The classification of lithology through remote sensing technology has become increasingly vital for geological surveys and resource exploration. However, accurately classifying lithology in areas covered by vegetation remains a significant challenge. In this article, we leverage the robust penetration and coherence capabilities of ALOS-1 dual-polarization synthetic aperture radar (SAR) data to propose a new deep learning network named polarimetric alpha-guided network (PolAlpha-Net), guided by polarization decomposition knowledge. This network utilizes the polarimetric decomposition parameter to learn stable features, enabling high-precision lithology segmentation in vegetated areas. Our study utilized images from June, September, and November, all of which were heavily vegetated, and performed segmentation tasks across four categories, including a background category, using existing geologic maps as ground truth. The results demonstrate that PolAlpha-Net obtains superior performance, achieving a mean accuracy of 86.79%. Specifically, the accuracy for the four categories-background, Q(4), Arjnd, and gamma(2(1))(5)-were 77.32%, 94.19%, 85.93%, and 89.71%, respectively. These outcomes confirm that PolAlpha-Net is adept at learning stable features even under dense vegetation cover and possesses excellent generalization capabilities across different conditions. Consequently, lithologic mapping with PolAlpha-Net yielded results with impressive spatial continuity. The outcomes of this research confirm that integrating PolAlpha-Net with ALOS-1 dual-polarization SAR data produces high-quality lithological segmentation in vegetated areas, underscoring its practical application value in remote sensing and geological analysis.

Highlights What are the main findings? This paper proposes a real-time UAV positioning framework for weak GNSS environments, which establishes high-precision cross-modal alignment between drone and satellite imagery via a "Multimodal features + LightGlue" strategy. A synergistic point-line-plane feature fusion module is designed to provide robust geometric constraints, effectively mitigating the feature sparsity and matching failures common in low-texture and high-dynamic scenarios. What are the implications of the main findings? The shift from GNSS-dependency to a robust multimodal geometric fusion paradigm offers a mission-critical solution for UAV autonomy, ensuring high-precision positioning continuity and operational security in GNSS-denied environments. The systematic analysis of how altitude, scale, and seasonal dynamics influence performance reveals that future robust positioning architectures should prioritize adaptive feature selection to ensure operational stability under variable deployment conditions.Highlights What are the main findings? This paper proposes a real-time UAV positioning framework for weak GNSS environments, which establishes high-precision cross-modal alignment between drone and satellite imagery via a "Multimodal features + LightGlue" strategy. A synergistic point-line-plane feature fusion module is designed to provide robust geometric constraints, effectively mitigating the feature sparsity and matching failures common in low-texture and high-dynamic scenarios. What are the implications of the main findings? The shift from GNSS-dependency to a robust multimodal geometric fusion paradigm offers a mission-critical solution for UAV autonomy, ensuring high-precision positioning continuity and operational security in GNSS-denied environments. The systematic analysis of how altitude, scale, and seasonal dynamics influence performance reveals that future robust positioning architectures should prioritize adaptive feature selection to ensure operational stability under variable deployment conditions.Abstract To address the challenges of unmanned aerial vehicle (UAV) navigation in weak global navigation satellite system (GNSS) environments, this study proposes a novel multimodal feature fusion framework for real-time positioning using a priori high-resolution satellite imagery. This framework utilizes georeferenced satellite images as matching sources and employs a "Multimodal features + LightGlue" algorithm to achieve high-precision cross-modal matching. By combining point, line, and plane features for enhanced robustness in low-texture scenarios, the system further integrates LightGlue's lightweight confidence classifier to accelerate inference while maintaining high accuracy on challenging image pairs. Consequently, the proposed method outperforms LoFTR, RoMa, SuperPoint + SuperGlue, and SuperPoint + LightGlue in matching performance. Experimental results demonstrate that at a flight altitude of 80 m, the average real-time positioning error is 0.73 m, which increases to 6.24 m at 480 m. Factors such as ground object type, seasonal changes, flight altitude, and satellite image scale significantly influence accuracy. This research demonstrates that the visual navigation system meets practical operational needs for real-time UAV positioning in GNSS-deprived environments.

Progressive slope instability in open-pit mines poses a persistent hazard to mining operations, yet its reliable identification remains challenging due to fragmented deformation pat terns, frequent engineering disturbances, and terrain-controlled limitations in radar observations. To address this challenge, we propose an InSAR observability guided multi-modal fusion framework (IOGMF) for open-pit slope hazard detection. The proposed framework integrates InSAR deformation time series with optical imagery and terrain-derived information to improve the reliability of slope hazard identification. Deformation trends extracted from multi-temporal InSAR time series are used to characterize long-term surface deformation evolution, while optical and terrain information provide complementary structural context for deformation interpretation. By jointly leveraging deformation evidence and multi-modal structural information, the framework effectively differentiates hazard-related precursory signals from low-magnitude operational displacements. The proposed method is trained on a multi-source remote sensing dataset collected from 26 large and super-large open-pit iron mines across China, covering diverse geomorphological settings, geological backgrounds, and mining conditions. To rigorously evaluate generalization under limited labels, we adopt a cross-site independent experimental protocol, where the model is tested on previously unseen mines that are not included in the 26 core study sites used for model development. Under this design, IOGMF achieves strong hazard detection performance, reaching an F1 score of 0.809 (Precision 0.917, Recall 0.742) and a Hazard IoU of 0.663, with improvements of 0.057 in F1-score and 0.066 in Hazard IoU over the strongest competing baseline. The results suggest that the proposed framework can capture deformation related hazard patterns on the evaluated unseen mining sites while reducing non-hazard responses associated with engineering disturbances. Furthermore, when only 50% of the training data is used, the proposed method still achieves an F1-score of 0.776 and a Hazard IoU of 0.607, demonstrating robustness under moderate data scarcity. The results highlight the importance of incorporating InSAR observability and multi-modal contextual information for reliable interpretation of deformation signals in complex mining terrains and provide useful support for slope hazard assessment using remote sensing data.

The “Mother’s Day superstorm” of 10–12 May 2024 was one of the most intense space weather events of Solar Cycle 25. Driven by multiple X-class solar flares and successive coronal mass ejections (CMEs) from AR13664, it produced strong magnetosphere–ionosphere coupling and severe global ionospheric disturbances. This paper reviews recent observations and research findings on the major response features and underlying mechanisms of the global ionosphere during this event. Particular attention is given to the relative roles of prompt penetration electric fields (PPEFs), disturbance dynamo electric fields (DDEFs), thermospheric neutral winds, and thermospheric composition changes, especially variations in the O/N2 ratio, across different latitudes, longitude sectors, and storm phases. The results indicate that the event produced a strongly nonuniform ionospheric response across latitude, longitude, local time, and storm phase. At low and equatorial latitudes, the storm produced a strengthened equatorial ionization anomaly (EIA), poleward expansion of the anomaly crests, and a rapid increase in total electron content (TEC), all of which are typical features of a positive ionospheric storm. At middle and high latitudes, sustained electron density depletion and negative storm effects developed under the combined influence of polar energy input, thermospheric circulation reconfiguration, and composition disturbances. At the same time, multiscale disturbance structures intensified markedly, including large-scale wave-like disturbances such as LSTIDs, F-region plasma density structures such as equatorial plasma bubbles (EPBs), and smaller-scale irregularities responsible for GNSS scintillation. Mechanistically, storm-time electric fields enhanced upward plasma drift over the equator and triggered a super-fountain effect, which was a key driver of the low-latitude positive storm. In contrast, thermospheric heating, global circulation reorganization, and a reduced O/N2 ratio were the main factors controlling the formation and delayed recovery of the negative storm at middle and high latitudes.

Road extraction from remote sensing imagery is crucial for transportation planning and operational management in open-pit mining scenarios. However, complex mining environments, like elongated road geometries, cloud and haze interference, and spectral confusion caused by residual ores, limit the performance of existing approaches. This study proposes RoadMamba-Net, a novel deep learning framework integrating Spatial-Aware Mamba (SA-Mamba), dynamic deformable strip convolution (DDSC), and graph-based topology refinement for robust road extraction in open-pit mines. The model adopts an encoder–decoder architecture, where DDSC enhances the capture of multi-orientation elongated road features, and SA-Mamba captures long-range spatial dependencies. Furthermore, a gated parallel fusion strategy is designed to effectively integrate SA-Mamba with the Convolutional Block Attention Module (CBAM), thereby improving feature representation capability. Additionally, a road topology optimization (RTO) module is introduced to refine structural continuity. Experiments on a self-constructed open-pit mine dataset demonstrate competitive overall performance, achieving an IoU of 73.55%, a Dice coefficient of 83.91%, and an APLS score of 83.18%, effectively addressing topological disconnections in challenging scenarios. Zero-shot cross-dataset validation on DeepGlobe-18 and an independent cross-mine evaluation on the Waitoushan iron mine further confirm its stable cross-domain adaptability. The proposed approach provides an effective solution for automated road mapping, supporting digital mining and intelligent mine management.

Accurate segmentation of open-pit mine road networks presents a critical challenge for mine digitization and autonomous driving applications. These roads are prone to mechanical compaction, geological erosion, and coverage by gravel dust, resulting in segmentation outcomes characterized by blurred boundaries, holes, fractures, and geometric deformations, which severely compromise measurement accuracy. To address these challenges, this paper proposes the Mining Road Segmentation Network (MRS-Net), which integrates local features with global semantics. First, a Residual Network Version 2 (ResNetV2)-Transformer cascaded encoder is constructed, employing residual connections to preserve sub-pixel-level edge details and multi-head self-attention to establish long-range dependencies, thereby enhancing the representation of weak texture features. Second, the Road Multi-scale Features Fusion Module (RMFF) was designed to extract local geometric features and global continuity features through progressive hollow convolution, enabling the model to extract multi-scale features and effectively suppress interference from gravel dust. Finally, a progressive decoding architecture incorporating bilinear interpolation is adopted to improve edge smoothness. MRS-Net is evaluated on an Unmanned Aerial Vehicle (UAV)-acquired road dataset from the Anshan open-pit iron mine in Liaoning Province, China. Results demonstrate that MRS-Net achieves superior segmentation performance compared to models such as DeepLabV3+ and TransUNet across three distinct scenarios: main roads, temporary roads, and abandoned roads. Specifically, it achieves Intersection over Union (IoU), Dice coefficient(Dice), and Kappa coefficient (Kappa) values of 89.4 % / 94.1 % / 87.2 %, 75.7 % / 83.3 % / 75.1 %, and 83.8 % / 90.0 % / 84.85 % respectively for these scenarios.

Ground-based synthetic aperture radar (GB-SAR) is widely used in several monitoring fields for its advantages of high deformation sensitivity. However, the limitations of its two-dimensional sector imaging mode make it difficult to accurately analyze and decipher its high-precision deformation results in a three-dimensional (3D) form, which has caused troubles in locating hazardous areas of open-pit mine slopes and disaster warning. For this purpose, this paper uses the laser point cloud as auxiliary data and performs coordinate definition and coordinate conversion on GB-SAR images, on the basis of which an image geocoding method that takes into account the original 3D point cloud matching is proposed. Firstly, the original 3D point cloud coordinate matching was performed based on the sector mesh. Then, for the pixels not matched to the point cloud, their planar coordinates x and y were reconstructed using the center of gravity weighted algorithm with the pixel center coordinates as the reference, and further, the elevation h was reconstructed based on the planar coordinate information using a radial basis function neural network. Finally, the accuracy of the pixels’ 3D coordinate reconstruction was quantitatively evaluated in the absence of the matched point cloud. The application of GB-SAR landslide monitoring in the Nanfen open-pit mine in Liaoning, China, verifies the reliability of this paper’s method, and its distance alignment root mean square error of 9.89 cm. This study can provide technical support for the interpretation of hazardous areas in open-pit mines and early warning of landslide disasters.

The accurate measurement of total electron content (TEC) is vital for ionospheric research and satellite navigation services. Nowadays, the advancement of multiple Global Navigation Satellite System (multi-GNSS) has resulted in over 100 seamlessly spaced satellites and thousands of publicly available GNSS stations globally, which enables precise estimation of global ionospheric TEC and differential code biases (DCBs). However, challenges persist in eliminating system differences and leveraging the advantages of multi-GNSS big data due to frequency variances and hardware delays. Here, we propose a novel estimation method that incorporates satellites, constellations, stations, and time constraints for joint global ionospheric TEC and DCB estimation to improve the consistency and reliability of the usage of global observation data from multi-GNSS systems. The results show that the global vertical TEC map derived from approximately 5000 global multi-GNSS sites significantly reduces discrepancies between different satellites, constellations, and receivers, enhancing temporal and spatial consistency, with a standard deviation improvement of over 20%. The long-term stability of satellite DCBs shows tiny fluctuations for the Global Positioning System (GPS), GALILEO, and BeiDou Navigation Satellite System (BDS) systems within +/- 0.5 ns, with slightly larger fluctuations for the GLONASS system within +/- 0.6 ns. The long-term stability of receiver DCBs shows that the standard deviation in mid- and high-latitude regions ranges from 0.1 to 0.2 ns, while in low-latitude regions ranging from 0.3 to 0.6 ns. The proposed method leverages multi-GNSS big data and pronouncedly improves the accuracy of global ionospheric TEC and DCB estimation, which provides a valuable tool for high-precision global ionosphere monitoring and related applications.

The gold mining resource holds significant economic and financial value, providing precipus metal resources for the country, driving economic growth, and enhancing currency stability and hedging capabilities in the international financial market. However, while precise, the current chemical analysis methods for measuring gold ore grades in mmes face issues such as long processing times, high costs, and reagent pollution, hindering the automation of ore grade and beneficiation method adjustments based on real-time grade information. In contrast, due to its efficiency, eco-friendliness, and in sin measurement advantag visible near infrared spectroscopy is gradually becoming an effective alternative for estimating melal grades in mining areas. First, the raw spectral data were processed using Savitzky-Golay (SG) smoothing to reduce noise, and the spectral characteristics of gold ores were analyzed. It was found that reflectance correlates with gold grade, and a gold absorption feature is present at 455 nm, Based on this finding, dimensionality reduction was performed on the raw spectral data using principal component analysis (PCA), isometric feature mapping (ISOMAP), and locally linear embedding (LLE), with the resulting dimensions reduced to 6. 5. and 5. respectively. Finally, prediction models for gold grade were established using random forest (RF), extremely randomized trees (ET), decision trees (DT), gradient boosting decision tree GBDT), adaptive boosting (Adaboost), extreme gradient boosting (XGBoost), and stacking ensemble learning algorithms of the dimensionally reduced data, Results indicated that the Stacking ensemble learning method outperformed single models in all aspects, Among them, the LLE-Stacking combined model achieved the highest accuracy, withR(2) of 0. 972. RPD of 5.935, and an average relative error of 0.231 between predicted and actual values, The method proposed in this study allows for rapid and accurate predictions of gold content in ore, significantly improving the inversion accuracy compared to traditional models, providing new technological means for the rapid and in-situ measurement of gold grades in mines, and holding great significance for efficient gold extraction.

Global climate change is leading to more severe rainfalls, greatly exacerbating the inundation risk of large-scale subsidence funnels worldwide. Here, we present a case of flood disaster in a large ground subsidence funnel at the Liaohe plain of China in 2022, triggered by continuous heavy rainfall under the ongoing global climate change. We conducted a retrospective investigation and analysis of the disaster event and its impacts through the comprehensive utilization of satellite remote sensing, numerical simulation, and risk assessment techniques. We observed the scale of the subsidence funnel for nearly 50 years, quantitatively analyzed the impact of subsidence on the flood disaster, and assessed the flood disaster risk. The results indicated that the ground subsidence has partially compromised the flood protection capacity of rivers, elevating flood risk in this region by 28.24%. Additionally, the subsidence has also damaged the flood control dam, leading to a breach and a secondary flooding disaster. Our work demonstrates that it's possible to characterize the inundation risk of large-scale subsidence funnels by spaceborne remote sensing in advance. This approach could enhance the capacity to cope with extreme weather events in similar ground subsidence regions worldwide.

The lunar shadow during a total solar eclipse significantly reduces energy input into the upper atmosphere, inducing notable changes in the Earth's ionosphere. This study presents two novel methods for non-uniform and non-linear ionospheric background corrections to precisely detect total electron content (TEC) changes using dense multi-GNSS observations. Using the total solar eclipse over North America on April 8, 2024, as a case study, we first propose a non-uniform background correction technique to accurately capture ionospheric eclipse-induced TEC depletions. Applying the non-uniform correction shows an adjustment of about 23

We statistically study 10 weak to moderate geomagnetic storms that occurred since the 25th solar activity cycle to evaluate the quality of satellite signal and positioning accuracy of Beidou Navigation Satellite System (BDS). The results show that the PAS (Percentage of Affected Satellites) index, which integrates three scenarios: satellite signal loss, satellite signal loss of lock, and carrier phase signal anomaly, appears significant anomaly changes during weak and moderate geomagnetic storms, and its changing trend is closely related to the geomagnetic storm Kp index and Dst index, the correlation coefficient is greater than 0.75. Based on the results of PAS signal quality evaluation, the effect of PAS on the positioning error of Precision Point Positioning (PPP) of BDS-2/BDS-3 systems is analyzed and verified in conjunction with PAS. The results show that there is a strong agreement between the time of PAS anomaly and the time of PPP positioning accuracy anomaly. The positioning error of PPP increases by 9 to 44 times compared to the quiet period. The statistical results indicate that the PAS index is useful for the signal quality evaluation of the Beidou Navigation Satellite System under different space weather conditions.

Referring to the current use of sparse representation algorithms to extract camouflaged targets from hyperspectral Images, the selection of the background dictionary is affected by the "same spectrum of different objects" of the hidden targets. resulting in the inability to detect camouflaged targets accurately. In this paper, we take the grassland camouflage net and desert tamouflage net as the research objects, collect the visible near infrared reflectance spectra of the camouflage net and airborne Ayperspectral images respectively, and analyze the spectral characteristics of the background pixels and camouflage target pixels In the camouflage net and airborne images measured outdoors. Taking advantage of the fact that the spectra of the camouflage pets measured outdoors and the background in the airborne images are different and the possibility of neighboring image elements belonging to the same feature is high, the background dictionary selection method based on the constraints of Euclidean distances And image homogeneity features is proposed, and the sparse representation of the background dictionary is further utilized to Identify the target, The results show that (1) in the wavelength range of 750 similar to 1.000 nm for grass camouflage, the reflectance of the background pixel spectrum in the image is higher than that of the camouflage net spectrum, For desert camouflage, in the tange of 550 similar to 700 nm, the reflectance of the background pixel spectrum in the image is higher than that of the camouflage net Apectrum, (2) By establishing spatial and spectral feature constraints with the maximum spectral Euclidean distance to the target mage element and the highest homogeneity with neighboring image elements, 413 background image elements in the grass amouflage image and 507 background image elements in the desert camouflage image were selected as the background Mictionary. (3) Based on the improved background dictionary selection method, the sparse representation algorithm is utilized to Hentify the camouflage targets, and the results can accurately discriminate the location and number of camouflage targets. The area under the curve (AUC) of the receiver operating characteristics for the detection of grass camouflage targets and desert amouflage targets reaches 0.96 and 0.98, respectively, indicating that the algorithm has good detection performance for both rass camouflage targets and desert camouflage targets.

Spaceborne interferometric synthetic aperture radar (InSAR) techniques are important for landslide detection and monitoring; however, several limitations and uncertainties, such as the unique north–south flying direction and side-look radar observing geometry, currently limit the ability of InSAR to credibly detect landslides, especially those related to high and steep slopes. Here, we conducted experimental and statistical analysis on the feasibility of time-series InSAR monitoring for steep slopes using ascending and descending SAR images. First, the theoretical (TGNSS), practical (PGNSS), and terrain (Hterrain) (T-P-H) indices for sensitivity evaluations of the slope displacement monitoring results from time-series InSAR were proposed for slope monitoring. Subsequently, two experimental and statistical studies were conducted for the cases with and without Global Navigation Satellite System (GNSS) monitoring data. Our experimental results of two high and steep open-pit mines showed that the defined theoretical and practical sensitivity indices can quantitatively evaluate the feasibility of ascending and descending InSAR observations in steep-slope deformation monitoring with GNSS data, and the terrain sensitivity index can qualitatively evaluate the feasibility of landslide monitoring results from ascending and descending Sentinel-1 satellite data without GNSS data. We further demonstrate the generalizability of these proposed indices using four landslide cases with both public GNSS and InSAR monitoring data and 119 landslide cases with only InSAR monitoring data. The statistical results indicated that greater indices correlated with higher reliability of the monitoring results, suggesting that these novel indices have wide suitability and applicability. This study can help to improve the practice of slope deformation monitoring using spaceborne InSAR, especially for high and steep slopes.

Mainly aiming at the research on the method of measuring the multi-source parameters of blasting boreholes in open pits, using the multi-source parameters of the measuring device to calculate the minimum resistance line distance. Through the intelligent analysis of the minimum resistance line distance, the blasting effect is improved, and the secondary blasting caused by the increase of the block rate can avoid the phenomenon of sudden increase in blasting cost and waste of resources caused by the increase of the mass rate. It ensures the safe production of the mine and the safety of the lives and property of the surrounding people.

At present, the traditional method of obtaining the minimum resistance line in mines is still to use the latitude and longitude instrument to draw the corresponding section map where the borehole is located and then use the circle solution method to find out the shortest distance from the borehole axis to the trapezoidal slope. This method is difficult to map in the open pit mine, with low accuracy and low automation level. An intelligent analysis method of minimum resistance line distance based on a 3D point cloud model is proposed. According to the structural characteristics of the open pit mine, the raster data is established by the point cloud data, and the free surface and non-free surface are automatically separated by the standard deviation of the height difference, and then the minimum resistance line is extracted. The engineering application shows that the method provides efficient data support for the 3D mining software to carry out mining design, block delineation, planning and blasting design of open pit mines, and significantly improves the efficiency and management level of mine blasting.

In this paper, we propose a new method to quantitatively evaluate the quality of the carrier phase observation signals of the BeiDou Navigation Satellite System (BDS) during weak and moderate geomagnetic storms. We take a moderate geomagnetic storm that occurred on 12 May 2021 during the 25th solar cycle as an example. The results show that the newly defined PAS (Percentage of Affected Satellites) index shows significant anomaly changes during the moderate geomagnetic storm. Its variation trend has good correlations with the geomagnetic storm Kp index and Dst index. The anomaly stations are mainly distributed in the equatorial region and auroral region in the northern and southern hemispheres. The proposed PAS index has a good indication for both BDS2 and BDS3 satellites. We further validated this index by calculating the Precise Point Position (PPP) positioning error. We found that the anomaly period of PAS has strong consistency with the abnormal period of PPP positioning accuracy. This study could provide methodological support for the evaluation of the signal quality and analysis of positioning accuracy for the BeiDou satellite navigation system under different space weather conditions.

Ground-based synthetic aperture radar (GB-SAR) interferograms contain a large amount of phase noise. The existing methods lose effective deformation information to varying degrees upon filtering, which seriously reduces the monitoring accuracy of GB-SAR. In this paper, a GB-SAR interferogram filtering method for open-pit mines is studied. First, permanent scatterer points (PSPs) are extracted through coarse filtering, and the distribution of the noise phase is analysed. Then, an improved Grubbs outlier discrimination criterion is introduced considering the uncertainty of the PSP distribution. A method based on the eight-neighbourhood outlier discrimination (EOD) criterion is proposed to identify noise points. High-coherence points (HCPs) are used to reconstruct the phase of the noise points to filter the GB-SAR interferogram. Finally, the filtering effect is quantitatively evaluated using the peak signal-to-noise ratio (PSNR) and structural similarity (SSIM). The superiority of the proposed filtering method in the image displacement region is verified based on simulated and real deformation data. The results show that the proposed method can accurately filter out the phase noise of interferograms and improve the accuracy of GB-SAR for slope monitoring.

根据矿区滑坡前地表变形的遥感监测方法,提出了不同轨道SAR(synthetic aperture radar)数据集监测结果可靠性的判别依据,并对诱发矿区滑坡的因素进行分析.以鞍山市鞍千哑巴岭露天采场边坡为研究对象,基于44景Sentinel-1雷达影像(两组升轨和一组降轨),利用时间序列InSAR方法分析了采场边坡在2019年11月25日滑坡前约6个月的地表运动特征.结果表明,不同轨道数据集针对同一研究区域所获取的监测结果具有差异性,滑坡区域顶部后缘位置在发生滑坡前的一段时间范围内(约45 d)呈异常强烈的加速变形现象.研究成果将为今后利用InSAR技术早识别矿区易滑坡危险区提供新的思路.