Search

Objectives: Convolutional neural network (CNN) was widely used in landslide susceptibility assessment because of its powerful feature extraction capability. However, with the demand for scene diversification and high accuracy, the algorithm of CNN was constantly improved. The practice of improving accuracy by deepening the network hierarchy or combining other models tended to increase the number of model parameters and the amount of calculation greatly, which led to the difficulty of model training or the overfitting of results, thus limiting its practical application. We want to solve the above problems from the model structure. of feature maps, one-dimensional CNN（1D-CNN） and two-dimensional CNN（2D-CNN） were connected to maintain network depth, limit model parameters, and reduce computation. The parameters sharing of multidimensional convolution kernels was used to capture the deep coupling features of different dimensions and different landslide factors, so as to make full use of features and avoid overfitting. Taking Sedongpu gully in Xizang,China as the experimental area, this paper selected 11 kinds of landslide influencing factors to analyze the landslide susceptibility. Results: The results show that the multi-dimensional CNN coupling structure had the same training efficiency as the shallow 2D-CNN with fewer parameters due to the reduced computational effort. While compared with the deep 2D-CNN with the approximate number of parameters, the training time of the proposed method was significantly reduced and the training cost was lower. In addition, the coupled model had a stronger feature learning ability than the independent 1D-CNN and 2D-CNN, and therefore obtained higher model accuracy. Under each confusion matrix metric of the testing data, the coupled model received higher scores, and thus obtained more reliable landslide susceptibility assessment results. Conclusions: The multidimensional CNN coupling model proposed in this paper is a reliable method applicable to landslide susceptibility assessment. This study provides new theoretical guidance and technical support for further landslide hazard monitoring and prevention [ABSTRACT FROM AUTHOR]

Objectives: In 2015, the Mw 7.8 earthquake in Nepal caused landslides and seriously impacted Nepal and its surrounding areas. Methods: In this paper, the Sindhupalchowk area, which is close to the epicenter of the earthquake, is selected as the study area. First, based on the L-band ALOS-2 and C-band Sentinel-1A synthetic aperture radar data, the detection and identification of post-earthquake landslides are carried out by using stacking-InSAR(interferometric synthetic aperture radar) technology. Second, the two-dimensional time series deformation characteristics of a typical landslide are obtained by combining the ascending and descending orbit data and multi-dimensional small baseline subset-InSAR technology. Results: Combined with optical images, 14 hidden points of landslide are delineated. The main deformation of the typical landslide occurs in the horizontal east-west direction, and the maximum deformation rate is −69 mm/a. Conclusions: By analyzing the trend and periodic term deformation signals in time series of the typical landslide, it is found that the earthquake has a significant acceleration effect on the landslide movement, and the increase of rainfall makes the landslide displacement show a periodic change from August to November every year. This paper can provide a reference for post-earthquake landslide monitoring re⁃ search. [ABSTRACT FROM AUTHOR]

Copyright of Geomatics & Information Science of Wuhan University is the property of Geomatics & Information Science of Wuhan University and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Objectives：In mid- to late- April 2024, an extreme heavy rainfall event occurred in Shaoguan City, Guangdong Province, inducing a large number of landslides in Jiangwan Town, Shaoguan. People lost connection with the outside world for nearly 36 hours, which aroused widespread social concern. Rapidly and accurately identifying the basic characteristics of landslides, development and distribution patterns and formation conditions is crucial for disaster emergency decision-making and risk elimination and disposal. Methods: Using the post-disaster optical remote sensing images and combining with deep learning model, the rainfall-induced landslides in Jiangwan Town, Shaoguan, were quickly and automatically identified. Results:After manually calibration, a total of 1 192 landslides were deciphered, with a total area of about 3.14 km². The scale of the landslides was dominated by small and medium-sized landslides, which were mainly distributed as an aggregated belt along the river in the northeast-southwest direction, with a significant characteristic of concentrated occurrence. Spatial statistical analysis showed that the landslides were mainly distributed on concave slopes with slopes of 10°-30° in the range of 200-300 m elevation. Further quantitative analysis of the geomorphic controlling factors of landslides using the random forest model and SHAP theory reveals that different topographic and geomorphic factors have different degrees of nonlinear effects on landslide formation, and that multiple factors such as elevation, slope, and catchment conditions are coupled to jointly control the formation of landslides. Conclusions: This paper highlights the great advantage of deep learning-based intelligent identification and analysis technology in the emergency investigation and formation conditions analysis of landslide disasters, which can provide important technical support for the rapid assessment of disaster losses and risk identification. [ABSTRACT FROM AUTHOR]

Objectives: The landslides disaster is one of the most important factors for the construction of major infrastructure in the western China. How to effectively and reliably carry out wide-area landslide susceptibility prediction has always been a frontier difficulty in domestic and foreign studies. However, the present data-driven methods for prediction of landslide susceptibility in large-scale and complex scenarios still face two major issues: (1) The difference between various landslide inducing environments in a wide range of scenarios, would cause the difficulty to applying a single model to account for multiple landslide phenomenon; (2) small sample problem: Complex environmental tasks require models with large capacity and strong representative power, but there is a lack of sufficient landslide samples in practice. Methods: In response to the above problems, this paper takes Qijiang and Fuling District of Chongqing City, China as an example, proposes a local prediction strategy, and introduces the idea of meta-training an intermediate representation suited to be generalized, that can be adapted to the landslide sensitivity prediction task corresponding to the current local area, with only a very small number of samples in the local area, and number of iterations. Thus, the two issues mentioned can be well settled. Results: The proposed method is different from traditional methods such as support vector machines, multilayer perceptrons, and random forests, which require a large number of samples and gradient iterations to train the supervised model. Instead, only a small sample is required to fine-tune the intermediate model, which still improves the global accuracy by 1%-5%, the precision by 1%-3%, the F1-score by 0.5%-6%, and the recall rate is close to the highest level of other methods. Conclusions: The meta-learning paradigm enables few-shot adaptation of landslide susceptibility prediction model with superior statistical performance. [ABSTRACT FROM AUTHOR]

Objectives: Global navigation satellite system (GNSS) technology is widely used in deformation monitoring such as dams, bridges, and landslides. In traditional monitoring methods, the reference stations need to be built, which leads to the increased monitoring cost. Methods: This paper proposes a GNSS deformation monitoring method based on virtual reference station (VRS), which uses the existing continuously operating reference stations (CORS) observation data to generate VRS instead of physical reference station to achieve high accuracy deformation monitoring. Firstly, the existing CORS observation data are used to obtain the atmospheric error correction models and generate VRS. Secondly, the baselines between VRS and each monitoring point are calculated. In order to further improve the accuracy of data processing, Delaunay triangulation networks are formed, and the baseline vector solution results are checked according to the criterion that the triangle closure error is zero. Finally, the checked baseline solution results are adjusted to obtain the final three-dimensional position coordinates of the monitoring points. Results: This paper uses about 30 monitoring points in Pingbian Miao Autonomous County, Yunnan Province for experimental analysis. The results show that the interpolation accuracis of the double-difference tropospheric and ionospheric delay are better than 10 mm if the distance between CORS reference stations is less than 40 km. The observation data longer than 8 hours can realize deformation monitoring with horizontal accuracy better than 5 mm and vertical accuracy better than 10 mm. Conclusions: The proposed method can realize the practical application of landslide monitoring. In the future, we can study how to improve the atmospheric error model and further improve the accuracy of data processing [ABSTRACT FROM AUTHOR]

The recognition of landslide structure and deformation behavior and the change with time are crucial in landslide hazard monitoring. Seismological methods have unique advantages in obtaining the internal structure of the Earth and the fault rupture and deformation. Passive seismic methods do not need artificial sources, and the field work is flexible. Passive seismic methods are safe, economical and effective to monitor landslide. Passive seismic methods have been used to obtain velocity structure, thickness and brittle de⁃ formation characteristics of the Xishancun landslide in Lixian County, Sichuan Province. This paper briefly describes the basic theory of these seismological methods and the progress in the study of the Xishancun landslide. [ABSTRACT FROM AUTHOR]

This study aims to explore building education capacity to improve disaster management in public awareness. We can improve public awareness of disaster risks through public education on landslide disasters. The current education in disaster courses is mainly aimed at students and ambulance personnel. Those education methods focus on textbook teaching and professional training, and the contents are too professional. There is no effective communication channel between educators and the public. Thus, this paper proposes a visual representation method of landslide disasters for public education. We prepared a concept map to organize and represent bits of knowledge of landslide, instantiating nodes of concept map as animation, then used animation to publicize. The landslide on June 24, 2017 in Diexi, Maoxian was selected to conduct a case test analysis. One short animation created by the method and one brief disaster report summarized by that animation is produced for a questionnaire. We collected and analyzed the results of the questionnaire. The results show that people tend to learn about landslides by watching animation, and they answer more questions correctly in the questionnaire after watching animation than reading the report. So the proposed method in this paper can effectively construct the visual content of landslide disaster knowledge and quickly help the public understand the landslide disaster knowledge. [ABSTRACT FROM AUTHOR]

Objectives: The Sichuan Tibet Railway crosses the Jinsha River specifically the section from Xiongsong Town to Shadong Town, Gongjue County, Tibet Autonomous Region, in which exist strong topography variations and many potential geohazards, and hence it urgent to detect the potential geohazards in this region to ensure the railway safety. Methods: In this paper, 61 Sentinel -1 ascending, 53 Sentinel -1 descending and 7 advanced land observing satellite 2 (ALOS 2) ascending images are used to derive the annual mean deformation rates in the satellite radar line of sight (LOS) with two advanced InSAR (interferometric synthetic aperture radar) approaches, namely generic atmospheric correction online service for InSAR (GACOS) assisted InSAR Stacking and LiCSBAS. The three LOS annual mean deformation rate maps are then used to determine 2D surface movements (one along the slope and the other perpendicular to the slope) in the study region. Results: The comparison between GACOS assisted InSAR Stacking and LiCSBAS results show that they agree with each other with correlation coefficients over 0.92 for the LOS deformation rates and over 0.85 for the 2D deformation rates, suggesting the reliability of GACOS assisted InSAR Stacking. The maximum annual deformation rate of 163 mm/a can be observed in the slope direction and seven landslides (A, B, C, D, E, F, G) can be clearly identified, which in turn lays the foundation for future real-time monitoring. Based on the detailed analysis of the seven regions by using InSAR and optical interpretation results, it is found that the seven landslides had relatively obvious deformation. Landslides A, B, C and D are active landslides, which might cause the result of river blocking due to overall instability. At present, landslides E, F and G are still in the stage of slow deformation. Conclusions: This study also find that GACOS assisted InSAR Stacking can effectively remove long band and topographic related atmospheric delay error with the help of GACOS, which has the advantages of being simple, effective, fast, and easy to popularize and apply. GACOS assisted InSAR Stacking technology can be used to quickly identify potential landslide hazards. [ABSTRACT FROM AUTHOR]

Objectives: Monitoring and early warning is an important means of geohazard prevention and mitigation. In the real ‑ time monitoring and early warning of landslide, the scheduling algorithm of early warning model is directly related to the success of early warning. The traditional way is to start the early warning service regularly, which is easy to realize and is also a common practice in the industry. However, due to the influence of the time interval, there is a certain delay in the issuance of the early warning conclusion, which cannot achieve real‑time early warning. How to get early warning level quickly through early warning model and real‑time monitoring data, and how to control the time and frequency of sending early warning information are two key problems to be solved. Methods: In order to shorten the time interval from data collection to early warning information sending, and win more time for the follow ‑ up early warning processing, this paper studies the early warning process and model scheduling algorithm. Based on the timedriven and data‑driven scheduling mode of early warning model, combined with the types and characteristics of real‑time monitoring data, the scheduling strategy of early‑warning model and the release strategy of early warning information are proposed to improve the warning accuracy of landslide, and a set of early warning level solver is developed. Results: For the early warning model scheduling of displacement monitoring data, the time‑driven mode can be used in the initial stage of landslide deformation, and the data‑driven mode can be used after the accelerated deformation stage. This can greatly reduce the number of early warning calculation, and ensure that the key deformation information can be captured in time. For the state monitoring data, the data‑driven mode can be used for early warning to ensure the timeliness of early warning. For rainfall data, time ‑ driven mode has a good effect. Here, the time interval can be determined by combining the data characteristics. Conclusions: The multi thread technology is used to construct the general computing framework of the early warning model based on strategy, to realize the real‑time process tracking and early warning of landslide, and give full play to the role of landslide early warning platform. and the general computing framework of early warning model based on strategy is constructed by using multi ‑ threading technology. [ABSTRACT FROM AUTHOR]

Objective: Large-scale human engineering activities induce and aggravate the landslide disaster, which seriously threatens the engineering safety and environmental safety. Landslide susceptibility assessment is the key technique of landslide monitoring and early warning. In view of the limited data sources of traditional landslide monitoring Methods: and the lack of effective Methods: for mining the spatial distribution of landslide hazards and their inducing factors, the main Objective: of this study was to investigate the prediction performances of synthetic minority oversampling technique (SMOTE) and convolutional neural networks (CNN) for landslide susceptibility assessment in the Wanzhou area of the Three Gorges, China. Methods:: The landslide susceptibility assessment Methods: used in this paper are SMOTE and CNN. SMOTE is an improved random oversampling algorithm, which was used to better the imbalance of input samples. CNN is a widely used deep learning network and differs from other neural networks mainly in convolution operation. CNN was used to analyze the nonlinear relationship between landslides and its influencing factors. Firstly, 22 influencing factors were extracted from multi-source data such as topographic, geological, and remote sensing data, and multicollinearity test was performed on these factors. Then, the grid unit with a spatial resolution of 25 m was selected as the mapping unit. 30% of the initial sample dataset was selected randomly for testing of the CNN, and the remaining 70% was expanded by using SMOTE to obtain 1:1 landslide and non-landslide dataset. 90% of the proportionally balanced dataset was randomly selected as the training set and the remaining 10% as the validation set. Thirdly, the obtained training sam - ple set is input to the CNN and six optimal model parameters were set to ensure the stability and efficiency of the model. Finally, the trained CNN was used to generate landslide susceptibility zoning map, and the receiver operating characteristic curves, testing and validating set were used for evaluating the accuracy and Results of the proposed model.Results The proposed SMOTE-CNN model with the highest predictive accuracy and generalization ability was trained and then used to calculate the landslide susceptibility indices. The index values of the grid cells vary from zero to one and correspond to landslide susceptibilities from low to high. The landslide susceptibility maps were created using these values. For easy visual interpretation, the susceptibility values were classified into five classes (very low,low,moderate,high,and very high) using Jenks natural breaks. Approximately, 59.57 % of the landslides lie in the very-high-susceptibility region, 18.74% of the landslides lie in the high-susceptibility region, 10.73% of the landslides lie in the moderatesusceptibility region, and 10.96 % of the landslides lie in the low- and very-low-susceptibility regions. The experimental Results demonstrate that the proposed model provides the best predictive accuracy. The model can effectively assess landslide susceptibility and provides a novel method for landslide prediction.Conclusions: The SMOTE and CNN were applied to quantitatively mapping the landslide susceptibility in Wanzhou area of the Three Gorges, China. The experimental Results: show that SMOTE and CNN have good stability, and the Results were consistent with field investigations and can provide a reference for landslide prevention and reduction in the Three Gorge, China. In addition, CNN has the characteristics of strong adaptability and is good at mining local features of data and extracting global training features and classification. However, the physical meaning of CNN model is not clear, and it takes a long time to process a large amount of data, so it relies on GPU acceleration. Therefore, it is necessary to carry out further research from the aspects of super-parameter optimization combination, model structure adjustment, applicability analysis etc., so as to improve the accuracy and robustness of the model. [ABSTRACT FROM AUTHOR]

Landslide evolution is a long‐term and complex process. Geo‐hazards have different characteristics in various data, and have certain applicability in different stages of landslide. This paper applies different characteristics and applicability of various data sources in landslide by the multi‐source data fusion method. The deformation and failure characteristics and temporal and spatial evolution law of Huangnibazi landslide during its dynamic evolution from pre‐sliding to mid‐sliding to post‐sliding are studied. The results show that the deformation and failure process of Huangnibazi landslide can be divided into four stages: startup stage, accelerated deformation (accelerated sliding) stage, front expansion (decelerated sliding) stage and gradual stabilization stage. Huangnibazi landslide is a creep ‐ pull ‐ split landslide formed under the influence of self‐weight, rainfall infiltration, earthquake and vibration caused by human engineering activities. The application of various data in different stages of landslide is summarized: interferometric synthetic aperture radar(InSAR) technology, optical image and topographic data are used to determine potential landslide body before sliding. Optical remote sensing image is used to observe the overall deformation and trend of landslide accumulation body during sliding. Global positioning system (GPS) is used to continuously observe local deformation after sliding. Field investigation is used to determine the characteristics of geological hazard body. [ABSTRACT FROM AUTHOR]

The pre-failure evolution of landslides is of great value for the analysis of landslide triggering factors and post-disaster stability assessment. At present, optical images are commonly employed to ana⁃ lyze the pre-failure evolution, but it is well-known that their data availability could be highly limited due to the presence of clouds. With the advance in radar remote sensing and interferometric synthetic aperture radar(InSAR), it could provide a new technical approach for landslide pre-failure detection. In this paper, the 2018 Nanyu landslide in Gansu Province is utilized to demonstrate the capability of InSAR to trace its pre-failure surface displacements using European Space Agency’s Sentinel-1 radar data with a temporal interval of 12 days in different tracks. The results show that the landslide began to deform in June of 2017, and the maximum cumulative deformation reached up to 77 mm in the 13 months before the occurrence of the landslide. The time series InSAR derive displacements and the rainfall data is consistent, suggesting that the rainfall should be one of the triggering factors for the landslide. The study demonstrates the potential of radar interferometry for landslide detection, which can provide insights on landslide triggering factors, landslide disaster prevention and mitigation, and even landslide monitoring and early warning work in the future. [ABSTRACT FROM AUTHOR]

In China，traditional methodology on early detection of natural terrain to landslides is challenging as zones most prone to slope failure are usually inaccessible due to high location and dense vegetation. This can lead to underestimation of potential landslide events to the degree of wrongly identifying unstable areas as stable. This paper provides a solution for these cases by proposing an integrated space~air-ground investigation system that allows for the early detection, real-time prediction, and warning of catastrophic geohazards. Firstly, high-resolution optical images and interferometric synthetic aperture radar (InSAR) data from satellites are employed to obtain a global panorama of a region，highlighting these problematic locations； yet results are detailed enough to provide reliable estimates of deformations at particular points along time spans of days and weeks. As consequence，it makes the compilation of long displacement time-histories feasible，contributing to the understanding of long-term landslide-driving phenomena in regions where it has been underestimated. This is called the general investigation. Then，detailed assessments can be done through the development of unmanned aerial vehicles (UAV) for elaborating high-resolution relief maps and photogrammetric representations based on both visual images and light laser detection and ranging (LiDAR) data. The system finally allows for precise tagging of locations that warrant real-time site monitoring of displacements using global navigation satellite system (GNSS) and crack gauges，validating expecting behavior of these critical, but previously hidden hazardous locations. The overall approach makes it possible to establish a four-level comprehensive early warning system, which meets the urgent needs of the country and promotes a practical and operational application of such system in the field of geohazard prevention [ABSTRACT FROM AUTHOR]

In order to promote the wider use of ground-based synthetic aperture radar (GB-SAR) in the field of surface micro-deformation monitoring, this paper summarizes the main types of GB-SAR and introduces the working principles and important parameters of some representative systems. By taking the S-SAR (slope-synthetic aperture radar) ground-based radar independently developed by the China Academy of Safety and Production Science as an example，the key processing techniques utilized by the GB-SAR are introduced. Experiments verify the accuracy of GB-SAR. It also shows the application of GB-SAR in open pit mine slope monitoring，mine dumping site and landslide emergency rescue. Finally，the problems and suggestions of GB-SAR in monitoring are proposed. [ABSTRACT FROM AUTHOR]

The landslide is one of the most common geological hazards in nature, the rapid development of unmanned aerial vehicle (UAV) technology and virtual reality (VR) technology provides important data resources and technical support for immersive simulation and visual analysis of the landslide. Firstly, this paper focuses on the dynamic construction and exploratory analysis of VR scene in landslide hazard，then some key technologies such as diversified organization of landslide data， dynamic fusion visualization of VR scene are addressed in detail，thus a VR scene interaction method based on handle ray is proposed. Finally > a prototype system is implemented and a case study region is selected for experimental analysis. The experimental results show that the proposed method can dynamically construct landslide VR scene based on UAV remote sensing data，and can support the user to carry out immersive interaction and disaster information analysis. [ABSTRACT FROM AUTHOR]

Based on the “Gaofen +，，，this paper explores the emergency analysis of the Jinsha River landslide disasters on October 11 and November 3，2018，with the “Gaofen + geological disaster" model，and uses the high-resolution remote sensing image of the disaster area before and after the disaster to interpret the landslide. According to the basic disaster information of the disaster, the characteristics of pregnancy disaster and deformation and creep before and after the disaster are analyzed. The hidden dangers are investigated in the Baige Lake and surrounding areas after the disaster. The suspected cracks and secondary landslide hazard points on both sides of the barrier lake are identified，and road accessibility analysis of disaster areas after the second landslide is carried out. The results show that the “ Gaofen + " model has great potential for the application of natural disasters such as the Jinsha River landslide，disaster monitoring and emergency rescue roads. [ABSTRACT FROM AUTHOR]

Satellite radar observations enable us not only to detect landslides with detailed sliding signals over broad spatial extents, but also to track landslide dynamics continuously，which has gradually been recognized by the earth observation and landslide communities. However, there are still several challenges in the landslide detection and monitoring with satellite radar observations due to their inherent limitations such as the phase decorrelation caused by heavy vegetation and/or large gradient surface movements，and the geometric distortion introduced by the side-looking orbit. In this paper，from landslide detection and monitoring perspective，the four major challenges of satellite radar technologies are discussed：① The phase decorrelation caused by heavy vegetation can be weakened by use of synthetic aperture radar (SAR) imagery with a long radar wavelength (e. g. S-band or L-band)，a short temporal resolution，and/or a high spatial resolution (e. g. 1 m or even higher)，and/or advanced interferometric SAR (InSAR) time series，and the phase decorrelation associated with large deformation gradients can be addressed by SAR offset tracking and range splitspectrum interferometry techniques.② Atmospheric effects represent a big challenge of conventional InSAR for landslide detection and monitoring, especially in mountain areas. The generic atmospheric correction online service (GACOS) which is developed at Newcastle University can be used to reduce atmospheric effects on radar observations and simplify the follow-on time series analysis.③ The geometric distortions such as shadows and layovers can be pre-analyzed using an external digital elevation model (DEM) for medium-spatial-resolution SAR data； in contrast，for high-resolution SAR data，a machine learning approach can be used to identify water bodies，shadow and layover areas without a requirement of a high-spatial-resolution DEM.④ Residual topographic phase exhibits in areas with high buildings or steep slopes，which could easily lead to phase unwrapping errors; this can be tackled by a baseline linear combination approach. In addition, a framework is proposed to combine satellite radar technologies with other earth observations (e. g. ground-based radar, LiDAR and GNSS) to develop an automated landslide detection and monitoring system. It is expected that this paper will help the earth observation and landslide communities clarify the technical pros and cons of the satellite radar technologies so as to promote them and guide their future development. [ABSTRACT FROM AUTHOR]

Landslide susceptibility prediction is the key technology in landslide monitoring, early warning, and assessment. The core problem in quantitative prediction of landslide hazards is the effective selection of conditioning factors and prediction models. In this paper, the Three Gorges Reservoir area was selected as a case study to predict landslide susceptibility. First, key landslide-related factors were selected as input variables using topographic, geological, and remote sensing data. Secondly, according to the nonlinear and uncertainty characteristics of landslides, a PSO-SVM model was trained and used to assess landslide susceptibility. Finally, the prediction results of grid-and object-based prediction models were validated by comparing them with known landslides using the classification accuracy. The results show that object-based PSO-SVM possesses high prediction accuracy with the area under curve of 0.841 5 and a Kappa coefficient of 0.849 0. These experimental results are consistent with field investigations and can provide a reference for landslide prevention and reduction in the Three Gorges, China. [ABSTRACT FROM AUTHOR]