Landslide identification using machine learning

Landslide identification is critical for risk assessment and mitigation.This paper proposes a novel machinelearning and deep-learning method to identify natural-terrain landslides using integrated geodatabases.First,landslide-related data are compiled,including topographic data,geological data and rainfall-related data.Then,three integrated geodatabases are established;namely,Recent Landslide Database(Rec LD),Relict Landslide Database(Rel LD)and Joint Landslide Database(JLD).After that,five machine learning and deep learning algorithms,including logistic regression(LR),support vector machine(SVM),random forest(RF),boosting methods and convolutional neural network(CNN),are utilized and evaluated on each database.A case study in Lantau,Hong Kong,is conducted to demonstrate the application of the proposed method.From the results of the case study,CNN achieves an identification accuracy of 92.5%on Rec LD,and outperforms other algorithms due to its strengths in feature extraction and multi dimensional data processing.Boosting methods come second in terms of accuracy,followed by RF,LR and SVM.By using machine learning and deep learning techniques,the proposed landslide identification method shows outstanding robustness and great potential in tackling the landslide identification problem.     

基于多期DEM和滑坡强度的滑坡风险评估

滑坡危险性定量评估是滑坡风险评估中的关键和难点,也是当前国际风险管理研究中的热点问题.以滑坡密集分布的黑方台南塬为研究区,以32处典型滑坡为研究对象,依据多期三维数字高程模型(DEM),提出了一种基于强度的滑坡危险性定量评估技术方法.根据多期三维地形信息的解译及野外调查,编制多期滑坡分布图,计算滑坡活动的频率.利用GIS技术,利用滑坡体积与速度的乘积计算滑坡强度.将滑坡危险性定义为滑坡频率和滑坡强度的乘积,同时调查和分析了黑方台地区各类承灾体的类型、价值及其在相应滑坡强度下的易损性,在此基础上开展了单体滑坡风险评估和黑方台南塬滑坡风险区划.     

Flood hazard assessment and mapping of River Swat using HEC-RAS 2D model and high-resolution 12-m TanDEM-X DEM (WorldDEM)

Natural Hazards - Floods are among the most devastating and recurring natural hazards and have caused extensive economic losses to human lives and infrastructures around the world. Swat valley in...     

平缓红砂岩滑坡特征及警示分析

通过调查研究朝阳坡平缓泥质粉砂岩滑坡,对其滑坡产生的地质条件、软弱夹层与斜坡和冲沟的关系、软弱夹层的强度指标、裂缝的发展等进行了分析。根据对红砂岩滑坡规律的认识,及时在滑坡发展阶段撤离该地区居民,避免了生命财产的重大损失。平缓红砂岩滑坡的内在因素是软弱夹层及其临空,外在因素是强降雨使其抗剪强度降低。通过实例加深了对滑坡机制的认识,为从事类似滑坡调查和相关工程的科研设计、施工、监理人员提供了参考。     

Correction to: Flood hazard assessment and mapping of River Swat using HEC-RAS 2D model and high-resolution 12-m TanDEM-X DEM (WorldDEM)

Natural Hazards - The original article was updated and corrected due to numbering errors in Figure 8’s subfigures and the placement of some of the article’s other figures. Additional...     

Three-dimensional DEM investigation of critical state and dilatancy behaviors of granular materials

The critical state is significant to the mechanical behaviors of granular materials and the foundation of the constitutive relations. Using the discrete element method (DEM), the mechanical behaviors of granular materials can be investigated on both the macroscopic and microscopic levels. A series of DEM simulations under true triaxial conditions have been performed to explore the critical state and dilatancy behavior of granular materials, which show the qualitatively similar macroscopic responses as the experimental results. The critical void ratio and stress ratio under different stress paths are presented. A unique critical state line (CSL) is shown to indicate that the intermediate principal stress ratio does not influence the CSL. Within the framework of the unique critical state, the stress–dilatancy relation of DEM simulations is found to fulfill the state-dependent dilatancy equations. As a microscopic parameter to evaluate the static determinacy of the granular system, the redundancy ratio is defined and investigated. The results show that the critical state is very close to the statically determinate state. Other particle-level indexes, including the distribution of the contact forces and the anisotropies, are carefully investigated to analyze the microstructural evolution and the underlying mechanism. The microscopic analysis shows that both the contact orientations and contact forces influence the mechanical behaviors of granular materials.     

Landslide dam failure and flood hydraulics. Part I: experimental investigation

Landslide dam failure can trigger catastrophic flooding in the downstream. However, field observation of such flooding is rarely available, while laboratory experimental studies are sparse. The mechanism of landslide dam failure and the flood has so far remained insufficiently understood. Here, we present an experimental investigation of landslide dam failure and the flood. A total of 28 runs of experiments are carried out in a flume of 80 m × 1.2 m × 0.8 m, with differing inflow discharge, dam composition, dam geometry, and initial breach dimension. An array of twelve automatic water-level probes is deployed to measure the stage hydrographs along the flume, and the video recording of the dam failure processes facilitates an estimation of the widening of initial breach. Under the present experimental conditions with dams composed of homogeneous materials, landslide dam failure is primarily caused by erosion of overtopping flow, and lateral mass collapse is also considerable during the cause of breach widening. Cohesive clay may act to mitigate the seepage through the dam and thus its subsidence and appreciably modulate the dam failure process and the flood. However, the impacts of clay may be readily overwhelmed by a large inflow discharge and initial breach. Gravels in the dam may appreciably depress the rate of the dam failure process and thus modify the flood. The present work provides new experimental data set for testing mathematical models of the flood flow due to landslide dam failure.     

Extraction and accuracy assessment of high-resolution DEM and derived orthoimages from ALOS-PRISM data over Sahel-Doukkala (Morocco)

In Sahel-Doukkala, which is characterized by lands of a relatively low relief, global DEMs and DEMs generated from digitizing topographic maps, have been the primary source of several multidisciplinary researches. Although these products present a great value of the conducted research, the level of the given accuracy is not sufficient enough for detailed geospatial analysis. These requirements led us to generate a high-resolution DEM as an alternative of available global DEMs or/and DEMs generated from digitizing topographic maps. In this study, we present a workflow to extract high-resolution DEM at 5 m resolution and derived orthoimages from ALOS-PRISM data over Sahel-Doukkala, through photogrammetric techniques, using a variation of GCPs obtained from topographic maps at scale 1:25,000. The accuracy of the generated products is reported according to NSSDA standards. Using ten GCPs, a PRISM-DEM with 3.88 m vertical accuracy and 11.60 m horizontal accuracy, both at 95% confidence level is obtained. This DEM will serve as base dataset for further detailed geospatial analysis and mapping applications in order to identify the relationship between surface parameters and groundwater, and also to assess and understand all factors influencing the development of karst landscapes and consequently subsurface stability in the investigated area.     

Geomorphic features extraction from high-resolution topography: landslide crowns and bank erosion

In recent years, new remote-sensed technologies, such as airborne and terrestrial laser scanner, have improved the detail and the quality of topographic information, providing topographical high-resolution and high-quality data over larger areas better than other technologies. A new generation of high-resolution (≤3 m) digital terrain models (DTMs) is now available for different areas and is widely used by researchers, offering new opportunities for the scientific community. These data call for the development of a new generation of methodologies for an objective extraction of geomorphic features, such as channel heads, channel networks, bank geometry, debris-flow channel, debris-flow deposits, scree slope, landslide and erosion scars, etc. A high-resolution DTM is able to detect the divergence/convergence of areas related to unchannelized/channelized processes with better detail than a coarse DTM. In this work, we tested the performance of new methodologies for an objective extraction of geomorphic features related to shallow landsliding processes (landslide crowns), and bank erosion in a complex mountainous terrain. Giving a procedure that automatically recognizes these geomorphic features can offer a strategic tool to map natural hazard and to ease the planning and the assessment of alpine regions. The methodologies proposed are based on the detection of thresholds derived by the statistical analysis of variability of landform curvature. The study was conducted on an area located in the Eastern Italian Alps, where an accurate field survey on shallow landsliding, erosive channelized processes, and a high-quality set of both terrestrial and airborne laser scanner elevation data is available. The analysis was conducted using a high-resolution DTM and different smoothing factors for landform curvature calculation in order to test the most suitable scale of curvature calculation for the recognition of the selected features. The results revealed that (1) curvature calculation is strongly scale-dependent, and an appropriate scale for derivation of the local geometry has to be selected according to the scale of the features to be detected; (2) such approach is useful to automatically detect and highlight the location of shallow slope failures and bank erosion, and it can assist the interpreter/operator to correctly recognize and delineate such phenomena. These results highlight opportunities but also challenges in fully automated methodologies for geomorphic feature extraction and recognition.     

Landslide detection using deep learning and object-based image analysis

Landslides - Recent landslide detection studies have focused on pixel-based deep learning (DL) approaches. In contrast, intuitive annotation of landslides from satellite imagery is based on...     

Landslide susceptible areas identification using IDW and Ordinary Kriging interpolation techniques from hard soil depth at middle western Central Java,Indonesia

Natural Hazards - Initial assessment of landslide susceptible areas is important in designing landslide mitigation measures. This study, a part of our study on the developing a landslide spatial...     

Landslide susceptibility zonation using GIS and evidential belief function model

The main goal of this paper is to generate a landslide susceptibility map through evidential belief function (EBF) model by using Geographic Information System (GIS) for Qianyang County, Shaanxi Province, China. At first, a detailed landslide inventory map was prepared, and the following ten landslide-conditioning factors were collected: slope angle, slope aspect, curvature, plan curvature, profile curvature, altitude, distance to rivers, geomorphology, lithology, and rainfall. The landslides were detected from the interpretation of aerial photographs and supported by field surveys. A total of 81 landslides were randomly split into the following two parts: the training dataset 70 % (56 landslides) were used for establishing the model and the remaining 30 % (25 landslides) were used for the model validation. The ArcGIS was used to analyze landslide-conditioning factors and evaluate landslide susceptibility; as a result, a landslide susceptibility map was generated by using EBF and ArcGIS 10.0, thus divided into the following five susceptibility classes: very low, low, moderate, high, and very high. Finally, when we validated the accuracy of the landslide susceptibility map, both the success-rate and prediction-rate curve methods were applied. The results reveal that a final susceptibility map has the success rate of 83.31 % and the prediction rate of 79.41 %.     

Landslide detection in the Himalayas using machine learning algorithms and U-Net

Landslides - Event-based landslide inventories are essential sources to broaden our understanding of the causal relationship between triggering events and the occurring landslides. Moreover,...     

基于机器学习的滑坡易发性预测建模及其主控因子识别

不同机器学习预测滑坡易发性的建模过程及其不确定性有所差异,另外如何有效识别滑坡易发性的主控因子意义重大.针对上述问题,以支持向量机(support vector machine,简称SVM)和随机森林(random forest,简称RF)为例探讨了基于机器学习的滑坡易发性预测及其不确定性,创新地提出了"权重均值法"来...     

文家坝大型单体滑坡地质特征遥感调查与稳定性分析

平武县的文家坝滑坡是由汶川地震诱发,并曾阻断河流形成文家坝堰塞湖,威胁到下游南坝、锰矿厂和公路的安全,所以有必要对其地质特征进行调查及稳定性分析。通过对震后平武县SPOT卫星影像的解译,识别出了文家坝滑坡的地貌要素及滑坡物质组成,表面植被发育度,水文条件等地质特征。可以认为,文家坝滑坡为大型牵引式滑坡,目前仍处于不稳定状态,在余震、强降雨等条件下,有复滑的可能,并针对潜在威胁给出了具体的防治建议。     

Modeling Clam-inspired Burrowing in Dry Sand using Cavity Expansion Theory and DEM

Acta Geotechnica - The Atlantic razor clam exhibits exceptional penetration performance in wet sands by periodically expanding and contracting its shell and foot during burrowing. Essentially, this...     

基于地貌特征的浅层崩滑体遥感自动识别

建立了一种基于地貌特征的浅层崩滑体遥感自动识别方法,通过对地物的遥感光谱特性、几何形状和顺坡性相关的6次布尔运算完成对浅层崩滑体的自动识别.并以碧罗雪山一处高山峡谷地区为例,采用10m空间分辨率的SPOT-5多光谱影像和1:5万地形图生成的DEM(数字高程模型)作为数据源,进行了浅层崩滑体自动识别效果的检验.实验结果表明:①该方法顾及了地貌对浅层崩滑体空间几何形态的影响,可以有效地提高浅层崩滑体遥感自动识别的正确率;②识别对象与数据源的关系明确、决策阙值容易确定,便于使用决策树进行分类;③该方法对数据源要求较低.只需中等以上比例尺的DEM和拥有红、近红外的遥感影像即可;④粘连图斑的分割是该方法面临的主要难题.     

Numerical study on stress states and fabric anisotropies in soilbags using the DEM

Wrapping granular soils in geosynthetic containers, such as soilbags, results in a considerable increase in the bearing capacity due to the effective restraint on the dilatancy of the soil. This paper numerically investigates the stress states and fabric anisotropies in the wrapped soil using the discrete element method, providing a novel perspective for new insights into the reinforcement mechanisms and the development of constitutive relations for soilbags. The two most anticipated loading conditions, namely, unconfined compression and simple shear, are considered, and numerical predictions are compared to experimental results. During unconfined compression, both global and local p–q stress paths evolve linearly, having the same slope until the global failure of the wrapping geosynthetic. Under simple shear, the global stress path approaches the critical state line first and then turns to the compression line of the wrapped soil. Some local loading–unloading stress paths are observed, which may account for the high damping of soilbags during cyclic shear. The reduced fabric anisotropies of the normal and tangential force chains suggest greater confinement from the lateral sides of the geosynthetic container in either loading course. The performance and mechanisms of the soilbag earth reinforcement method, i.e., confinement and interlocking, can be better understood based on these new findings on the stress states and fabric anisotropies.