Ground-based radar technology has substantially alleviated the limitations of conventional measurement techniques in landslide monitoring. Nevertheless, the absence of reliable methodologies for deformation extraction and early warning has impeded the further advancement and practical deployment of this technology. In this study, radar deformation images were employed to extract the primary landslide body through the method combining Otsu and region growing (Otsu-RGM), thereby establishing a foundation for subsequent deformation data extraction and early warning applications. Furthermore, by extracting multidimensional deformation data from the landslide body, we developed a comprehensive standardized regional deformation (CSRD) and a comprehensive standardized regional deformation index (CSRDI) early warning model, which were validated using case studies of mining landslides. The results show that the Otsu-RGM method has certain stability in extracting the deformation of ground-based radar images. The deformation data of the landslide surface, as monitored by ground-based radar, reveal distinct statistical characteristics. The CSRD curve effectively captures the temporal evolution of landslide movement and demonstrates strong sensitivity during the acceleration stage. The CSRDI early warning model can effectively warn of impending landslides. The Otsu-RGM method and CSRDI model provide novel insights into the application of ground-based radar in landslide monitoring.

Fractional vegetation cover (FVC) is a key indicator for assessing vegetation growth and ecological conditions in grasslands. Its estimation holds significant value in smart pastoral management and global ecological research. The radar vegetation index (RVI), obtained from polarimetric synthetic aperture radar (PolSAR) data, has been widely used to characterize the vegetation distribution and growth conditions. Conventional approaches are primarily designed for dense vegetation types such as forests and croplands, where RVI is typically constructed by statistically analyzing the polarimetric characteristics within a local window. However, such approaches face limitations in sparse areas, particularly in grasslands, where the vegetation cover is not only highly sparse but also displays considerable heterogeneity. Moreover, due to the weaker scattering signals from the short canopy layer, traditional RVI exhibits relatively low sensitivity in grasslands. To tackle such issues, an enhanced RVI (ERVI) targeting the monitoring of FVC was proposed in this study. The proposed ERVI is developed within pixels that share similar polarimetric signatures, rather than spatially neighboring pixels, effectively characterizing the sparse and heterogeneous nature of the grassland regions. Furthermore, the volume scattering component is introduced to improve the responsiveness of ERVI to variations in grassland FVC. Two Gaofen-3 PolSAR datasets obtained in typical grassland areas in northern China are adopted to examine the FVC estimation ability of the proposed ERVI. The results show that the correlation between ERVI and reference FVC increased by 0.22 and 0.20 compared to the traditional RVI. Meanwhile, in FVC estimation, the coefficients of determination (R2) increased from 0.43 to 0.78 and from 0.40 to 0.69, respectively, demonstrating the effectiveness and applicability of ERVI for grassland FVC monitoring.

The speckle noise limits the interpretation of polarimetric synthetic aperture radar (PolSAR) data. The non-local means (NLM) algorithms rely on the self-similarity of the noisy SAR data itself to calculate matching weights. However, speckle noise often leads to inaccurate pixel matching, failing to balance noise reduction and structural preservation. To address these limitations, this paper proposes a structure-guided non-local means (SG-NLM) algorithm, which utilizes cross-modal optical guidance. The proposed approach introduces rigorously co-registered optical images as a speckle-noise-free semantic reference, providing structural constraints in the filter weight calculation. A semantically consistent guidance map is generated via super-pixel segmentation and multi-dimensional feature clustering to provide reliable spatial constraints. By integrating semantic boundary constraints into the NLM framework, polarimetric statistical similarity calculations are strictly confined within homogeneous areas. This mechanism prevents cross-boundary smoothing and preserves ground features. Both qualitative and quantitative evaluations using Radarsat-2 data from the Xilin Gol grassland, Inner Mongolia, demonstrate that SG-NLM significantly outperforms existing methods. Specifically, the edge preservation index (EPI) in natural regions improved by 11.39%, while the measured polarization fidelity increased by up to 45.54%. These results confirm that the proposed method effectively suppresses speckle while preserving the key physical scattering mechanisms.

Aiming at the issue that traditional slope landslide early warning methods for mining areas fail to effectively consider the influence of terrain characteristics on landslide processes, this study proposes an adaptive early warning threshold method based on dynamic time decay weight (DTDW). First, before processing radar image data, an automatic outlier detection and correction algorithm is applied to reduce data anomalies and enhance monitoring data quality. Next, the processed time-series radar deformation images are analyzed to extract the maximum deformation area as the study region, and multiple monitoring parameters (including velocity, acceleration, and coherence) within this area are calculated. Subsequently, the dynamic time decay weight (DTDW) is introduced, which assigns varying degrees of decay weights to deformation image parameters from historical moments, thereby enabling continuous influence on the parameter estimation at the current moment. Finally, a three-level warning mechanism is established, along with an adaptive warning threshold adjustment method based on a sliding window analysis strategy. This method dynamically monitors parameter trends and optimizes warning thresholds in real time. By continuously updating and calibrating warning criteria for different parameters, the system’s responsiveness to abnormal deformations and its adaptability to environmental changes are improved. Experimental results based on real mining area landslide data show that, compared with traditional fixed-threshold warning methods, the proposed approach can identify landslide signs earlier and issue warning signals more promptly. It demonstrates significant advantages in both timeliness and accuracy, effectively enhancing the monitoring capability for landslide risks.

To address low vegetation height inversion, this paper proposes an adaptive polarimetric decomposition (APD) algorithm based on the degree of anisotropy (A). The algorithm builds an adaptive volume scattering model, which adjusts A to simulate the scattering characteristics of low vegetation canopies (such as rice and crops) composed of randomly oriented disk-like particles. Validation is conducted using PolInSAR data of 1-meter-high rice from the LAMP microwave anechoic chamber under zero temporal baseline and controlled spatial baseline conditions. Experimental results show that the algorithm achieves a stable root mean square error (RMSE) of 0.1815–0.2874 m at full incidence angles, and significantly improves the reliability of phase extraction in low vegetation areas under optimal incidence angle conditions.

Grassland aboveground biomass (gAGB) is an important indicator for assessing ecosystem services. Traditional ground plot harvesting is accurate but costly and spatially sparse, limiting its ability to represent accuracy and heterogeneity. Existing models seldom consider scale dependence or the effects of management and disturbance. This study examined enclosed and grazed grasslands in Xilingol and Hulunbuir, Inner Mongolia. Centimetric Unmanned Aerial Vehicle (UAV) Light Detection and Ranging (LiDAR) data were used to derive the canopy height model (CHM), canopy gap fraction (CGF), and leaf area index (LAI) at spatial resolutions from 0.05 m to 0.25 m. Spatial heterogeneity was quantified using geostatistical semi-variograms (as nugget, sill, range) based on canopy structure parameter and linked to the accuracy of gAGB models analysis. Three modelling approaches were evaluated: linear regression (LR), random forest (RF), and backpropagation neural network (BPNN) in gAGB estimation. The compared results show that grazing disturbance produced a more homogenized canopy structure across all spatial resolutions. Compared with CGF and LAI, CHM shown lower heterogeneity and higher stability, and its inclusion contributes more to enhancing inversion accuracy. RF model achieved higher and more stable accuracy than LR and BPNN. The results also showed that finer resolutions (0.05 m) captured micro-scale variation more effectively and were suitable for high-precision monitoring. The proposed workflow provides spatially continuous, centimetric structural information and applies heterogeneity diagnostics to guide the selection of suitable gAGB measuring predictors and spatial scales. Based on the grazing sample plots and dataset analysed, this approach significantly improved gAGB modelling accuracy and offers a practical basis for fine-scale biomass monitoring and grassland management.

Spaceborne DBF-SAR systems achieve highresolution wide-swath (HRWS) imaging through multi-channel technology, but Direction of Arrival (DOA) deviations caused by undulating terrain lead to severe coherent processing gain loss. To address this issue, this paper first establishes an observation geometric model under undulating terrain and quantitatively analyzes the impact of pointing deviations on system gain. Subsequently, an adaptive DBF method based on prior DEM assistance and a dual-beam flip iteration mechanism is proposed. This method utilizes rough DEM information for beam initialization and, combined with an adaptive step-size search strategy, effectively balances tracking speed and convergence precision. Finally, simulation experiments of point targets and complex terrain scenarios verify the effectiveness of the proposed method. The results demonstrate that, compared to traditional spectral estimation methods, this algorithm has lower computational complexity and higher spaceborne real-time processing efficiency under typical terrain errors; meanwhile, the method also exhibits excellent noise robustness in low Signal-to-Noise Ratio (SNR) environments.

Highlights What are the main findings? The presence of space-variant baseline errors affects the deformation retrieval results of ground-based synthetic aperture radar. Space-variant baseline errors not only compromise the accuracy of individual deformation measurements but also accumulate over time, leading to larger errors. What are the implications of the main findings? Accuracy of deformation inversion can be further improved by using Taylor expansion and Singular Value Decomposition to estimate and compensate for space-variant baseline errors. Applying this method enables the timely identification of the actual deformation area and expands the scope of landslide monitoring.Highlights What are the main findings? The presence of space-variant baseline errors affects the deformation retrieval results of ground-based synthetic aperture radar. Space-variant baseline errors not only compromise the accuracy of individual deformation measurements but also accumulate over time, leading to larger errors. What are the implications of the main findings? Accuracy of deformation inversion can be further improved by using Taylor expansion and Singular Value Decomposition to estimate and compensate for space-variant baseline errors. Applying this method enables the timely identification of the actual deformation area and expands the scope of landslide monitoring.Abstract Leveraging differential interferometric techniques, ground-based synthetic aperture radar (GB-SAR) delivers highly accurate displacement measurements, typically reaching submillimeter scales. However, in practical engineering, minor platform instability induced by environmental factors gives rise to space-variant baseline errors, which affects the deformation value. In response to this issue, this paper presents a method combining Taylor expansion and singular value decomposition for estimation and compensation of the space-variant baseline error. Initially, the Gaussian Mixture Model (GMM) is employed to adaptively select high-quality Permanent Scatterers (PSs) to facilitate robust data provision for the following error parameter estimation. Subsequently, a three-dimensional multi-parameter model for the space-variant baseline error is established via Taylor expansion, followed by parameter estimation using Singular Value Decomposition (SVD). Experiments indicate that the proposed approach effectively mitigates the error phase arising from platform vibration, thereby enhancing the precision of GB-SAR deformation inversion.

Phase unwrapping in interferometric synthetic aperture radar (InSAR) aims to recover a continuous phase field from wrapped observations, which enable accurate topographic reconstruction and surface deformation measurements. With the recent advances in deep learning (DL), several DL-based unwrapping approaches have shown promising performance. However, deep learning networks suffer from inconsistent feature representations between encoder and decoder stages. This leads to incompatible skip connections that provide limited benefits and even degrade reconstruction quality. To overcome this limitation, we propose ResUCTransNet that integrates residual learning with transformer-based feature modeling. The network employs a multi-scale residual backbone derived from Res_UNet to extract stable deep features. Then, to replace conventional skip connections, a channel transformer (CTrans) module is introduced that composed of channel-wise cross fusion transformer (CCT) and channel-wise cross attention (CCA). This design effectively reduces the semantic gap in different network stages, which allows adaptive integration of local CNN features and global transformer representations. Experiments on the public InSAR-DLPU dataset demonstrate that ResUCTransNet effectively reduces model complexity and achieves substantial improvements over existing deep learning models and classical unwrapping algorithms. Specifically, the proposed method attains the best performance in terms of RMSE and SSIM (RMSE = 1.6247, SSIM = 0.7741). Compared with the second-best model, Res_Unet (RMSE = 2.8409, SSIM = 0.7733), ResUCTransNet achieves an approximately 42.8% reduction in RMSE while maintaining nearly identical structural similarity. The proposed method provides higher reconstruction accuracy and better structural fidelity, while maintaining strong robustness and generalization in complex terrain or severe noise conditions.

Ground-based radar (GBR) delivers rich multi-component observations, including spatially continuous surface-deformation fields and coherence maps. Yet conventional early-warning models rely heavily on single-point displacement time series, underutilizing GBR’s inherent spatial information and proving inadequate for complex landslides deviating from classic creep behavior. To address this, we propose an improved landslide early-warning and forecasting method centered on a novel Extent of Channel Deformation (EOCD) index. First, we formalize “channel-deformation data” — conceptualizing the four-dimensional spatiotemporal evolution of a slope’s deformation zone as an integrated data channel capturing macroscopic intensity and scale. Second, EOCD fuses three complementary parameters: cumulative channel deformation, deformation-zone area, and mean coherence — with coherence inverted quadratically to enhance physical sensitivity to pre-failure acceleration, per the theoretical coherence–deformation-velocity relationship. Third, the EOCD tangent angle enables fine-grained identification of acceleration stages; terminal-stage channel-deformation velocities, combined with the inverse-velocity method, yield precise failure-time forecasts. Validated using data from two western China open-pit mines with contrasting geology and deformation modes, results show: for a classic creep landslide, EOCD achieves a 6-minute prediction error (ahead), outperforming the improved-tangent-angle method (21-minute lag); for a composite landslide where single-point analysis fails entirely, EOCD delivers effective warning 36 min prior, with only 4-minute error (behind). By fully leveraging GBR’s area-monitoring capability, EOCD overcomes single-point limitations under atypical deformation, offering a more robust framework for integrated multi-component radar-data interpretation.

Abstract. The inverse velocity method (INV) based on the onset of acceleration (OOA) point is widely used in landslide time prediction. However, the selection of OOA point affects the accuracy of INV prediction results. This study proposed a deformation standard deviation-OOA (DSD-OOA) point identification method based on the statistical characteristics of ground-based radar landslide area deformation data. By introducing a controllable variable, a modified INV method was derived. The OOA point identified by DSD-OOA was substituted into the modified INV method for landslide time prediction and compared with the prediction results of the moving average-OOA (MA-OOA) point method. Results show that compared to MA-OOA, the inverse velocity time series after the OOA point identified by DSD-OOA exhibits smaller fluctuations and is closer to linear change. The INV predictions using MA-OOA (MA-OOA-INV) consistently lag behind the actual landslide time, while the INV predictions using DSD-OOA (DSD-OOA-INV) are more stable and consistently precede the actual landslide time of failure. Furthermore, the root mean square error (RMSE) and coefficient of determination (R²) indicate that the DSD-OOA-INV method predicts landslide lifetime with higher accuracy, suggesting that OOA points identified by the DSD-OOA method can more precisely predict landslide time of failure.

Conventional adaptive beamforming algorithms often suffer from significant performance degradation when steering vector mismatches and covariance matrix estimation errors occur. To address this problem, this paper proposes an adaptive robust beamforming algorithm based on an improved generalized linear combination (GLC) framework. The proposed method first applies singular spectrum analysis to the received signals to suppress noise components. A diagonal loading coefficient function related to the received signal snapshots is then constructed, and a generalized diagonally loaded covariance matrix is formed using the denoised data. Finally, by exploiting spatial integration and subspace projection within a predefined angular uncertainty set, the actual direction of arrival of the desired signal is accurately estimated, and the steering vector is corrected accordingly. Simulation results demonstrate that, compared with traditional SMI, LSMI, GLC and improved GLC algorithms, the proposed method achieves a 3-5 dB higher output signal-to-interference-plus-noise ratio (SINR) across the entire input signal-to-noise ratio (SNR) range under steering vector mismatch, and reaches an output SINR close to the optimal level with only 100 snapshots, exhibiting excellent robustness against steering vector mismatch and limited snapshot conditions.

The critical sliding tangential angle (the tangential angle of the displacement-time curve prior to landslide failure) is a crucial indicator for predicting landslide instability. However, existing early warning methods struggle to accurately predict its threshold. Based on a modified Nishihara constitutive model (a rheological model for geomaterial deformation), this study established a quantitative relationship between the critical sliding tangential angle and the average velocity during the constant deformation stage of landslides, proposing a new early warning method. Through statistical analysis of 25 typical landslide cases, a critical sliding tangential angle warning model was developed. The model can predict the critical sliding tangential angle when landslides enter the constant deformation stage. To validate its effectiveness, two open-pit mine landslide cases were selected for verification. Results indicate that the relative errors between the model's predicted and calculated values are all less than 1 %, demonstrating high prediction accuracy. The model can predict the threshold of the critical sliding tangential angle before landslide occurrence, providing sufficient response time for early warning personnel. This method enables dynamic adjustment of landslide warning thresholds, significantly improving accuracy. The findings provide a new technical approach for early identification and warning of landslide hazards.

The conventional Capon beamforming algorithm can achieve a high gain in the direction of desired signals and zero-trapping in the direction of interfering signals, providing a high output signal-to-interference-plus-noise ratio (SINR). However, when the steering vector of the desired signal is mismatched, the performance of the Capon beamforming algorithm degrades. In addressing this challenge, the present research introduces a refined algorithm. The core of the proposed robust Capon beamforming technique lies in leveraging the orthogonality between the steering vector and the noise space, the estimated expected signal steering vector is corrected. Based on this feature, the proposed algorithm meticulously optimizes the predicted steering vector of the desired signal, which can mitigate the problem of performance degradation caused by the mismatch in the steering vector. Moreover, the covariance matrix is corrected using the desired signal elimination method, which can overcome the problem of signal self-cancelation. Furthermore, through the optimization process, the proposed algorithm can maintain high robustness in complex environments and under the condition of different input signals, its beam pattern performance is more excellent. The results of simulation experiments show that the proposed algorithm demonstrates greater robustness compared to the currently available algorithms, can achieve a higher output SINR, and is insensitive to steering vector mismatch.

The inability of traditional offset tracking to adapt the feature window to changes in surface feature density limits its effectiveness in monitoring complex surface deformations. This study addresses this issue by proposing a local feature extraction method based on different coordinate grid areas and combining multiscale analysis and multiangle dynamic thresholds. The study uses principal component analysis to select the main variation components for local feature values and dynamically adjust the initial feature window method in conjunction with the variation coefficient. Finally, the study separately employs the structural time series model and spatial autocorrelation to evaluate and compare the performance of the dynamic window and the fixed window, using the offset results to analyze potential causes of landslides. The research indicates that the dynamic window size can effectively capture the complex characteristics of areas with severe deformation and ensure data integrity in landslide edge areas.

Space–time adaptive processing (STAP) based on sparse recovery (SR-STAP) has demonstrated remarkable clutter suppression performance under insufficient sample conditions. However, the main aim of sparse recovery is to solve the norm minimization problem. To this end, this study proposes a weighted STAP algorithm based on a greedy block coordinate descent method to address the problems of slow convergence speed and insufficient estimation accuracy in the existing l2,1-norm minimization methods. First, the weights are estimated using the multiple signal classification (MUSIC) algorithm. Then, a greedy block selection rule that favors sparsity is used, prioritizing the update of the weighted block that has the greatest impact on sparsity. Although the proposed algorithm in this paper is greedy in nature, it is globally convergent. Finally, the accuracy of clutter covariance matrix estimation and the convergence speed of the SR-STAP algorithm are enhanced by reasonably estimating the noise power and selecting appropriate regularization parameters. The results of simulation experiments indicate that the proposed algorithm can effectively suppress clutter ridge expansion, achieving excellent clutter suppression and target detection performance compared with the existing methods, as well as satisfactory convergence properties.

Ground-based synthetic aperture radar (GBSAR) has been widely used in the fields of early warning of geologic hazards and deformation monitoring of engineering structures due to its characteristics of high spatial resolution, zero spatial baseline, and short revisit period. However, in the continuous monitoring process of GBSAR, due to the sudden failure of radar equipment, such as power failure, or the influence of alternating work between multiple regions, it often leads to discontinuous data collection, and this problem caused by missing data is collectively called “inspection mode”. The problem of missing data in the inspection mode not only destroys the spatial and temporal continuity of the data but also affects the accuracy of the subsequent deformation analysis. In order to solve this problem, in this paper, we propose a data reconstruction method that combines Sage–Husa Kalman adaptive filtering and the Rauch–Tung–Striebel (RTS) smoothing algorithm. The method is based on the principle of Kalman filtering and solves the problem of “model mismatch” caused by the fixed noise statistics of traditional Kalman filtering by dynamically adjusting the noise covariance to adapt to the non-stationary characteristics of the observed data. Subsequently, the Rauch–Tung–Striebel (RTS) smoothing algorithm is used to process the preliminary filtering results to eliminate the cumulative error during the period of missing data and recover the complete and smooth deformation time series. The experimental and simulation results show that this method successfully restores the spatial and temporal continuity of the inspection data, thus improving the overall accuracy and stability of deformation monitoring.

A Sliding-Spotlight SAR provides high azimuth resolution but suffers from severe efficiency loss when the target scene is obliquely oriented or curved with respect to the satellite ground track. Such geometric mismatch restricts the effective swath and often requires multiple repeated passes for full coverage. To address this limitation, this paper proposes a spaceborne sliding-spotlight imaging architecture featuring two-dimensional beam steering, in which the antenna beam is jointly controlled in both azimuth and range to align the radar footprint with arbitrarily oriented scenes in real time. A multi-segment adaptive steering strategy is further introduced to support smoothly varying scene trajectories while maintaining Doppler continuity and stable imaging geometry. Simulation results demonstrate that the proposed approach achieves orientation-adaptive beam tracking, ensures consistent resolution across segments, and significantly improves coverage efficiency compared with conventional sliding-spotlight operations.

Previous polarimetric target decomposition methods have been widely used in forests and buildings and have achieved good results. However, there is a lack of corresponding modeling for the study of grassland scattering characteristics. To improve the accuracy of grassland target decomposition, this article fully considers the characteristics of grassland vegetation and abstracts the grassland vegetation canopy into a shape-adaptive random quasi-ellipsoid particle cloud. The random state of grassland vegetation is simulated from the three dimensions of spin angle, tilt angle, and rotation angle, and the shape of grassland vegetation is adaptively simulated by the anisotropy degree A to generalize the applicability of the model so that the particle cloud is more in line with the overall state of grassland vegetation. To better deal with unnatural distributed scatterers, that is, considering the correlation between copolarization and cross-polarization, a helix scattering mechanism is introduced. The derived volume scattering model is constructed with the helix scattering model and the Freeman two-component ground scattering model to form a novel three-component scattering model suitable for grasslands, and a polarimetric target decomposition algorithm is proposed based on this model. Experimental comparisons with several other excellent polarimetric target decomposition algorithms are carried out on X-band Cosmos-SkyMed, C-band GF-3, and L-band AfriSAR fully polarimetric data. The experimental results show that this article’s algorithm performs better than other algorithms for decomposition effect on different data, and is more reflective of the scattering characteristics of different grassland vegetation regions.

Sliding spotlight SAR suffers from fixed angular velocity, making it difficult to achieve adaptive, high-resolution imaging in complex terrain. This paper introduces an adaptive azimuth beam-steering method for non-uniform scenes, where adjustable parameters a1 and a2 enable dynamic control of angular velocity and variable-resolution imaging. The geometric configuration and Doppler behavior of the sliding spotlight mode are analyzed to clarify how the proposed steering law affects resolution. Simulation experiments with point targets verify that the proposed method flexibly adjusts azimuth resolution according to terrain requirements, improving imaging adaptability and overall performance.