Due to a unique imaging mechanism, Synthetic Aperture Radar (SAR) images typically exhibit degradation phenomena. To enhance image quality and support real-time on-board processing capabilities, we propose a lightweight deep generative network framework, namely, the Lightweight Super-Resolution Generative Adversarial Network (LSRGAN). This method introduces Depthwise Separable Convolution (DSConv) in residual blocks to compress the original Generative Adversarial Network (GAN) and uses the SeLU activation function to construct a lightweight residual module (LRM) suitable for SAR image characteristics. Furthermore, we combine the LRM with an optimized Coordinated Attention (CA) module, enhancing the lightweight network’s capability to learn feature representations. Experimental results on spaceborne SAR images demonstrate that compared to other deep generative networks focused on SAR image super-resolution reconstruction, LSRGAN achieves compression ratios of $74.68\%$ in model storage requirements and $55.93\%$ in computational resource demands. In this work, we significantly reduce the model complexity, improve the quality of spaceborne SAR images, and validate the effectiveness of the SAR image super-resolution algorithm as well as the feasibility of real-time on-board processing technology.

Due to its capability for all-weather, all-time information acquisition, Synthetic aperture radar (SAR) plays a vital role in the field of Earth observation. However, the specificity of the radar sensor and the complexity of electromagnetic scattering imaging physics result in SAR images lacking the intuitiveness of optical images, making them unsuitable for interpretation by non-experts. A common approach to tackle this challenge is to use conditional generative adversarial network (cGAN) to translate SAR images into optical images, thereby enhancing readability and assisting non-experts in interpretation while filling the gaps in optical data due to acquisition constraints. Nevertheless, traditional cGAN-based methods are limited by inadequate global semantic information extraction and poor detail preservation, leading to translated images with incoherent texture and color, and blurred edge. To address these issues, we propose a hybrid vision transformer conditional generative adversarial network (HVT-cGAN) for SAR-to-optical image translation (S2OIT). In our proposed HVT-cGAN, the generator utilizes a convolutional stem for patch embedding and encoding. The parallel CNN branch and vision transformer branch are employed for the extraction and mapping of local and global information, respectively. Moreover, we propose a novel attention-based feature fusion module, named the Convolutional Attention Fusion Module (CAFM), which can adaptively aggregate local and global information from parallel branches by learning both channel-wise and spatial-wise relations. Benefiting from these improvements, our method achieves superior performance in both qualitative and quantitative comparisons with other methods on the SEN1-2 dataset. Additionally, the results of multiple ablation experiments validate the effectiveness of the proposed method.

TianQin is a geocentric space-based gravitational wave detecting mission that relies on three drag-free spacecraft with a mean inter-spacecraft distance of approximately 1.7 × 10⁵ km. Considering the efficiency requirement of orbit optimisation, we apply a gradient descent method based on variational gradient calculation and the damped Newton iterative method to find the most stable solution of the constellation near the given initial position and velocity of the spacecraft. The efficiency was improved and the number of orbital integrals required for gradient calculation was significantly reduced by over 83%. Based on this optimisation method, we analysed the influence of perturbation on the stability of the TianQin constellation, and Moon’s three-body perturbation was the main factor. In addition, the influence of the orbit elements on the optimisation results was studied. The influence of the initial mean anomaly can be ignored, and the stability of the optimised constellation decreases with an increase in the semi-major axis. Working in the ‘3 months on +3 months off’ observation scheme for 2 years, the optimisation result is the ideal direction when the normal direction of the orbital plane of the TianQin constellation points to RXJ0806.3 + 1527.

Satellite laser ranging (SLR) is the space geodetic technique with the highest degree of range, measuring precision and distances right down to the millimeter level. Thanks to the improvement of SLR station layouts and the advance of SLR technology, in recent years, more research has been conducted to determine Global Navigation Satellite System (GNSS) satellite orbits using SLR data. The primary goal of this contribution is to investigate the accuracy of BeiDou Navigation-3 (BDS-3) Satellite precise orbit determination (POD) using solely SLR data, as well as explore the impact of various factors on that accuracy. Firstly, we used actual SLR data to make the POD for BDS-3 satellites, and the POD accuracy was positively connected with the orbital arc lengths. The 9-day median root mean square (RMS) in radial (R), along-track (T), and cross-track (N) directions were estimated at 4.7–8.2, 22.1–35.2, and 27.4–43.8 cm, respectively, for comparison with WUM precise orbits. Then, we explored the impact of SLR observations and stations on POD accuracy. For 9-day orbital arc lengths, five station or 20 observation arcs may offer an orbit with a 1 m precision. Six to eight stations or 30–35 observation arcs allow an improved orbit accuracy up to approximately 0.5 m. Furthermore, we examined how measurement errors and orbit modeling errors affect the SLR-only POD accuracy using simulated SLR data. For orbital arc lengths of 9 days, each cm of random error leads to a 9.3–11.0 cm decrease in orbit accuracy. The accuracy of an orbit is reduced by 10.1–15.0 cm for every 1 cm of systematic error. Moreover, for solar radiation pressure (SRP) errors, the effect of POD accuracy is 20.5–45.1 cm, respectively.

Compared to traditional real-valued neural networks, which process only amplitude information, complex-valued neural networks handle both amplitude and phase information, leading to superior performance in polarimetric synthetic aperture radar (PolSAR) image classification tasks. This paper proposes a multi-scale feature extraction (MSFE) method based on a 3D complex-valued network to improve classification accuracy by fully leveraging multi-scale features, including phase information. We first designed a complex-valued three-dimensional network framework combining complex-valued 3D convolution (CV-3DConv) with complex-valued squeeze-and-excitation (CV-SE) modules. This framework is capable of simultaneously capturing spatial and polarimetric features, including both amplitude and phase information, from PolSAR images. Furthermore, to address robustness degradation from limited labeled samples, we introduced a multi-scale learning strategy that jointly models global and local features. Specifically, global features extract overall semantic information, while local features help the network capture region-specific semantics. This strategy enhances information utilization by integrating multi-scale receptive fields, complementing feature advantages. Extensive experiments on four benchmark datasets demonstrated that the proposed method outperforms various comparison methods, maintaining high classification accuracy across different sampling rates, thus validating its effectiveness and robustness.

The GRACE-FO satellite's laser ranging interferometer (LRI) measurements suffer from a significant number of phase anomalies, which can directly affect the gravity field recovery. To address this issue, we proposed an improved difference method (ImDFM) for phase anomaly processing, which enhances the traditional difference method (DFM) through two improvements: (1) combine the phase smoothing technique to handle cycle slips, single event upsets, and Mega phase jumps (PJs) which DFM hard to process; and (2) propose a novel time-tag synchronization offsets estimation method based on PJs to rectify the accuracy degradation caused by time-tag synchronization anomalies. Compared to the mainstream template method (TPM), ImDFM offers a significant advantage in conciseness. We evaluated the accuracy of ImDFM and residual spike characteristics by comparing the ImDFM product (S11) with the template-based product (S10) and the official JPL RL04 product in phase anomaly, LRI1B and gravity field results. The results show that (1) ImDFM can process phase anomalies with an accuracy better than 600 \boldmath $ \mathrm{pm/\sqrt{Hz}@1Hz}$ and the residual spikes mainly affect frequency band above 2 Hz; (2) the accuracy of estimated time-tag synchronization offset is better than 4 $\mu$ s, which can effectively correct the loss of accuracy; (3) the residual spikes in S11 are below 7nm/s, which does not affect the accuracy of gravity field recovery. The RL04 is still affected by phase anomalies, resulting in relatively higher noise above degree 30. In summary, the ImDFM proposed in this paper provides a high precision and concise approach for processing phase anomalies.