Real-world environment-derived point clouds invariably exhibit noise across varying modalities and intensities. Hence, point cloud denoising (PCD) is essential as a preprocessing step to improve downstream task performance. Deep learning (DL)-based PCD models, known for their strong representation capabilities and flexible architectures, have surpassed traditional methods in denoising performance. To our best knowledge, despite recent advances in performance, no comprehensive survey systematically summarizes the developments of DL-based PCD. To fill the gap, this paper seeks to identify key challenges in DL-based PCD, summarizes the main contributions of existing methods, and proposes a taxonomy tailored to denoising tasks. To achieve this goal, we formulate PCD as a two-step process: outlier removal and surface noise restoration, encompassing most scenarios and requirements of PCD. Additionally, we compare methods in terms of similarities, differences, and respective advantages. Finally, we discuss research limitations and future directions, offering insights for further advancements in PCD.

Impulse ultrawideband (UWB) synthetic aperture radar (SAR) combines high-azimuth-range resolution with robust penetration capabilities, making it ideal for applications such as through-wall detection and subsurface imaging. In such systems, the performance of UWB antennas is critical for transmitting high-power, large-bandwidth impulse signals. However, two primary factors degrade radar imaging quality: (1) inherent limitations in antenna radiation efficiency, which lead to low-frequency signal loss and subsequent time-domain ringing artifacts; (2) impedance mismatch at the antenna terminals, causing standing wave reflections that exacerbate the ringing phenomenon. This study systematically analyzes the mechanisms of ringing generation, including its physical origins and mathematical modeling in SAR systems. Building on this analysis, we propose a Bayesian ringing suppression algorithm based on sparse optimization. The method effectively enhances imaging quality while balancing the trade-off between ringing suppression and image fidelity. Validation through numerical simulations and experimental measurements demonstrates significant suppression of time-domain ringing and improved target clarity. The proposed approach holds critical importance for advancing impulse UWB SAR systems, particularly in scenarios requiring high-resolution imaging.

The introduction of tomographic synthetic aperture radar (TomoSAR) technology into the field of through-wall perception enables high-resolution 3D imaging of behind-wall targets, enhancing reconnaissance capabilities for applications such as urban counter-terrorism and disaster response. Unlike airborne tomographic SAR systems, vehicle-mounted through-wall tomographic SAR (TW-TomoSAR) operates at shorter ranges and involves propagation through layered media, which induces wavefront curvature errors (WCE) and medium path errors (MPE). These errors manifest as target defocusing and positional inaccuracies in reconstructed images. This paper proposes an efficient processing architecture for TW-TomoSAR. Following 2D SAR imaging, wavenumber-domain MPE compensation is performed using phase correction functions derived from equivalent transmission models. Subsequently, WCE correction is accomplished using keystone transformation in the complex image domain, allowing the height reconstruction process to be separated, simplifying the 3D imaging workflow. Finally, Bayesian super-resolution reconstruction incorporating Laplace priors is adopted to overcome antenna size limitations in elevation resolution. The practical implementation of the proposed TW-TomoSAR demonstrates high-resolution imaging of targets behind walls.

With the growing demand for High-Resolution and Wide-Swath (HRWS) imaging in Synthetic Aperture Radar (SAR) missions, digital beamforming (DBF) in the range dimension has become a key enabling technology. However, under wide-swath imaging conditions, existing DBF algorithms—such as SCORE, single/multi-delay DBF, multi-frequency DBF, and scaling-function DBF—often suffer from performance degradation near the swath edges, including target displacement, mainlobe broadening, and signal-to-noise ratio (SNR) reduction, which decrease imaging quality. This paper proposes an enhanced DBF algorithm based on multi-delay and scaling compensation. The proposed method introduces multi-delay correction, scaling compensation, and a data fusion strategy to effectively suppress mainlobe broadening while enhancing image SNR, all with a relatively low computational complexity of O(N×Nr×log(N_r)). Extensive validation using both satellite simulation data and airborne measured data demonstrates the superiority of the proposed method: compared to existing algorithms, it achieves up to 6 dB SNR improvement and 0.4098 m mainlobe suppression in the near range, and up to 2 dB SNR improvement and 0.0819 m suppression in the far range. This method offers an important technical reference for efficient beam shaping and robust imaging in future wide-swath SAR systems.

The measurement of the target azimuth angle using forward-looking radar (FLR) is widely applied in unmanned systems, such as obstacle avoidance and tracking applications. This paper proposes a semi-supervised support vector regression (SVR) method to solve the problem of small sample learning of the target angle with FLR. This method utilizes function approximation to solve the problem of estimating the target angle. First, SVR is used to construct the function mapping relationship between the echo and the target angle in beamspace. Next, by adding manifold constraints to the loss function, supervised learning is extended to semi-supervised learning, aiming to improve the small sample adaptation ability. This framework supports updating the angle estimating function with continuously increasing unlabeled samples during the FLR scanning process. The numerical simulation results show that the new technology has better performance than model-based methods and fully supervised methods, especially under limited conditions such as signal-to-noise ratio and number of training samples.

Three-dimensional synthetic aperture radar (SAR) point clouds provide valuable geometric observations of urban scenes, but they often suffer from severe noise and layer-like artifacts caused by the low signal-to-noise ratio and tomographic imaging mechanism. These degradations make SAR point cloud denoising significantly more challenging than conventional LiDAR point cloud denoising. In this paper, we propose a Geometry-guided Diffusion SAR Point Cloud Denoising (GDSD) framework to recover geometrically coherent building surfaces from noisy SAR point clouds.The key idea is to exploit relatively clean LiDAR point clouds as geometry priors while avoiding the need for paired SAR–LiDAR supervision or clean SAR ground truth. Specifically, we introduce a Forward Gaussian Noising Process to disrupt the intrinsic layer-like artifacts of SAR point clouds and reduce the input-level discrepancy between SAR and LiDAR domains. We further design a geometry prototype-based alignment module that projects SAR and LiDAR bottleneck features into a shared LiDAR-dominated latent space, enabling geometry-aware conditional reverse diffusion. A DiT-3D-based denoising network is then trained with LiDAR-domain diffusion supervision and applied to SAR point clouds using the aligned SAR geometry condition. To evaluate the proposed method, we construct a SAR point cloud denoising benchmark based on the MV3DSAR dataset with CAD-derived reference surfaces. Experimental results show that GDSD significantly improves the quality of noisy SAR point clouds and clearly outperforms the previous conventional LiDAR point cloud denoising baseline, producing more continuous and geometrically coherent SAR building point clouds.