation Method

Polarimetric radar provides useful information on scattering mechanisms, and advances in polarimetric synthetic aperture radar (POLSAR) technology have shown the effect of polarization information on various applications. The disadvantage of POLSAR is low azimuth resolution capability. Since the effective pulse repetition frequency (PRF) in each polarimetric channel of POLSAR is half the total PRF, the resolution is twice as low as the single-polarization SAR. This drawback may be mitigated by employing a superresolution method designed for polarimetric radar such as polarimetric bandwidth extrapolation (PBWE), previously proposed by the authors. Thus far, the authors only empirically showed that polarization information helps in improving resolution via numerical simulations. In this paper, theoretical analysis on the resolution of a polarimetric radar is provided. Along with the resolution, estimation accuracy of the polarization property of the target is studied. In the analysis, we employ the recently proposed metric called the statistical resolution limit (SRL) that provides the highest resolution achievable by any unbiased parametric spectral estimator. The SRL of polarimetric radar is derived and compared with that of single-polarization radar. The Cramér-Rao bound (CRB) of the polarization estimation error of the polarimetric radar is also derived. The complexified full Fisher information matrix and the CRBs of signal parameters are derived first for the general multichannel signal in nonwhite Gaussian noise. Then, the SRL and the polarization estimation error are derived using the CRBs of signal parameters for the cases where two point targets are closely located. It is shown analytically and numerically that the polarization information helps in improving the SRL. It is also shown that the polarimetric processing significantly improves the accuracy of the polarization estimation when two targets are located closer than the Fourier resolution cell. Finally, the performance of the PBWE is evaluated via simulations and is compared to the SRL and the CRB for the polarization estimation.

Evaluation of elevation derived from interferometric SAR data with DEM

Interferometric SAR is a stereoscopic method to obtain contours of ground elevation. A lot of studies have been done to derive the elevation map from SAR data, however, only few studies have been reported to examine the difference between the elevation map and DEM (Digital Elevation Model). In this paper, the authors compare the surface elevation map derived from interferometric SAR data with DEM, and evaluate the error distribution of the elevation map. The root mean squared error is found to be 33 m, and it agrees with the absolute vertical accuracy of DEM. The error distribution is also evaluated, and is verified to be approximately Gaussian. The large errors tend to appear locally at shadowed or steep sloped area, where phase noise or phase unwrapping error is large.