Diffusion-based image compression has demonstrated impressive perceptual performance. However, it suffers from two critical drawbacks: (1) excessive decoding latency due to multi-step sampling, and (2) poor fidelity resulting from over-reliance on generative priors. To address these issues, we propose SODEC, a novel single-step diffusion image compression model. We argue that in image compression, a sufficiently informative latent renders multi-step refinement unnecessary. Based on this insight, we leverage a pre-trained VAE-based model to produce latents with rich information, and replace the iterative denoising process with a single-step decoding. Meanwhile, to improve fidelity, we introduce the fidelity guidance module, encouraging output that is faithful to the original image. Furthermore, we design the rate annealing training strategy to enable effective training under extremely low bitrates. Extensive experiments show that SODEC significantly outperforms existing methods, achieving superior rate-distortion-perception performance. Moreover, compared to previous diffusion-based compression models, SODEC improves decoding speed by more than 20×.

Pretrained latent diffusion models have shown strong potential for lossy image compression, owing to their powerful generative priors. Most existing diffusion-based methods reconstruct images by iteratively denoising from random noise, guided by compressed latent representations. While these approaches have achieved high reconstruction quality, their multi-step sampling process incurs substantial computational overhead. Moreover, they typically require training separate models for different compression bit-rates, leading to significant training and storage costs. To address these challenges, we propose a one-step diffusion codec across multiple bit-rates. termed OSCAR. Specifically, our method views compressed latents as noisy variants of the original latents, where the level of distortion depends on the bit-rate. This perspective allows them to be modeled as intermediate states along a diffusion trajectory. By establishing a mapping from the compression bit-rate to a pseudo diffusion timestep, we condition a single generative model to support reconstructions at multiple bit-rates. Meanwhile, we argue that the compressed latents retain rich structural information, thereby making one-step denoising feasible. Thus, OSCAR replaces iterative sampling with a single denoising pass, significantly improving inference efficiency. Extensive experiments demonstrate that OSCAR achieves superior performance in both quantitative and visual quality metrics. The code and models are available at https://github.com/jp-guo/OSCAR/.

Over-the-horizon radar (OTHR) utilizes the ionospheric refraction and reflection in high-frequency band for air-sea targets detection. However, non-stationary ionospheric dynamics induce inhomogeneous distortions in sea clutter and targets, significantly degrading detection and localization performance. To address this challenge, we present a comprehensive investigation on OTHR full-link modeling and simulation to systematically analyze the impact of various trans-ionospheric propagation effects on echo signals. In Part I, we develop a unified framework for full-link modeling of sea clutter that incorporate background ionospheric and oceanic conditions, enabling simulation and analysis of sea clutter characteristics. Firstly, we establish the models for radar cross section of sea clutter and ionospheric propagation effects. Key parameters, including the wave propagation range, actual range, group delay, phase disturbance, and propagation loss, are calculated based on the Appleton-Hartree formula and ray tracing technique. Secondly, we construct the echo signal models in the fast- and slow-time domain and derive the corresponding range-Doppler spectrum. The origin mechanism and intrinsic cause of Doppler shifting, broadening, and splitting, as well as range localization errors are theoretically analyzed. Finally, simulation experiments of three scenarios are designed to produce OTHR sea clutter data in sea and air modes, which are validated in comparison with the real data. The typical phenomena of Doppler shifting, broadening, and splitting observed in real data are reproduced, and the results indicate that the multi-mode propagation is the main obstacle to range localization and ionospheric decontamination. The full-link sea clutter model provides critical insights for the subsequent signal processing tasks including ionospheric decontamination, clutter suppression, target detection, and localization.

Synthetic aperture radar tomography (TomoSAR) is widely used for 3-D imaging in urban environments. Reference network (RN) technology has emerged as an effective solution in TomoSAR, leveraging a spatially correlated network of persistent scatterers (PSs) to mitigate the atmospheric phase screen (APS), a significant source of error in multipass spaceborne SAR systems. An effective RN should connect sufficient PSs with reliable edges, thereby ensuring comprehensive tomographic imaging of PSs and other scatterers around PSs. However, RN construction faces challenge in simultaneously addressing both network connectivity and quality during RN construction. To resolve this issue, this article introduces an innovative approach for constructing an optimal RN based on the minimum spanning tree (MST-RN). In this approach, the paradigm for determining the optimal RN is transformed into the problem of finding the MST in which the relative elevation estimation quality of the edge in RN is regarded as their weight in MST. Furthermore, standardized metrics are proposed to comprehensively evaluate the effectiveness of RN. The proposed MST-RN method demonstrates superior effectiveness on the metrics including overall connectivity quality, PS candidate (PSC) coverage, and network redundancy, compared to existing methodologies, which underscores the significant progress made in the RN construction. Finally, the effectiveness of the proposed method is validated through experiments conducted on the TerraSAR-X dataset acquired in Shenzhen, China, demonstrating accurate height estimation of PSs in urban areas.

Abstract Bayesian compressive sensing (BCS) is an important sub‐class of sparse signal reconstruction algorithms. In this paper, a modified complex multitask Bayesian compressive sensing (MCMBCS) algorithm using the Laplacian scale mixture (LSM) prior is proposed. The LSM prior is first introduced into the complex BCS framework by exploiting its better sparse characteristic and flexibility than traditional Laplacian prior. Furthermore, by integrating out the noise variance analytically, the MCMBCS algorithm significantly improves the signal recovery performance than the original CMBCS. More importantly, the authors not only present the iterative algorithm but also develop the sub‐optimal fast implementation method based on the marginal likelihood maximisation, which dramatically reduce the computational complexity. Finally, sufficient numerical simulations validate the better performance of the proposed algorithm in reconstruction accuracy and computational effectiveness than existing work. It is revealed that the proposed algorithm has great potential in the complex‐valued signal processing field.

The amplitude stripes induced by the ionospheric scintillation have been frequently observed in spaceborne L-band synthetic aperture radar (SAR) images, such as ALOS PALSAR, which may impede SAR applications. From another perspective, the presence of ionospheric stripes implies a reverse guidance for ionospheric sounding. In this paper, a methodology for correcting and measuring ionospheric stripes from SAR images is proposed. Firstly, based on prior information of the stripes’ orientation, the stripe components are automatically detected and extracted using a band-rejection spectral filter. Secondly, the extracted amplitude errors are further applied to measure ionospheric scintillation parameters by fitting the power spectrum. The methodology is tested on several PALSAR acquisitions. The results indicate that the methodology can automatically corrects the stripes and has a great potential of measuring ionospheric scintillation from the SAR images with amplitude stripes.

The phase scintillation induced by ionospheric irregularities can reduce azimuth decorrelation and cause defocusing in the low-frequency spaceborne synthetic aperture radar (SAR) imaging. Thus, it demands for an autofocus approach to correct the scintillation phase error (SPE) and to refocus the deteriorated SAR image. However, the azimuth variation of the SPE has been rarely considered and poorly tackled in previous research studies. In this letter, an ionospheric phase screen (PS) projection method of phase gradient autofocus (PSP-PGA) is proposed to deal with this issue and compensate the azimuth-varying SPE in SAR images. The PSP is used to project the SAR image onto the PS, by using the part decompression and spectral analysis, so as to realize the decoupling of the SAR data with SPE. Then, the projected image can be directly applied to the PGA to accurately estimate the SPE. The proposed PSP-PGA is validated by using several simulation processing experiments and its performance is finally evaluated with respect to the parameter of the PS height.

Spaceborne polarimetric synthetic aperture radar (Pol-SAR) systems operating at low frequencies, such as P-band, are significantly influenced by ionospheric Faraday rotation (FR). A novel theoretical model of the effect of FR dispersion on Pol-SAR images is proposed. Three components of the range compression response are derived with the main function accompanied by two accessories. Simulation results indicate that two accessory functions in existence of FR dispersion can be comparable with the main for P-band ultrawideband systems. It can not only lead to range imaging deteriorations but also bring about extra polarimetric distortions. Finally, an airborne P-band Pol-SAR real scene is used for validation.

The spaceborne low-frequency (L-band and below) synthetic aperture radar (SAR) is very susceptible to ionospheric scintillation. The scintillation phase error (SPE) is a vital factor that brings about the azimuth decorrelation and leads to the imaging degradation. In this article, a methodology is described, which exploits the accurate SPE estimation to measure scintillation parameters from SAR images. First, the phase gradient autofocus (PGA) is applied to achieve the believable SPE estimate from the local area that contains a strong scatterer. Second, based on the estimated staggered index of two SPE estimates, the irregularity altitude can be derived by solving an established equation using an iteration process. Third, three parameters, including the spectrum index, the outer scale, and the integrated turbulence strength, can be fitted when the theoretical spectrum expression mostly approximates the derived spectrum. The methodology is validated on an ALOS-2 PALSAR-2 spotlight data, and its azimuth imaging was seriously degraded by ionospheric scintillation. Two existent corner reflectors ensure the accuracy of SPE estimates from PGA. The measurement results indicate that the local ionospheric irregularities are extremely turbulent and located at a lower altitude of about 220 km in the F2-layer.

Domain adaptation (DA) aims at transferring knowledge from a labeled source domain to an unlabeled target domain. Though many DA theories and algorithms have been proposed, most of them are tailored into classification settings and may fail in regression tasks, especially in the practical keypoint detection task. To tackle this difficult but significant task, we present a method of regressive domain adaptation (RegDA) for unsupervised keypoint detection. Inspired by the latest theoretical work, we first utilize an adversarial regressor to maximize the disparity on the target domain and train a feature generator to minimize this disparity. However, due to the high dimension of the output space, this regressor fails to detect samples that deviate from the support of the source. To overcome this problem, we propose two important ideas. First, based on our observation that the probability density of the output space is sparse, we introduce a spatial probability distribution to describe this sparsity and then use it to guide the learning of the adversarial regressor. Second, to alleviate the optimization difficulty in the high-dimensional space, we innovatively convert the minimax game in the adversarial training to the minimization of two opposite goals. Extensive experiments show that our method brings large improvement by 8% to 11% in terms of PCK on different datasets.

An L-band geosynchronous synthetic aperture radar (GEO SAR) system has to be confronted by an intractable issue of the decorrelations imposed by ionospheric irregularities. On the one hand, the phase and amplitude scintillations will bring about the decorrelation within the synthetic aperture and result in azimuth-imaging degradation. On the other hand, the imposed scintillation history is spatially decorrelated across the ultra-large GEO SAR scene. In this article, a signal model of the GEO SAR acquisitionis established with the two-way ionospheric transfer function (ITF) modulation to incorporate these two types of decorrelations. This model meanwhile takes the anisotropic and flowing irregularities into account. By using this model, the L-band GEO SAR azimuth-imaging is evaluated in terms of five indexes, whose performances are dependent on nine ionospheric parameters. Furthermore, the spatial correlation of the phase and intensity scintillation histories is investigated for the L-band GEO SAR scene, both in simulation and statistics. The statistical result implies a sized scene, in which the phase scintillation history tends to be consistent. Finally, the interferometric performance is investigated between the pure and contaminated GEO SAR images. The simulation result shows that the degradation of the interferometric coherence results from the in-aperture decorrelation.

There has been a burgeoning prospect in developing a spaceborne P-band synthetic aperture radar (SAR) mission for its stronger penetrability through foliage and subsurface than the higher-frequency system. However, the transionospheric signals operating at P-band are more susceptible to scintillation impacts, which may bring about the decorrelation of the signal amplitude, phase, and frequency introduced by ionospheric irregularities. In this article, a comprehensive numerical model of the generalized ambiguity function (GAF) is established to evaluate SAR imaging deterioration. On the one hand, an improved two-frequency and two-position coherence function (TFTPCF) is integrated in the GAF to include the amplitude scintillation derived from the diffraction. On the other hand, the anisotropic characteristics of the irregular structure is introduced into TFTPCF by adopting the Rino's 2-D spectrum. Furthermore, the ambiguous resolution is redefined for a more strict numeration. Numerical analyses about the ionospheric coherence and an ambiguous resolution are performed to investigate their sensitivity to scintillation parameters. Results show that this model is capable of depicting more comprehensive effects of the anisotropic irregular ionosphere, which may exhibit a rod-like structure, including elongation by anisotropic scale, rotation by geomagnetic heading, and projection by geomagnetic inclination. At last, the signal-level simulations are operated to verify the effectiveness of numerical conclusions, which further confirm that the structural configuration of anisotropic irregularities has a complicated effect on spaceborne P-band SAR image resolution.

In this paper, the anisotropic generality of the ionospheric irregularities is incorporated in the generalized ambiguity function (GAF) to evaluate its impact on spaceborne P-band synthetic aperture radar (SAR) imaging. The configuration of the anisotropic ionosphere exhibits as a rod-like structure, which is elongated by anisotropic scale, rotated by magnetic heading and projected by geomagnetic inclination. Aiming at these three parameters, numerical analysis is implemented in terms of the coherence and ambiguous resolution. At last, signal-level simulation is carried out to validate the effectiveness of the numerical results.

As was indicated in the mentioned paper, the ionospheric stripes, in general, aligned well with the orientation of the projected ambient geomagnetic field vector. However, it is shown in our study that the calculated striping heading is not only dependent upon the orientation of the projected ambient geomagnetic field vector, namely, the geomagnetic heading but also the geomagnetic inclination, the system incident, and squint angle. It also confirms that the changing direction of the visible stripes in the mentioned Phased Array-type L-band Synthetic Aperture Radar (PALSAR) data is mainly due to the variation of the geomagnetic inclination, while the geomagnetic heading is nearly constant along the orbit. Therefore, the denotation and presentation in that research that the projected geomagnetic field vector elongates in the direction of the stripe orientation might not be feasible.

It is well known that spaceborne synthetic aperture radar (SAR) systems, which operate at L- and P-bands, are significantly influenced by the ionospheric effects. One of the severe effects is Faraday rotation (FR), which has a potential trend to cause the polarimetric distortion. Therefore, performing FR estimation is a prerequisite for the polarimetric SAR (PolSAR) application. A set of FR estimators have been proposed to solve this issue. Based on the Bickel and Bates's method, an improved FR estimator has been briefly mentioned, but has not been specified. The improved estimator's performance is verified by real SAR data, and experimental results indicate that the improved estimator shows the most robust performance in terms of system noise compared with other proposed estimators, which implies an applicable method for FR estimation.

The performance of spaceborne linearly polarized synthetic aperture radar (SAR) systems operating at low frequencies, such as L-band and P-band, is significantly affected by Faraday rotation (FR) effects. FR indicates the rotation of the polarization plane, which deteriorates the polarimetry accuracy of the polarization scattering matrix. A set of FR estimators have been proposed to solve this issue, but all estimators suffer from the ambiguity error of FR angle (FRA). Based on Bickel and Bates's estimator and the scatter distribution diagram of FR estimation, a novel correction method for eliminating FRA ambiguity error is proposed, which is divided into pixel-level correction and image-level correction. Experimental results verify the effectiveness of the novel correction method by processing the real data.