Extent Of Landslides Research Articles

Predicting landslide extent of satellite images based on image classification extraction

Abstract Landslide is a common geological disaster and the most destructive natural disaster at home and abroad. Therefore, automatic interpretation of landslide high-risk areas and potential landslide areas through different models and methods is one of the basic ways to prevent and reduce possible disasters. In this study, we improve the NDWI method by using the multispectral image data from the French satellite SPOT 5. Combined with infrared band sensors, the Normalized Difference Water Index (NDWI) of large exposed landslides is extracted by improving the NDWI index method by utilizing the characteristics of near-infrared (NIR) and short-wave infrared (SWIR) bands and replacing the green-wave band or the mid-infrared (MIR) band with the short-wave infrared (SWIR) band, which can be applied to the water resources management of high-risk areas, potential landslide areas, and potential landslide areas, and the water environment monitoring of water resources and disaster emergency management. We calculate the 2014 satellite images with a total area of 11351.6 ha, and after the NDWI index, the dark blue color represents the high-risk landslide area of 1609.1 ha (14175%), the light blue color represents the potential landslide area of 3971.773 ha (34.989%), and the brown color represents the vegetation area of 5770.727 ha (50.836%).

A three-dimensional smoothed particle hydrodynamics analysis of multiple retrogressive landslides in sensitive soil

Numerical modelling and analysis of three-dimensional multiple retrogressive landslides of Saint-Jude landslide in sensitive clay is conducted. The investigations of the three-dimensional effects in determining the extent of landslides are concerned. To accurately simulate the extremely large deformation caused by progressive soil failure, a special form of SPH formulation known as finite particle method (FPM) is adopted. A hypoelastic–plastic model with strain softening is implemented to capture the key feature of the mechanical behaviour of the sensitive clay in undrained condition. The multiple progressive failure process, the final run-out stretch, and the retrogression distance of the Saint-Jude landslide are compared quantitatively with the data from site investigation. Furthermore, three-dimensional effects and influences of various factors, including slope morphology and soil properties, are discussed for better understanding of the development of retrogressive landslides in sensitive soil.

Mapping Post-Earthquake Landslide Susceptibility Using U-Net, VGG-16, VGG-19, and Metaheuristic Algorithms

Landslides are among the most frequent secondary disasters caused by earthquakes in areas prone to seismic activity. Given the necessity of assessing the current seismic conditions for ensuring the safety of life and infrastructure, there is a rising demand worldwide to recognize the extent of landslides and map their susceptibility. This study involved two stages: First, the regions prone to earthquake-induced landslides were detected, and the data were used to train deep learning (DL) models and generate landslide susceptibility maps. The application of DL models was expected to improve the outcomes in both stages. Landslide inventory was extracted from Sentinel-2 data by using U-Net, VGG-16, and VGG-19 algorithms. Because VGG-16 produced the most accurate inventory locations, the corresponding results were used in the landslide susceptibility detection stage. In the second stage, landslide susceptibility maps were generated. From the total measured landslide locations (63,360 cells), 70% of the locations were used for training the DL models (i.e., convolutional neural network [CNN], CNN-imperialist competitive algorithm, and CNN-gray wolf optimizer [GWO]), and the remaining 30% were used for validation. The earthquake-induced landslide conditioning factors included the elevation, slope, plan curvature, valley depth, topographic wetness index, land cover, rainfall, distance to rivers, and distance to roads. The reliability of the generated susceptibility maps was evaluated using the area under the receiver operating characteristic curve (AUROC) and root mean square error (RMSE). The CNN-GWO model (AUROC = 0.84 and RMSE = 0.284) outperformed the other methods and can thus be used in similar applications. The results demonstrated the efficiency of applying DL in the natural hazard domain. The CNN-GWO predicted that approximately 38% of the total area consisted of high and very high susceptibility regions, mainly concentrated in areas with steep slopes and high levels of rainfall and soil wetness. These outcomes contribute to an enhanced understanding of DL application in the natural hazard domain. Moreover, using the knowledge of areas highly susceptible to landslides, officials can actively adopt steps to reduce the potential impact of landslides and ensure the sustainable management of natural resources.

On the estimation of landslide intensity, hazard and density via data-driven models

Maps that attempt to predict landslide occurrences have essentially stayed the same since 1972. In fact, most of the geo-scientific efforts have been dedicated to improve the landslide prediction ability with models that have largely increased their complexity but still have addressed the same binary classification task. In other words, even though the tools have certainly changed and improved in 50 years, the geomorphological community addressed and still mostly addresses landslide prediction via data-driven solutions by estimating whether a given slope is potentially stable or unstable. This concept corresponds to the landslide susceptibility, a paradigm that neglects how many landslides may trigger within a given slope, how large these landslides may be and what proportion of the given slope they may disrupt. The landslide intensity concept summarized how threatening a landslide or a population of landslide in a study area may be. Recently, landslide intensity has been spatially modeled as a function of how many landslides may occur per mapping unit, something, which has later been shown to closely correlate to the planimetric extent of landslides per mapping unit. In this work, we take this observation a step further, as we use the relation between landslide count and planimetric extent to generate maps that predict the aggregated size of landslides per slope, and the proportion of the slope they may affect. Our findings suggest that it may be time for the geoscientific community as a whole, to expand the research efforts beyond the use of susceptibility assessment, in favor of more informative analytical schemes. In fact, our results show that landslide susceptibility can be also reliably estimated (AUC of 0.92 and 0.91 for the goodness-of-fit and prediction skill, respectively) as part of a Log-Gaussian Cox Process model, from which the intensity expressed as count per unit (Pearson correlation coefficient of 0.91 and 0.90 for the goodness-of-fit and prediction skill, respectively) can also be derived and then converted into how large a landslide or several coalescing ones may become, once they trigger and propagate downhill. This chain of landslide intensity, hazard and density may lead to substantially improve decision-making processes related to landslide risk.

Borehole investigation on steep sloping ground

The Department of Transport and Main Roads (TMR), Engineering and Technology's (E&T) Geotechnical Section, have completed a research-based field trial aimed at developing a repeatable methodology to carry out geotechnical boreholes on steeply sloping ground. This has been presented in the context of slope failures, whereby the utilisation of this practise will provide subsurface information, which is directly representative of the area of interest, rather than that which is outside the landslide extent (current general practise within Australia). It is expected that the information gained from adopting this practice will enable optimum design and remediation treatments for landslides within Australia, and subsequently contribute strongly to sustainable project outcomes.

A Multi-Criteria Decision Analysis (MCDA) Approach for Landslide Susceptibility Mapping of a Part of Darjeeling District in North-East Himalaya, India

Landslides are the nation’s hidden disaster, significantly increasing economic loss and social disruption. Unfortunately, limited information is available about the depth and extent of landslides. Therefore, in order to identify landslide-prone zones in advance, a well-planned landslide susceptibility mapping (LSM) approach is needed. The present study evaluates the efficacy of an MCDA-based model (analytical hierarchy process (AHP)) and determines the most accurate approach for detecting landslide-prone zones in one part of Darjeeling, India. LSM is prepared using remote sensing thematic layers such as slope, rainfall earthquake, lineament density, drainage density, geology, geomorphology, aspect, land use and land cover (LULC), and soil. The result obtained is classified into four classes, i.e., very high (11.68%), high (26.18%), moderate (48.87%), and low (13.27%) landslide susceptibility. It is observed that an entire 37.86% of the area is in a high to very high susceptibility zone. The efficiency of the LSM was validated with the help of the receiver operating characteristics (ROC) curve, which demonstrate an accuracy of 96.8%, and the success rate curve showed an accuracy of 81.3%, both of which are very satisfactory results. Thus, the proposed framework will help natural disaster experts to reduce land vulnerability, as well as aid in future development.

Deep Learning Method of Landslide Inventory Map with Imbalanced Samples in Optical Remote Sensing

Landslide inventory mapping (LIM) is a key prerequisite for landslide susceptibility evaluation and disaster mitigation. It aims to record the location, size, and extent of landslides in each map scale. Machine learning algorithms, such as support vector machine (SVM) and random forest (RF), have been increasingly applied to landslide detection using remote sensing images in recent decades. However, their limitations have impeded their wide application. Furthermore, despite the widespread use of deep learning algorithms in remote sensing, for LIM, deep learning algorithms are limited to less unbalanced landslide samples. To this end, in this study, full convolution networks with focus loss (FCN-FL) were adopted to map historical landslides in regions with imbalanced samples using an improved symmetrically connected full convolution network and focus loss function to increase the feature level and reduce the contribution of the background loss value. In addition, K-fold cross-validation training models (FCN-FLK) were used to improve data utilization and model robustness. Results showed that the recall rate, F1-score, and mIoU of the model were improved by 0.08, 0.09, and 0.15, respectively, compared to FCN. It also demonstrated advantages over U-Net and SegNet. The results prove that the method proposed in this study can solve the problem of imbalanced sample in landslide inventory mapping. This research provides a reference for addressing imbalanced samples in the deep learning of LIM.

Monitoring fotogrametryczny z wykorzystaniem bezzałogowego statku powietrznego na przykładzie osuwiska zlokalizowanego na zboczach poeksploatacyjnego zbiornika wodnego w KWB Adamów

The article presents the results of monitoring of landslides related to opencast mining. The landslides formed on the slopes of an artificial water reservoir in the Adamów open-cast lignite mine. The photogrammetric method was used for monitoring. The study of landslide activity was performed on the basis of six photogrammetric surveys made with the DJI Phantom 4 Advanced UAV. Changes in the landslide extent and vertical displacements were determined in five measurement periods. For four periods, differential digital terrain models were made and the volume of displaced earth masses was calculated. The future extent of the landslide was determined on the basis of cracks visible in high-resolution optical photos. The research has shown that UAVs equipped with an optical camera are very useful in monitoring the stability of slopes of post-mining water reservoirs. The applied method provided detailed data on the size of displacements without compromising on safety of people involved in field surveys.

Analysis of Landslide Susceptibility using Quantitative Multivariate Prediction Models in Jenelata Sub-Watershed, Jeneberang Watershed South Sulawesi Indonesia

Landslides have taken an essential role in the evolution of landscapes, are present almost everywhere, and have become one of the catastrophic subjects of frequent natural disasters. The landslide disaster in Indonesia became one of the disasters that resulted in many casualties. Quantitative multivariate prediction models are statistical methods that allow us to research more than two variables simultaneously. Multivariate analysis is used because, in reality, the problem cannot be solved by simply connecting two variables or looking at the influence of one variable on another. The Jenelata Sub-Watershed of the Jeneberang Watershed is the upper reaches of the Jeneberang watershed, with areas that have undergone many changes in land use and experienced a very high erosion rate. It can be seen by the river flow, which always carries soil sedimentation. The 650 landslide events detected in Google Earth imagery serve as the basis for assessing the extent of landslides. Discrimination of causal factors in frequency ratio (FR) indicates that slope factors, river density, and slope direction are the leading causes of landslides. The NGO was then built using a logistic regression (LR) model. The ROC curve in the model is more inclined towards the sensitivity axis with the success of the AUC and predictions of 0.809 model validation and 0.801 prediction validation, respectively.

A national landslide inventory for Denmark

Abstract. Landslides are a frequent natural hazard occurring globally in regions with steep topography. Additionally, landslides play an important role in landscape evolution by transporting sediment downslope. Landslide inventory mapping is a common technique to assess the spatial distribution and extent of landslides in an area of interest. High-resolution digital elevation models (DEMs) have proven to be useful databases to map landslides in large areas across different land covers and topography. So far, Denmark had no national landslide inventory. Here, we create the first comprehensive national landslide inventory for Denmark derived from a 40 cm resolution DEM from 2015 supported by several 12.5 cm resolution orthophotos. The landslide inventory is created based on a manual expert-based mapping approach, and we implemented a quality control mechanism to assess the completeness of the inventory. Overall, we mapped 3202 landslide polygons in Denmark with a level of completeness of 87 %. The complete landslide inventory is freely available for download at https://doi.org/10.6084/m9.figshare.16965439.v2 (Svennevig and Luetzenburg, 2021) or as a web map (https://data.geus.dk/landskred/, last access: 6 June 2022) for further investigations.

Framework for the identification of shallow ground movement in modified slopes (an expert opinion)

This paper has discussed and analyzed some significant ways to quickly find the root course of landslides on slopes, especially on shallow artificial slopes. Many mechanisms that trigger ground movement leading to landslides in slopes are discussed here. The general idea was to study the occurrence of slips on a modified slope under certain environmental conditions for a specific duration and develop a simple framework that can help to identify landslides’ causative factors subjectively. This whole idea was to help locals and other experts easily identify the triggering mechanisms of landslides to reveal the possible time when the slides will occur and control their impact. Furthermore, a framework was developed to help us understand the critical factors leading to shallow landslides on modified slopes. Establishing the mechanism that triggers landslides in an area helps greatly identify the landslides-prone locations, especially for the locals living in high-risk areas. The research also presented significant forms of landslides with their presumed causes. Although the general causes of landslides are not limited to what was presented in this text, the environment is the main factor deciding the nature and extent of landslides in an area.

A Century of Landslides in Seattle, Washington: Coalescing and Digitizing the City's Historic Landslide Inventories

ABSTRACT Landslides have impacted the built environment of the city of Seattle for over a century. We present a new historic landslide inventory for Seattle built from over 100 years of landslide records. Seattle has tracked landslide occurrence since the 1890s and has commissioned various studies to examine where and why landslides occur. During this tenure, methods for collection, display, and dissemination of landslide data have varied, resulting in a complex, non-uniform, and rich dataset. For the new database completed in 2019, multiple inventories and thousands of historic documents were combined to meet the City of Seattle's objectives of keeping landslide areas precise with respect to property boundaries. Compared with prior maps, the new map provides more information from the historical record, maps landslide extent more accurately, and tags fewer private properties with landslide features than prior maps. In addition, digitized historic documents are now attached to landslide features in the City of Seattle's public geographic information system (GIS) map. The historic inventory is complementary to a light detection and ranging (LiDAR) inventory covering the same area. This case study shows how historic data can be used to record landslides in urban areas where geomorphological signatures may be removed by development. In Seattle, access to information about historic landslides will streamline workflows for public and private engineers and geologists and will serve property owners and the general public. The map is currently available at https://www.arcgis.com/apps/webappviewer/index.html?id = f822b2c6498c4163b0cf908e2241e9c2 = f822b2c6498c4163b0cf908e2241e9c2.

Combining Statistical Design with Deterministic Modelling to Assess the Effect of Site-Specific Factors on the Extent of Landslides

Combining Statistical Design with Deterministic Modelling to Assess the Effect of Site-Specific Factors on the Extent of Landslides

Landslide size matters: A new data-driven, spatial prototype

The standard definition of landslide hazard requires the estimation of where, when (or how frequently) and how large a given landslide event may be. The geoscientific community involved in statistical models has addressed the component pertaining to how large a landslide event may be by introducing the concept of landslide-event magnitude scale. This scale, which depends on the planimetric area of the given population of landslides, in analogy to the earthquake magnitude, has been expressed with a single value per landslide event. As a result, the geographic or spatially-distributed estimation of how large a population of landslide may be when considered at the slope scale, has been disregarded in statistically-based landslide hazard studies. Conversely, the estimation of the landslide extent has been commonly part of physically-based applications, though their implementation is often limited to very small regions.In this work, we initially present a review of methods developed for landslide hazard assessment since its first conception decades ago. Subsequently, we introduce for the first time a statistically-based model able to estimate the planimetric area of landslides aggregated per slope units. More specifically, we implemented a Bayesian version of a Generalized Additive Model where the maximum landslide size per slope unit and the sum of all landslide sizes per slope unit are predicted via a Log-Gaussian model. These “max” and “sum” models capture the spatial distribution of (aggregated) landslide sizes. We tested these models on a global dataset expressing the distribution of co-seismic landslides due to 24 earthquakes across the globe. The two models we present are both evaluated on a suite of performance diagnostics that suggest our models suitably predict the aggregated landslide extent per slope unit. In addition to a complex procedure involving variable selection and a spatial uncertainty estimation, we built our model over slopes where landslides triggered in response to seismic shaking, and simulated the expected failing surface over slopes where the landslides did not occur in the past.What we achieved is the first statistically-based model in the literature able to provide information about the extent of the failed surface across a given landscape. This information is vital in landslide hazard studies and should be combined with the estimation of landslide occurrence locations. This could ensure that governmental and territorial agencies have a complete probabilistic overview of how a population of landslides could behave in response to a specific trigger. The predictive models we present are currently valid only for the 25 cases we tested. Statistically estimating landslide extents is still at its infancy stage. Many more applications should be successfully validated before considering such models in an operational way. For instance, the validity of our models should still be verified at the regional or catchment scale, as much as it needs to be tested for different landslide types and triggers. However, we envision that this new spatial predictive paradigm could be a breakthrough in the literature and, in time, could even become part of official landslide risk assessment protocols.

A multi-temporal mapping approach for improving the temporal and spatial characterization of landslide activity in clay-rich terrain

In mountain, clay-rich physiographic settings of the Northern Apennines, earth slides and earth flows are dominant processes of hillslope sediment transfer, which cause hazard and pose risk to lives and man-made structures. In Emilia-Romagna (E-R), information drawn from a region-wide inventory is customarily used for assessing landslide hazard potential and partly guides decision making on land management and urban planning. However, the inventory's historical (i.e., post-1954) degree of completeness is presently unknown. To address this gap, we integrate and compare the existing E-R inventory with a multi-temporal landslide mapping approach at four selected study sites of the Sillaro River basin, where we perform visual interpretation of thirteen sequential aerial photo sets in the 1954- 2020 period and conduct confirmatory fieldwork. Specific objectives include evaluating the advantages of higher resolution mapping for characterizing (i) landslide extent, geometry and activity through time; and (ii) landslide interaction with the drainage network. With reference to the E-R inventory, multi-temporal mapping allowed reducing mapping uncertainty through the identification of additional first occurrences and recurrences (up to a factor of 5), additional mapped landslide area (up to a factor of 12), and combined footprint area (up to a factor of 9). Our findings indicated that the proposed approach does not always result in mapping larger (or smaller) landslide polygons, or previously undetected landslides, but provides insights on process understanding. Therefore, an added value of this mapping approach lies in its ability to detect different styles of landsliding and define most likely scenarios of evolution at the site scale. Insights were drawn through detection of: (i) headscarp upslope migration (or revegetation); (ii) newly formed gully channels and/or revegetation on landslide depositional features; and (iii) changes at landslide terminus, which allowed identifying new sediment pulses, exact delivery targets (e.g., fan, valley floor, and stream channel), as well as induced changes in channel bed texture and geometry.

Random Forests for Landslide Prediction in Tsengwen River Watershed, Central Taiwan

Landslides have been identified as one of the costliest and deadliest natural disasters, causing tremendous damage to humans and societies. Information regarding the spatial extent of landslides is thus important to allow officials to devise successful strategies to mitigate landslide hazards. This study aims to develop a machine-learning approach for predicting landslide areas in the Tsengwen River Watershed (TRW), which is one of the most landslide-prone areas in Central Taiwan. Various spatial datasets were collected from 2009 to 2015 to derive 36 predictive variables used for landslide modeling with random forests (RF). The results of landslide prediction, compared with ground reference data, indicated an overall accuracy of 91.4% and Kappa coefficient of 0.83, respectively. The findings achieved from estimates of predictor importance also indicated to officials that the land-use/land-cover (LULC) type, distance to previous landslides, distance to roads, bank erosion, annual groundwater recharge, geological line density, aspect, and slope are the most influential factors that trigger landslides in the study region.

Multi-Regional landslide detection using combined unsupervised and supervised machine learning

Landslide detection is concerned with delineating the extent of landslides. Most of existing works on landslide detection have limited geographical extents. Therefore, the models developed out of these studies might perform poorly when applied to regions with different characteristics. This study investigates an Object-Based Image Analysis methodology built on unsupervised and supervised Machine Learning to detect the location of landslides occurred in multiple regions across the world. The utilized data includes Sentinel-2 multi-spectral satellite imagery and ALOS Digital Elevation Model. In the segmentation stage, pre and post-landslide images undergo segmentation using K-means clustering. Following the segmentation stage and dataset preparation and removing highly-correlated features from the dataset, two Random Forest classifiers (RF1 and RF2) are trained and tested on two different datasets to measure the generalization level of the algorithms with RF1 dataset spanning over more geographical diversities than RF2 dataset. The results show that the RF models can successfully detect landslide segments with test precision = 0.96 and recall = 0.96 for RF1 and test precision = 0.90 and recall = 0.87 for RF2. Further validation shows that, compared to RF2, RF1 results in less mislabelled non-landslide segments.

DETERMINATION OF LAND COVER AS LANDSLIDE FACTOR BASED ON MULTITEMPORAL RASTER DATA IN MALANG REGENCY

Malang Regency is one of the regency administration areas in East Java Province. The complexphysical condition of Malang Regency, especially its land reliefs accompanied by high rainfall, causes MalangRegency obtain floods and landslides frequently. Based on the BPBD data in 2017, it is stated that around 80%of Malang Regency is disaster-prone area of floods and landslides, especially in the southern part of MalangRegency. This is caused by several natural factors, one of which is slope, and human factors called changes inland cover. This study aims to examine the land cover factors that influence the extent of landslides that occurin the Southern Malang Regency. The independent variable used is builtup area, dense tree vegetation,vegetation of rare trees, shrubs, grass, open space, slope, and distance. Based on the results of multiple linearregression analysis, it can be concluded that the builtup area cover with coefficient value 0.062 and open spacewith coefficient value 0.020 are variables that are directly proportional to the area of the landslide that has ahigher percentage of the area of land cover being built and open, the area of landslides will increase, while thevariable cover of vegetated land has properties inversely proportional to the extent of landslides which meansthat it can reduce the extent of landslides if the percentage of vegetated land cover increases.

Landslide Detection Using Multi-Scale Image Segmentation and Different Machine Learning Models in the Higher Himalayas

Landslides represent a severe hazard in many areas of the world. Accurate landslide maps are needed to document the occurrence and extent of landslides and to investigate their distribution, types, and the pattern of slope failures. Landslide maps are also crucial for determining landslide susceptibility and risk. Satellite data have been widely used for such investigations—next to data from airborne or unmanned aerial vehicle (UAV)-borne campaigns and Digital Elevation Models (DEMs). We have developed a methodology that incorporates object-based image analysis (OBIA) with three machine learning (ML) methods, namely, the multilayer perceptron neural network (MLP-NN) and random forest (RF), for landslide detection. We identified the optimal scale parameters (SP) and used them for multi-scale segmentation and further analysis. We evaluated the resulting objects using the object pureness index (OPI), object matching index (OMI), and object fitness index (OFI) measures. We then applied two different methods to optimize the landslide detection task: (a) an ensemble method of stacking that combines the different ML methods for improving the performance, and (b) Dempster–Shafer theory (DST), to combine the multi-scale segmentation and classification results. Through the combination of three ML methods and the multi-scale approach, the framework enhanced landslide detection when it was tested for detecting earthquake-triggered landslides in Rasuwa district, Nepal. PlanetScope optical satellite images and a DEM were used, along with the derived landslide conditioning factors. Different accuracy assessment measures were used to compare the results against a field-based landslide inventory. All ML methods yielded the highest overall accuracies ranging from 83.3% to 87.2% when using objects with the optimal SP compared to other SPs. However, applying DST to combine the multi-scale results of each ML method significantly increased the overall accuracies to almost 90%. Overall, the integration of OBIA with ML methods resulted in appropriate landslide detections, but using the optimal SP and ML method is crucial for success.

GIS-based landslide susceptibility mapping using hybrid integration approaches of fractal dimension with index of entropy and support vector machine

The loess area in the northern part of Baoji City, Shaanxi Province, China is a region with frequently landslide occurrences. The main aim of this study is to quantitatively predict the extent of landslides using the index of entropy model (IOE), the support vector machine model (SVM) and two hybrid models namely the F-IOE model and the F-SVM model constructed by fractal dimension. First, a total of 179 landslides were identified and landslide inventory map was produced, with 70% (125) of the landslides which was optimized by 10-fold cross-validation being used for training purpose and the remaining 30% (54) of landslides being used for validation purpose. Subsequently, slope angle, slope aspect, altitude, rainfall, plan curvature, distance to rivers, land use, distance to roads, distance to faults, normalized difference vegetation index (NDVI), lithology, and profile curvature were considered as landslide conditioning factors and all factor layers were resampled to a uniform resolution. Then the information gain ratio of each conditioning factors was evaluated. Next, the fractal dimension for each conditioning factors was calculated and the training dataset was used to build four landslide susceptibility models. In the end, the receiver operating characteristic (ROC) curves and three statistical indexes involving positive predictive rate (PPR), negative predictive rate (NPR) and accuracy (ACC) were applied to validate and compare the performance of these four models. The results showed that the F-SVM model had the highest PPR, NPR, ACC and AUC values for training and validation datasets, respectively, followed by the F-IOE model. Finally, it is concluded that the F-SVM model performed best in all models, the hybrid model built by fractal dimension has advantages than original model, and can provide reference for local landslide prevention and decision making.