Structure-from-motion (SfM) is the concept of estimating both sensor pose and three-dimensional (3D) scene structure from input images. The difficulty with synthetic aperture radar (SAR) SfM stems from the non-linearity of radar imaging, making it hard to decouple and calculate radar pose and structure. Existing methods typically address this issue by simplifying the imaging process or introducing auxiliary data. In this paper, we propose to jointly solve radar pose and structure by formulating an optimization problem, which only needs two-dimensional (2D) SAR observations as input. The objective function is written as the summation of all reprojection errors established by the precise range-Doppler (RD) imaging model, with radar poses embedded into the transformations between the sensor and scene coordinate systems. The rotational components in radar poses are represented as matrices and constrained on the special orthogonal groupSO(3). A special orthogonal Riemannian conjugate gradient algorithm (SO-RCG) is then proposed to solve the optimization problem. The proposed algorithm preserves the orthogonality of rotation matrices and updates all variables iteratively. Furthermore, we deduce that only five degrees of freedom (DoFs) in radar pose are involved in determining the imaging results of targets. We also discuss the ambiguity issue in multiview SAR observation leading to local minimums and introduce strategies against ambiguity. Experimental results on real-measured and simulated datasets show that the proposed algorithm is effective on both near- and far-field cases, and can be applied to both side-looking and squint SAR.

This paper proposes a fast two-step method which aims to address the SAR pose estimation problem and enable Unmanned Aerial Vehicles (UAVs) the self-localization capability under harsh conditions. Firstly, the monocular SAR pose estimation is formulated as a Circular-n-Point (CnP) problem based on frequency-domain SAR imaging mechanism. Then, we propose to decouple the motion components of radar platform and solve the overdetermined equations of direction and location sequentially. Experimental results validate the effectiveness, accuracy, and efficiency of the proposed method.

Structure-from-Motion (SfM) is the process of estimating 3D scene structure and sensor pose from a set of 2D inputs. In this paper, we present a matrix factorization scheme for solving SAR SfM problem. First, SAR imaging model is linearized at local scene and the SfM problem is converted to a problem that decomposes data matrices into the product of two kinds of matrices, one of which satisfies Stiefel constraint. We then propose a Riemannian conjugate gradient descent algorithm leveraging the Stiefel constraint. SAR SfM is solved by the proposed algorithm via alternating iteratively estimating radar pose and scene structure. Finally, the feasibility and accuracy of the proposed scheme are verified through experiments.

A working video synthetic aperture radar (VideoSAR) can obtain continuous observations of a region of interest from different viewpoints, meaning those observations naturally contain 3D information for positioning. This paper reports our preliminary studies on SAR scene structure-from-motion using several frames extracted from a VideoSAR sequence. In contrast to classic stereo-radargrammetric workflow, the radar pose is estimated only with information measured by radar. The coordinates of each scattering point are calculated by least-square optimization. We also develop an affine transformation aware dense matching framework to accurately measure pixel-level correspondence between frames. Experiments on real VideoSAR data demonstrate the validity of our approach.

Moving targets can be detected in SAR images using the difference between two looks because, in the two looks, the images of a stationary target are similar, but the images of a moving target are different. However, the sidelobes in the two looks deteriorate the similarity of the stationary ground between the two looks and thus degrade the performance of this method. In this paper, a scaled Hanning window is designed to generate the two looks with lower sidelobes. The shape parameter of the window is automatically chosen according to the spread of the Doppler spectrum. The results show that this method has a better performance in the suppression of stationary targets than the original method.

The along-track interferometry (ATI) uses the phase difference between the images from different channels to detect moving targets in multi-channel SAR systems. However, the phase differences of weak clutters have large variances, and this may cause false alarms. This paper presents an improved ATI method. In this method, the phase difference is weighted by the normalized interferometric amplitude to detect moving targets. Since the phase differences of weak clutters are suppressed, the false alarm rate is decreased. The results of real data show the effectiveness of this algorithm.

A novel algorithm is presented for warhead recognition in the defense of ballistic missiles. The deviation of a range profile from a standard range profile is measured by the 2-norm of the difference between the two range profiles. A target is recognized as a warhead if the deviation of its range profile from the range profile of the warhead is less than a threshold. The threshold is chosen such that the detection rate is constant. Since the threshold is automatically calculated according to the signal- to- noise ratio and the detection rate, this algorithm has a larger applicability than the traditional methods based on range-profile matching.

In order to overcome the disadvantage that traditional multichannel SAR systems are only sensitive to the radial velocities of moving targets, a variable-boresight mode has been proposed, whose kernel is to detect and estimate the velocity component of the azimuth velocity at two nodes. For purpose of improving the estimation accuracy of the azimuth velocity, we introduce more nodes with different squint angles, so as to obtain more velocity components to carry out the estimation by the least square method. Considering that only few tiny angles can be ignored, while introducing more nodes are likely to bring about nonnegligible squint angles, the previously used signal processing suitable for side-looking mode is modified. Finally, numerical experimental results are provided to verify the effectiveness.

Sidelobe suppression is an issue of interest in SAR imaging. In order to obtain low-sidelobe SAR images, various sidelobe suppression methods have been proposed. SVA (Spatially Variant Apodization) method can both maintain image resolution and suppress sidelobes. However, this method assumes that the Doppler centroid is zero and the Doppler bandwidth is close to the pulse repetition frequency (PRF). Actually, when the radar has a radial velocity, the Doppler centroid may deviate from zero and the Doppler bandwidth may be significantly smaller than the PRF. In such case, the method has poor performance in suppressing sidelobes. This new algorithm proposed in this paper not only inherits the advantage that the SVA maintains image resolution, but also considers arbitrary Doppler centroids and bandwidths. The result of simulated and real SAR images indicates the advantage of this new algorithm over the traditional methods.

A novel scheme is proposed for the velocity estimation of ground moving targets in multi-channel synthetic aperture radar (SAR) systems. This scheme exploits the statistics of the interferometric phase. In fact, a moving target consists of multiple point scatterers instead of only one. Therefore, the velocity estimation should be based on the interferometric phase of all the scatterers rather than only one. It is proved that the probability density function of the interferometric phase is a Gaussian function, whose peak corresponds to the radial velocity of the moving target. The peak is estimated via a smoothed histogram of the interferometric phase. Finally, the effectiveness of our proposed scheme is verified by simulated and real airborne SAR data.

With more degrees of freedom in the along-track axis, multichannel SAR systems are widely investigated for the purpose of ground moving target indication and motion parameter estimation. However, since conventional multichannel SAR system usually works at the side-looking mode and only the radial velocity is considered, it fails to detect targets only moving in azimuth. In this paper, the variable-boresight configuration is designed for multichannel SAR and overlapping imaging mode is applied to detect targets moving in azimuth. Besides, this system can also provide the precise velocity information of the moving target. Finally, the effectiveness of the proposed approach is verified with simulated multichannel SAR data.

This paper presents a new scheme for moving target velocity estimation in conventional high-resolution wide-swath (HRWS) SAR systems. A generalized steering vector is designed for azimuth ambiguity suppression and the velocity parameter of the moving target is involved in this modified steering vector. The low PRF in HRWS-SAR systems will result in serious azimuth ambiguity. When the adopted parameter is matched with the real velocity of the moving target, its azimuth ambiguities will be suppressed clearly. Entropy can be used to measure the performance for azimuth ambiguity suppression. The velocity can be estimated through searching the matched velocity. Finally, the effectiveness of the proposed approach is verified by simulated multichannel SAR data.

This letter presents a new scheme for ground moving target indication in high-resolution wide-swath (HRWS)-synthetic aperture radar (SAR) systems. The asymmetry of the Doppler spectra is measured to extract the range bins with moving targets. To improve the computational efficiency, only the extracted range bins are used to restore the unambiguous Doppler spectra. The two-look processing technique is then applied to generate two looks and moving targets are indicated by comparing the difference between the two looks. In this detection scheme, the configuration of the conventional HRWS-SAR system remains unchanged and no additional receiving channels are needed. The experimental results show the effectiveness of this detection scheme.

In this paper, a new scheme is presented for ground moving target indication for multichannel synthetic aperture radar (SAR) systems using the asymmetric distribution of the spatial spectrum. Two feature images are generated via accumulating energy intensities of the positive and the negative spatial spectrum, respectively. In the two images, the stationary backgrounds have the same intensity while moving targets are not due to their across-track velocities. Coherent subtraction is carried out to implement clutter suppression and potential moving targets can be indicated through the constant-false-alarm-rate (CFAR) test. When dealing with slow weak moving targets, this developed approach can still obtain good detection performance. Finally, the effectiveness of the proposed approach is verified by real airborne SAR data.

An automatic DPCA technique is presented for SAR Ground Moving Target Indication (GMTI). We note that there exists a shift and a phase difference between the images from two channels. Therefore, SAR-GMTI can be implemented in the following steps: Image registration, phase compensation, image subtraction, and CFAR detection. In our technique, these steps are carried out automatically, and thus no precise information is needed about the length of the baseline and the velocity of the platform. We utilize a set of real data to demonstrate the accuracy and the robustness of our algorithm.

In this paper, a novel synthetic aperture radar (SAR) imaging algorithm based on two-dimensional (2-D) compressed sensing (CS) is presented. First, a complex image is generated by applying a traditional imaging method to raw echoes. Then, the basis pursuit (BP) method is applied to a CS framework to obtain the scattering coefficients of target points from the complex image. Since CS is applied jointly in range and azimuth, this algorithm utilizes the sparsity of the target scene fully. Experimental results using the real field data demonstrate the effectiveness of our algorithm.

A new algorithm is presented to indicate the ground targets moving in azimuth in synthetic aperture radar (SAR) images. This algorithm is based on local two-Look similarity. In the two looks of the scene, a stationary target has the same positions, but a moving target has different positions. Therefore, moving targets can be indicated by the patch-by-patch similarity detection between the two looks. We use normalized correlation to characterize similarity. In the similarity detection, for each patch in the first look, only the normalized correlations with three patches in the second look are calculated, and thus this algorithm is computationally efficient.

A new algorithm is presented to indicate ground moving targets in synthetic aperture radar images. Two filters, which differ only in the signs of the phase responses, are used to defocus the complex image respectively. In the two defocused images, each stationary target is blurred to the same extent, but each moving target is blurred to different extents. Therefore, moving targets can be indicated by the patch-by-patch sharpness comparison of the two defocused images. The results of the simulated and real data show that this algorithm is effective and efficient.

It is necessary to measure the sharpness of distributions in many situations. A class of functions is investigated in this paper. First, the relation between this class and sharpness is clarified, and this justifies this class as sharpness measures. Then, we analyze the performance of different sharpness measures and present a guide to select the sharpness measure. In addition, the relation of this class to the sparsity measure is addressed, which leads to a deeper understanding about sparsity. Finally, we show and discuss the application of this class in inverse synthetic aperture radar imaging.

In this paper, we present a novel SAR (synthetic aperture radar) imaging algorithm based on compressed sensing (CS). We first obtain complex SAR images from raw SAR echoes, and then search all targets in complex-image domain by an improved MP (matching pursuit) algorithm. Unlike other CS-based imaging models, our algorithm has several advantages: 1) our algorithm is a two-dimensional (2-D) imaging system, rather than one-dimensional (1-D) imaging models introduced in most existing CS-based imaging algorithms; 2) unlike low efficiencies of most reconstruction algorithms, our method has an efficient reconstruction rate; 3) our imaging model applies existing SAR imaging systems, and inherits current SAR imaging results. The experimental results using real SAR echo data demonstrate the effectiveness of our algorithm.

In this paper, we propose a CS-based SAR imaging algorithm after analyzing the model of echoes from point target. Our reconstruction algorithm is two-dimensional, unlike current one-dimensional CS-based SAR imaging algorithms. It can reconstruct targets with high resolution from relatively small number of echoes and effectively improves the efficiency of reconstruction. The processing results of simulated data demonstrate the effectiveness of our algorithm.

A new scheme is presented for the velocity estimation of moving targets in synthetic aperture radar (SAR) imaging. First, the moving target is imaged using a range-Doppler algorithm, where clutter lock, range correction and focus filtering are automatically done. Then, the parameters obtained in this algorithm are used to estimate the velocity of the moving target. The Doppler centroid is used to estimate the velocity of the moving target in range. The out of-focus coefficient is used to estimate the velocity of the moving target in azimuth. Especially, the ambiguity of the Doppler centroid is resolved according to the range correction coefficient and the out-of-focus coefficient. This scheme is accurate and computationally efficient.

A new technique is developed for the automatic correction of range migration in synthetic aperture radar imaging. In the range-Doppler domain, samples at a Doppler frequency constitute a Doppler slice. It is noted that, when there is no range migration, the Doppler slices have similar envelopes, i.e., the envelope of one Doppler slice roughly equals the envelope of another Doppler slice multiplied by a constant. Thus, range migration can be corrected by shifting the Doppler slices such that their envelopes are similar. The technique applies even when the radar moves irregularly or the target is moving.

This paper presents an improved step transform subaperture approach for spotlight Synthetic Aperture Radar (SAR). The step transform algorithm will obtain desired high azimuth resolution only if the linear FM rate and the sampling rate of the SAR signal satisfy a certain constraint. In order to obtain a high quality image, the minimum entropy autofocus using an adaptive order polynomial model is extended and applied before the coherent summation step in this paper. The extended autofocus processing not only compensates quadratic and higher order phase errors but also shifts the coarse-resolution response functions to their correct positions. The simulation results of 1-dimension and 2-dimension are given, and the validity of this algorithm is demonstrated.

A new technique is presented for range-migration correction in SAR imaging. In the range-Doppler domain, the samples at a Doppler frequency constitute a Doppler slice. Different Doppler slices have similar envelopes. Based on this similarity, the Doppler slices are shifted and aligned in range to correct range migration. The technique applies even if the prior knowledge about the relative motion between the radar and the target is unavailable.

The time-frequency method is a promising phase adjustment technique for inverse synthetic aperture radar (ISAR). Usually, the time-frequency method estimates the translational Doppler frequency and thus the translational Doppler phase by detecting the peaks of the time-frequency representation at different times. Unfortunately, it has some defects, such as the sensitivity to noise and target scintillation, and the aliasing of the time-frequency representation due to undersampling. In order to remove these limitations, we develop an improved time-frequency method. It estimates the translational Doppler frequency from the envelope correlation of the instantaneous slices of the time-frequency representation. Compared with the traditional time-frequency method, this technique can estimate the translational Doppler frequency more accurately. Moreover, we develop a preprocessing method to avoid the aliasing of the time-frequency representation due to undersampling.

A new algorithm is presented for autofocus in synthetic aperture radar imaging. Entropy is used to measure the focus quality of the image, and better focus corresponds to smaller entropy. The phase response of the focus filter is modeled as a specially designed polynomial, and the coefficients of this polynomial are adjusted in sequence to minimize the entropy of the image. Because the order of this polynomial is adaptive, this algorithm applies more widely than the minimum-entropy algorithms with a fixed-order polynomial model

The global range alignment has a good performance in inverse synthetic aperture radar (ISAR) imaging. The shifts made to the echoes are modeled as a polynomial, and the coefficients of this polynomial are chosen to optimize a quality measure of range alignment. The shift in the time domain is implemented by introducing a phase ramp in the frequency domain in order to remove the limitation of integer steps. However, in this method, the quality measure of range alignment need be calculated in the time domain, which is time consuming. In this paper, we present a new expression for the quality measure of range alignment so that it can be calculated directly in the frequency domain. This raises the computational efficiency of the algorithm significantly.

A new idea is presented to correct range migration in SAR imaging. In the range-Doppler domain, all the samples at a given Doppler frequency constitute a Doppler slice. Different Doppler slices are found to have similar envelopes. According to this similarity, the Doppler slices are shifted in range to correct range migration. This technique applies even without the prior information about the relative motion between the radar and the target