Method, device and computer equipment for assessing the risk of landslides caused by rainfall clusters

Abstract

The present application relates to a method, device, and computer equipment for evaluating the risk of rainfall-induced landslide groups. The method includes: obtaining historical rainfall-induced landslide data of historical heavy rainfall events to construct landslide sample data; analyzing based on landslide attribute parameters and the distribution information of rainfall-induced landslide data in the landslide sample data to obtain landslide development characteristics; according to landslide geographical characteristics and landslide geometric characteristics, using landslide sample data and multiple candidate landslide influence indicators to analyze the background law of landslide generation, and establishing a landslide risk evaluation index system; through correlation analysis of multiple candidate landslide influence indicators, determining the target landslide influence indicator, and combining the target landslide influence indicator and the landslide risk evaluation index system to obtain a landslide risk assessment model constructed based on an automated machine learning framework. Using this method can improve the efficiency of rainfall-induced landslide risk assessment and effectively improve the accuracy of rainfall-induced landslide risk assessment.

Description

Technical Field

[0001] The present application relates to the field of computer technologies, and particularly to a method, apparatus, computer device, computer-readable storage medium, and computer program product for evaluating the risk of rainfall-induced mass landslides. Background Art

[0002] In the prevention and control of geological disasters, landslide disasters induced by rainfall are issues that require key attention. At present, qualitative methods or quantitative methods are usually used for landslide risk assessment.

[0003] In the related art, the traditional qualitative method relies on the experience judgment of experts, which has subjective errors; the traditional quantitative method analyzes the landslide occurrence mechanism through physical models to predict landslides, but it requires a large number of geotechnical parameters as model inputs, restricting its application in large-scale areas. Moreover, with the complexity of the model in structure and technology, the professional technical background requirements for users are getting higher and higher, affecting the efficiency of rainfall-induced mass landslide risk assessment and resulting in poor assessment effects. Summary of the Invention

[0004] Based on this, it is necessary to provide a method, apparatus, computer device, computer-readable storage medium, and computer program product for evaluating the risk of rainfall-induced mass landslides, which can improve the efficiency and effect of rainfall-induced mass landslide risk assessment, aiming at the above technical problems.

[0005] In a first aspect, the present application provides a method for evaluating the risk of rainfall-induced mass landslides, including:

[0006] Obtaining historical rainfall-induced mass landslide data of historical heavy rainfall events and constructing landslide sample data;

[0007] Analyzing based on landslide attribute parameters and the distribution information of rainfall-induced mass landslide data in the landslide sample data to obtain landslide development characteristics; the landslide development characteristics include landslide geographical characteristics and landslide geometric state characteristics;

[0008] According to the landslide geographical characteristics and the landslide geometric characteristics, using the landslide sample data and multiple candidate landslide influence indicators to analyze the background laws of landslide generation, and establishing a landslide risk assessment index system;

[0009] Through performing a correlation analysis on the multiple candidate landslide influence indicators, determining target landslide influence indicators, and combining the target landslide influence indicators and the landslide risk assessment index system, obtaining a landslide risk assessment model constructed based on an automated machine learning framework; the landslide risk assessment model is used for evaluating the risk of rainfall-induced mass landslides for heavy rainfall events.

[0010] In one embodiment, obtaining historical rainfall-induced landslide data of historical heavy rainfall events and constructing landslide sample data includes:

[0011] Obtaining the historical rainfall-induced landslide data based on satellite images of the historical heavy rainfall events;

[0012] Through data verification processing and data supplementation processing of the historical rainfall-induced landslide data, landslide sample data containing the distribution of rainfall-induced landslide data is obtained.

[0013] In one embodiment, analyzing the distribution information of rainfall-induced landslide data based on landslide attribute parameters and the landslide sample data to obtain landslide development characteristics includes:

[0014] Based on the distribution information of rainfall-induced landslide data in the landslide sample data, obtaining the distribution pattern of rainfall-induced landslide data in the geographical space, and determining the landslide density parameter and the landslide geometric parameter as the landslide attribute parameters;

[0015] Analyzing the landslide mobility of the landslide sample data according to the landslide attribute parameters to obtain the landslide geographical characteristics and the landslide geometric characteristics as the landslide development characteristics.

[0016] In one embodiment, according to the landslide geographical characteristics and the landslide geometric characteristics, using the landslide sample data and multiple candidate landslide influencing indicators to analyze the background law of landslide generation and establishing a landslide hazard evaluation index system includes:

[0017] Obtaining multiple candidate landslide influencing indicators associated with the regional scope of the historical heavy rainfall event;

[0018] Combining the landslide geographical characteristics and the landslide geometric characteristics, analyzing the landslide sample data and the multiple candidate landslide influencing indicators to determine the landslide spatial distribution law and the landslide disaster-forming environment;

[0019] According to the landslide spatial distribution law, the landslide disaster-forming environment, and the multiple candidate landslide influencing indicators, constructing the landslide hazard evaluation index system.

[0020] In one embodiment, through correlation analysis of the multiple candidate landslide influencing indicators to determine the target landslide influencing indicator, and combining the target landslide influencing indicator and the landslide hazard evaluation index system to obtain a landslide hazard assessment model constructed based on an automated machine learning framework, includes:

[0021] Through correlation analysis of the multiple candidate landslide influencing indicators, taking the selected candidate landslide influencing indicators as the target landslide influencing indicators;

[0022] From the landslide sample data, landslide samples and non-landslide samples are selected according to the target landslide impact indicators as training sample data.

[0023] Combining the target landslide impact indicators and the landslide hazard assessment index system, an initial evaluation model is constructed using an automated machine learning framework.

[0024] The initial evaluation model is trained based on the training sample data to obtain the landslide hazard assessment model.

[0025] In one embodiment, the method further includes:

[0026] Using preset model evaluation indicators to detect the performance of the landslide hazard assessment model.

[0027] Based on the landslide hazard assessment model that passes the detection, the rainfall-induced landslide hazard of the historical heavy rainfall event is evaluated to obtain a landslide hazard assessment result.

[0028] According to the landslide hazard assessment result, a rainfall-induced landslide hazard prompt image of the historical heavy rainfall event is displayed.

[0029] Among them, the model evaluation indicators include any one or more of the following:

[0030] Area under the curve indicator, accuracy indicator, precision indicator, recall indicator.

[0031] In a second aspect, the present application also provides a rainfall-induced landslide hazard assessment device, including:

[0032] A landslide sample data construction module, configured to obtain historical rainfall-induced landslide data of historical heavy rainfall events and construct landslide sample data.

[0033] A landslide development feature acquisition module, configured to analyze based on landslide attribute parameters and the distribution information of the rainfall-induced landslide data in the landslide sample data to obtain landslide development features; the landslide development features include landslide geographical features and landslide geometric state features.

[0034] A landslide evaluation system establishment module, configured to analyze the background rules of landslide generation according to the landslide geographical features and the landslide geometric features, using the landslide sample data and multiple candidate landslide impact indicators, and establish a landslide hazard assessment index system.

[0035] The landslide hazard assessment model obtaining module is used to determine the target landslide impact indicators by performing correlation analysis on the multiple candidate landslide impact indicators, and combine the target landslide impact indicators and the landslide hazard evaluation index system to obtain a landslide hazard assessment model constructed based on the automated machine learning framework; the landslide hazard assessment model is used to perform rainfall-induced landslide hazard assessment on heavy rainfall events.

[0036] In a third aspect, the present application also provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0037] Obtain historical rainfall-induced landslide data of historical heavy rainfall events and construct landslide sample data;

[0038] Analyze based on the landslide attribute parameters and the distribution information of the rainfall-induced landslide data of the landslide sample data to obtain landslide development characteristics; the landslide development characteristics include landslide geographical characteristics and landslide geometric state characteristics;

[0039] According to the landslide geographical characteristics and the landslide geometric characteristics, use the landslide sample data and multiple candidate landslide impact indicators to analyze the background rules of landslide generation, and establish a landslide hazard evaluation index system;

[0040] Determine the target landslide impact indicators by performing correlation analysis on the multiple candidate landslide impact indicators, and combine the target landslide impact indicators and the landslide hazard evaluation index system to obtain a landslide hazard assessment model constructed based on the automated machine learning framework; the landslide hazard assessment model is used to perform rainfall-induced landslide hazard assessment on heavy rainfall events.

[0041] In a fourth aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0042] Obtain historical rainfall-induced landslide data of historical heavy rainfall events and construct landslide sample data;

[0043] Analyze based on the landslide attribute parameters and the distribution information of the rainfall-induced landslide data of the landslide sample data to obtain landslide development characteristics; the landslide development characteristics include landslide geographical characteristics and landslide geometric state characteristics;

[0044] According to the landslide geographical characteristics and the landslide geometric characteristics, use the landslide sample data and multiple candidate landslide impact indicators to analyze the background rules of landslide generation, and establish a landslide hazard evaluation index system;

[0045] By performing a correlation analysis on the multiple candidate landslide impact indicators, the target landslide impact indicator is determined. Combining the target landslide impact indicator and the landslide hazard assessment index system, a landslide hazard assessment model constructed based on an automated machine learning framework is obtained. The landslide hazard assessment model is used to assess the hazard of rainfall-induced mass landslides for heavy rainfall events.

[0046] In a fifth aspect, the present application also provides a computer program product, including a computer program which, when executed by a processor, implements the following steps:

[0047] Obtain historical rainfall-induced mass landslide data of historical heavy rainfall events and construct landslide sample data;

[0048] Analyze based on the landslide attribute parameters and the distribution information of the rainfall-induced mass landslide data in the landslide sample data to obtain landslide development characteristics. The landslide development characteristics include landslide geographical characteristics and landslide geometric state characteristics;

[0049] According to the landslide geographical characteristics and the landslide geometric characteristics, use the landslide sample data and multiple candidate landslide impact indicators to analyze the background laws of landslide generation, and establish a landslide hazard assessment index system;

[0050] By performing a correlation analysis on the multiple candidate landslide impact indicators, the target landslide impact indicator is determined. Combining the target landslide impact indicator and the landslide hazard assessment index system, a landslide hazard assessment model constructed based on an automated machine learning framework is obtained. The landslide hazard assessment model is used to assess the hazard of rainfall-induced mass landslides for heavy rainfall events.

[0051] The above-mentioned method, device, computer equipment, computer-readable storage medium, and computer program product for assessing the hazard of rainfall-induced mass landslides obtain historical rainfall-induced mass landslide data of historical heavy rainfall events, construct landslide sample data, analyze based on the landslide attribute parameters and the distribution information of the rainfall-induced mass landslide data in the landslide sample data to obtain landslide development characteristics, which include landslide geographical characteristics and landslide geometric state characteristics. Then, according to the landslide geographical characteristics and the landslide geometric characteristics, use the landslide sample data and multiple candidate landslide impact indicators to analyze the background laws of landslide generation, establish a landslide hazard assessment index system, and further determine the target landslide impact indicator by performing a correlation analysis on the multiple candidate landslide impact indicators. Combining the target landslide impact indicator and the landslide hazard assessment index system, a landslide hazard assessment model constructed based on an automated machine learning framework is obtained. The landslide hazard assessment model is used to assess the hazard of rainfall-induced mass landslides for heavy rainfall events, realizing the optimization of the assessment of the hazard of rainfall-induced mass landslides, being able to improve the efficiency of the assessment of the hazard of rainfall-induced mass landslides, and effectively improving the accuracy of the assessment of the hazard of rainfall-induced mass landslides. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for the description of the embodiments of the present application or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0053] Figure 1 It is a schematic flowchart of a method for evaluating the risk of rainfall-induced landslide clusters in an embodiment;

[0054] Figure 2 It is a schematic diagram of the process for evaluating the risk of rainfall-induced landslide clusters based on an automated machine learning framework in an embodiment;

[0055] Figure 3 It is a schematic flowchart of a method for evaluating the risk of rainfall-induced landslide clusters in another embodiment;

[0056] Figure 4 It is a structural block diagram of a device for evaluating the risk of rainfall-induced landslide clusters in an embodiment;

[0057] Figure 5 It is an internal structure diagram of a computer device in an embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0058] In order to make the objectives, technical solutions and advantages of the present application clearer, the following further details the present application in conjunction with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0059] In an exemplary embodiment, as Figure 1 shown, a method for evaluating the risk of rainfall-induced landslide clusters is provided. In this embodiment, it is exemplified that the method is applied to a terminal. It can be understood that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps 101 to 104. Among them:

[0060] Step 101, obtain historical rainfall-induced landslide data of historical heavy rainfall events, and construct landslide sample data.

[0061] In practical applications, high-resolution satellite images of historical heavy rainfall events can be obtained for data extraction, and the extracted landslide data can be verified, supplemented and other data preprocessing to obtain landslide sample data containing the distribution of rainfall-induced landslide data.

[0062] Step 102: Analyze the rainfall-induced landslide data distribution information based on the landslide attribute parameters and the landslide sample data to obtain the landslide development characteristics.

[0063] As an example, the landslide development characteristics may include the landslide geographical characteristics and the landslide geometric state characteristics.

[0064] In a specific implementation, landslide attribute parameters such as landslide density, landslide length, landslide width, and landslide height can be used to determine the landslide geographical characteristics and landslide geometric characteristics, and obtain the landslide development characteristics that characterize the landslide development law.

[0065] Step 103: According to the landslide geographical characteristics and the landslide geometric characteristics, use the landslide sample data and multiple candidate landslide influence indicators to analyze the background law of landslide generation, and establish a landslide hazard assessment index system.

[0066] Exemplarily, data on internal and external influence factors of landslides within the area of historical heavy rainfall events can be collected to determine multiple candidate landslide influence indicators. Furthermore, according to the landslide geographical characteristics and the landslide geometric characteristics, the landslide sample data and multiple candidate landslide influence indicators can be used to analyze the background law of landslide generation. For example, the landslide disaster-forming environment and the landslide spatial distribution law can be analyzed to determine the potential conditions for landslide occurrence and the distribution characteristics in the geographical space, and establish a landslide hazard assessment index system.

[0067] Step 104: Through correlation analysis of the multiple candidate landslide influence indicators, determine the target landslide influence indicator, and combine the target landslide influence indicator and the landslide hazard assessment index system to obtain a landslide hazard assessment model constructed based on the automated machine learning framework.

[0068] Among them, the landslide hazard assessment model can be used to assess the hazard of rainfall-induced landslides in heavy rainfall events.

[0069] In an example, correlation tests can be performed on multiple candidate landslide influence indicators to screen out the target landslide influence indicator. Furthermore, a model based on the automated machine learning framework can be established and model training can be carried out by combining the target landslide influence indicator and the landslide hazard assessment index system to obtain a landslide hazard assessment model. Thus, based on the heavy rainfall scenario, by comprehensively considering the rainfall-induced landslide data and the internal and external factors affecting landslides, and using the integrated model in the automated machine learning framework for model construction, the hazard assessment of rainfall-induced landslides is realized.

[0070] In yet another example, with the rapid development of artificial intelligence, machine learning can be applied to solve problems of non-linear relationships, especially in the aspect of ensemble learning models. Machine learning methods can learn and discover hidden and unknown patterns in databases with high precision, can process and analyze a large amount of complex data, such as various influencing factors like rainfall, terrain, vegetation, etc., and it has strong adaptability and flexibility. Based on the automated machine learning framework, it can well alleviate the difficulties that researchers may face when using machine learning, and using the automated machine learning framework can broaden the application scope of machine learning models and reduce excessive interaction with users.

[0071] Compared with traditional methods, the technical solution of this embodiment extracts rainfall-induced mass landslide data based on high-resolution satellite images, combines landslide density, geometric parameters, etc. to reveal the geographical development characteristics and geometric characteristics of landslides, and under the condition of collecting internal and external influencing factors of landslides, conducts in-depth analysis of the spatial distribution law of landslides, establishes a landslide hazard assessment index system, and then can carry out the rainfall-induced mass landslide hazard assessment based on the automated machine learning framework. It can effectively solve the deficiencies of traditional landslide hazard assessment methods in terms of efficiency and accuracy, provides support for carrying out the mass landslide hazard assessment in strong rainfall scenarios, and meets the needs of geological disaster assessment and prevention in rainfall-prone areas.

[0072] In the above rainfall-induced mass landslide hazard assessment method, by obtaining the historical rainfall-induced mass landslide data of historical strong rainfall events, constructing landslide sample data, analyzing based on the landslide attribute parameters and the rainfall-induced mass landslide data distribution information of the landslide sample data to obtain the landslide development characteristics, and then according to the landslide geographical characteristics and landslide geometric characteristics, using the landslide sample data and multiple candidate landslide influencing indicators to analyze the background law of landslide generation, establishing a landslide hazard assessment index system, and then through the correlation analysis of multiple candidate landslide influencing indicators, determining the target landslide influencing indicator, combining the target landslide influencing indicator and the landslide hazard assessment index system, obtaining a landslide hazard assessment model constructed based on the automated machine learning framework, realizing the optimization of the rainfall-induced mass landslide hazard assessment, being able to improve the efficiency of the rainfall-induced mass landslide hazard assessment and effectively improve the accuracy of the rainfall-induced mass landslide hazard assessment.

[0073] In an exemplary embodiment, the obtaining the historical rainfall-induced mass landslide data of historical strong rainfall events and constructing landslide sample data may include the following steps:

[0074] Obtaining the historical rainfall-induced mass landslide data according to the satellite images of the historical strong rainfall events; through data verification processing and data supplementation processing of the historical rainfall-induced mass landslide data, obtaining landslide sample data containing the rainfall-induced mass landslide data distribution.

[0075] In practical applications, high-resolution satellite images within a preset time range can be obtained according to the occurrence date of historical heavy rainfall events. For example, images for one month before and after the occurrence date can be obtained. As Figure 2 shown, historical rainfall-induced landslide data can be extracted through visual interpretation by human-computer interaction. Furthermore, by verifying and supplementing high-resolution satellite images within the area of historical heavy rainfall events, an objective and detailed distribution of rainfall-induced landslide data can be obtained, and a rainfall-induced landslide database (i.e., landslide sample data) can be constructed.

[0076] Optionally, based on the visual interpretation method of human-computer interaction, the image features (such as tone, color, shape, size, shadow, texture, etc.) and spatial features (such as position and layout) of the image can be utilized, and combined with a variety of non-remote sensing information materials for comprehensive analysis and logical reasoning to extract data.

[0077] In an alternative embodiment, for the data preprocessing process of verification and supplementation, historical rainfall-induced landslide data can be collected, such as meteorological data including but not limited to rainfall amount, rainfall intensity, rainfall duration, temperature, etc., topographic and geomorphic data such as slope, aspect, altitude, etc., geological data such as soil type, rock property, vegetation cover, etc., and historical landslide information such as location, scale, occurrence time, etc. The historical rainfall-induced landslide data is subjected to data cleaning, such as removing duplicate, incorrect or abnormal data records, and data conversion, such as normalization processing, to facilitate further processing by machine learning algorithms.

[0078] In an exemplary embodiment, the analysis of the rainfall-induced landslide data distribution information based on the landslide attribute parameters and the landslide sample data to obtain the landslide development characteristics may include the following steps:

[0079] Based on the rainfall-induced landslide data distribution information of the landslide sample data, obtain the distribution pattern of rainfall-induced landslide data in the geographical space, determine the landslide density parameter and the landslide geometric parameter as the landslide attribute parameters; analyze the landslide mobility of the landslide sample data according to the landslide attribute parameters to obtain the landslide geographical characteristics and the landslide geometric characteristics as the landslide development characteristics.

[0080] Specifically, as Figure 2 shown, GIS (Geographic Information System) tools and statistical methods can be used to obtain the distribution pattern of landslide sample data in the geographical space, and it can be characterized based on the quantification indexes of landslide point density and landslide surface density. Furthermore, by statistically analyzing parameters such as landslide density, landslide length, landslide width, landslide height (i.e., landslide density parameters and landslide geometric parameters), and using the equivalent friction coefficient to analyze the landslide mobility, the landslide development law, that is, the landslide development characteristics, can be obtained.

[0081] In an exemplary embodiment, according to the landslide geographical features and the landslide geometric features, using the landslide sample data and a plurality of candidate landslide impact indicators to analyze the background law of landslide generation, and establishing a landslide hazard evaluation index system may include the following steps:

[0082] Obtain a plurality of candidate landslide impact indicators associated with the regional scope of the historical heavy rainfall event; combine the landslide geographical features and the landslide geometric features, analyze the landslide sample data and the plurality of candidate landslide impact indicators, and determine the landslide spatial distribution law and the landslide disaster-forming environment; according to the landslide spatial distribution law, the landslide disaster-forming environment, and the plurality of candidate landslide impact indicators, construct the landslide hazard evaluation index system.

[0083] In practical applications, as Figure 2 shown, by collecting data on internal and external impact factors of landslides, using methods such as correlation analysis and principal component analysis to screen out the impact factors most closely related to landslides, a plurality of candidate landslide impact indicators can be determined. Then, based on the spatial analysis function of the GIS tool, the landslide sample data and the plurality of candidate landslide impact indicators can be superimposed and analyzed to study the distribution law of rainfall-induced landslide disasters and the landslide disaster-forming environment. Furthermore, a landslide hazard evaluation index system containing a plurality of candidate landslide impact indicators can be constructed.

[0084] For example, the landslide hazard evaluation index system may include multiple impact factors such as 5-day cumulative rainfall, elevation, slope, aspect, terrain humidity index, normalized difference vegetation index, formation lithology, distance from water system, distance from road, and surface cover.

[0085] In an exemplary embodiment, by performing a correlation analysis on the plurality of candidate landslide impact indicators to determine the target landslide impact indicators, and combining the target landslide impact indicators and the landslide hazard evaluation index system, a landslide hazard assessment model constructed based on an automated machine learning framework may include the following steps:

[0086] Perform a correlation analysis on the plurality of candidate landslide impact indicators, and use the selected candidate landslide impact indicators as the target landslide impact indicators; in the landslide sample data, select landslide samples and non-landslide samples according to the target landslide impact indicators as training sample data; combine the target landslide impact indicators and the landslide hazard evaluation index system, and use an automated machine learning framework to construct an initial assessment model; train the initial assessment model based on the training sample data to obtain the landslide hazard assessment model.

[0087] In specific implementation, as Figure 2As shown, the Pearson correlation coefficient (an index used in statistics to measure the linear correlation degree between two variables, which can reflect the strength and direction of the linear relationship between the two variables) can be used to perform correlation analysis on multiple candidate landslide impact indicators to screen out the target landslide impact indicators. Then, landslide samples and non-landslide samples can be selected based on a 1:1 ratio to obtain training sample data. Furthermore, in combination with the target landslide impact indicators and the landslide hazard assessment index system, an initial evaluation model can be constructed using the automated machine learning framework AutoGluon, and the landslide hazard assessment model can be obtained through training based on the training sample data.

[0088] In one example, the automated machine learning framework can include multiple ensemble methods, such as random forest, gradient boosting tree, etc. Ensemble learning can improve the overall prediction performance by combining the prediction results of multiple base learners to construct an ensemble model. In the automated machine learning framework, one or more ensemble models can be selected for training. By adjusting the hyperparameters and optimizing the model structure, the prediction accuracy and generalization ability of the model can be improved, and its performance can be evaluated through methods such as cross-validation.

[0089] In an exemplary embodiment, the following steps can also be included:

[0090] Use a preset model evaluation index to detect the model performance of the landslide hazard assessment model; based on the landslide hazard assessment model that passes the detection, perform rainfall-induced landslide hazard assessment on the historical heavy rainfall events to obtain the landslide hazard assessment result; according to the landslide hazard assessment result, display the rainfall-induced landslide hazard prompt image of the historical heavy rainfall events.

[0091] Among them, the model evaluation index can include any one or more of the following: area under the curve index, accuracy index, precision index, recall index.

[0092] In practical applications, for the landslide hazard assessment model, AUC (area under the curve, the area under the ROC curve, a model evaluation index), accuracy, precision, and recall (i.e., the area under the curve index, accuracy index, precision index, recall index) can be used to detect the model performance; as Figure 2 shown, after establishing the landslide hazard assessment model, rainfall-induced landslide hazard evaluation can be performed on historical heavy rainfall events based on this landslide hazard assessment model.

[0093] In one example, GIS tools can be used to visually display the landslide hazard assessment results output by the model on a map, as Figure 2As shown, according to the results of the landslide hazard assessment, the natural breakpoint method can be used to divide the landslide hazard levels (such as extremely low, low, medium, high, and extremely high levels), and a rainfall-induced landslide hazard level map (i.e., a rainfall-induced landslide hazard warning image) can be generated for display, such as displaying the rainfall-induced landslide hazard warning image corresponding to the geographical area of historical heavy rainfall events.

[0094] In an exemplary embodiment, as Figure 3 shown, a flowchart of another rainfall-induced landslide hazard assessment method is provided. In this embodiment, the method includes the following steps:

[0095] In step 301, based on the satellite images of historical heavy rainfall events, historical rainfall-induced landslide data is obtained. By performing data verification processing and data supplementation processing on the historical rainfall-induced landslide data, landslide sample data containing the distribution of rainfall-induced landslide data is obtained. In step 302, based on the distribution information of the rainfall-induced landslide data in the landslide sample data, the distribution pattern of the rainfall-induced landslide data in the geographical space is obtained, and the landslide density parameter and the landslide geometric parameter are determined as the landslide attribute parameters. In step 303, according to the landslide attribute parameters, the landslide mobility of the landslide sample data is analyzed to obtain the landslide geographical characteristics and the landslide geometric characteristics as the landslide development characteristics. In step 304, a plurality of candidate landslide influence indicators associated with the regional scope of the historical heavy rainfall event are obtained. Combining the landslide geographical characteristics and the landslide geometric characteristics, the landslide sample data and the plurality of candidate landslide influence indicators are analyzed to determine the landslide spatial distribution law and the landslide disaster-forming environment. In step 305, according to the landslide spatial distribution law, the landslide disaster-forming environment, and the plurality of candidate landslide influence indicators, a landslide hazard evaluation index system is constructed. In step 306, through the correlation analysis of the plurality of candidate landslide influence indicators, the selected candidate landslide influence indicators are used as the target landslide influence indicators. In the landslide sample data, the landslide samples and non-landslide samples are selected according to the target landslide influence indicators as the training sample data. In step 307, combining the target landslide influence indicators and the landslide hazard evaluation index system, an initial evaluation model is constructed using an automated machine learning framework, and the initial evaluation model is trained based on the training sample data to obtain a landslide hazard assessment model. It should be noted that the specific limitations of the above steps can refer to the specific limitations of a rainfall-induced landslide hazard assessment method described above and will not be elaborated here.

[0096] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0097] Based on the same inventive concept, an embodiment of the present application also provides a rainfall-induced landslide hazard assessment device for implementing the above-mentioned rainfall-induced landslide hazard assessment method. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the rainfall-induced landslide hazard assessment device provided below can refer to the limitations on the rainfall-induced landslide hazard assessment method in the above text, and will not be repeated here.

[0098] In an exemplary embodiment, as Figure 4 shown, a rainfall-induced landslide hazard assessment device is provided, including:

[0099] A landslide sample data construction module 401, configured to obtain historical rainfall-induced landslide data of historical heavy rainfall events and construct landslide sample data;

[0100] A landslide development feature acquisition module 402, configured to analyze based on landslide attribute parameters and the rainfall-induced landslide data distribution information of the landslide sample data to obtain landslide development features; the landslide development features include landslide geographical features and landslide geometric state features;

[0101] A landslide evaluation system establishment module 403, configured to analyze the background rules of landslide generation according to the landslide geographical features and the landslide geometric features, and use the landslide sample data and multiple candidate landslide influence indicators to establish a landslide hazard evaluation index system;

[0102] A landslide hazard assessment model obtaining module 404, configured to determine target landslide influence indicators by performing correlation analysis on the multiple candidate landslide influence indicators, and combine the target landslide influence indicators and the landslide hazard evaluation index system to obtain a landslide hazard assessment model constructed based on an automated machine learning framework; the landslide hazard assessment model is used to evaluate the rainfall-induced landslide hazard of heavy rainfall events.

[0103] In one embodiment, the landslide sample data construction module 401 includes:

[0104] A data extraction sub-module, configured to obtain the historical rainfall-induced landslide data according to the satellite images of the historical heavy rainfall events;

[0105] A data processing sub-module, configured to obtain landslide sample data including the distribution of rainfall-induced landslide data by performing data verification processing and data supplementation processing on the historical rainfall-induced landslide data.

[0106] In one embodiment, the landslide development feature acquisition module 402 includes:

[0107] A landslide attribute parameter obtaining sub-module, configured to obtain the distribution pattern of rainfall-induced landslide data in the geographical space based on the distribution information of rainfall-induced landslide data in the landslide sample data, and determine the landslide density parameter and the landslide geometric parameter as the landslide attribute parameters;

[0108] A feature analysis sub-module, configured to analyze the landslide fluidity of the landslide sample data according to the landslide attribute parameters to obtain the landslide geographical features and the landslide geometric features as the landslide development features.

[0109] In one embodiment, the landslide evaluation system establishment module 403 includes:

[0110] An influence index acquisition sub-module, configured to acquire a plurality of candidate landslide influence indexes associated with the regional scope of the historical heavy rainfall events;

[0111] An index analysis sub-module, configured to analyze the landslide sample data and the plurality of candidate landslide influence indexes in combination with the landslide geographical features and the landslide geometric features to determine the landslide spatial distribution law and the landslide disaster-forming environment;

[0112] An index system construction sub-module, configured to construct the landslide hazard evaluation index system according to the landslide spatial distribution law, the landslide disaster-forming environment, and the plurality of candidate landslide influence indexes.

[0113] In one embodiment, the landslide hazard assessment model acquisition module 404 includes:

[0114] An index screening sub-module, configured to perform correlation analysis on the plurality of candidate landslide influence indexes, and use the screened candidate landslide influence indexes as the target landslide influence indexes;

[0115] A sample selection sub-module, configured to select landslide samples and non-landslide samples from the landslide sample data according to the target landslide influence indexes as training sample data;

[0116] A model construction sub-module, configured to construct an initial evaluation model by using an automated machine learning framework in combination with the target landslide impact index and the landslide hazard assessment index system;

[0117] A model training sub-module, configured to train the initial evaluation model based on the training sample data to obtain the landslide hazard assessment model.

[0118] In one embodiment, the device further includes:

[0119] A model performance detection module, configured to perform model performance detection on the landslide hazard assessment model by using a preset model evaluation index;

[0120] A landslide hazard assessment module, configured to perform rainfall-induced landslide hazard assessment on the historical heavy rainfall event based on the landslide hazard assessment model that passes the detection, to obtain a landslide hazard assessment result;

[0121] A landslide hazard prompt image display module, configured to display a rainfall-induced landslide hazard prompt image of the historical heavy rainfall event according to the landslide hazard assessment result;

[0122] Wherein, the model evaluation index includes any one or more of the following:

[0123] Area under the curve index, accuracy index, precision index, recall index.

[0124] Each module in the above rainfall-induced landslide hazard assessment device can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in the processor in the computer device in the form of hardware or be independent of it, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0125] In an exemplary embodiment, a computer device is provided. The computer device can be a terminal, and its internal structure diagram can be as Figure 5As shown in the figure. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit, and an input device. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface, the display unit, and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, near field communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a method for evaluating the risk of rainfall-induced mass landslides.

[0126] Those skilled in the art can understand that Figure 5 the structure shown in the figure is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0127] In an exemplary embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory. When the processor executes the computer program, the following steps are implemented:

[0128] Obtain the historical rainfall-induced mass landslide data of historical heavy rainfall events and construct landslide sample data;

[0129] Analyze based on the landslide attribute parameters and the rainfall-induced mass landslide data distribution information of the landslide sample data to obtain landslide development characteristics; the landslide development characteristics include landslide geographical characteristics and landslide geometric state characteristics;

[0130] According to the landslide geographical characteristics and the landslide geometric characteristics, use the landslide sample data and multiple candidate landslide impact indicators to analyze the background rules of landslide generation, and establish a landslide risk evaluation index system;

[0131] By performing a correlation analysis on the multiple candidate landslide impact indicators, determine the target landslide impact indicator, and combine the target landslide impact indicator and the landslide risk evaluation index system to obtain a landslide risk assessment model constructed based on an automated machine learning framework; the landslide risk assessment model is used to evaluate the risk of rainfall-induced mass landslides for heavy rainfall events.

[0132] In one embodiment, when the processor executes the computer program, it also implements the steps of the rainfall-induced mass landslide hazard assessment method in the above-mentioned other embodiments.

[0133] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0134] Obtain the historical rainfall-induced mass landslide data of historical heavy rainfall events and construct landslide sample data;

[0135] Analyze based on the landslide attribute parameters and the distribution information of the rainfall-induced mass landslide data of the landslide sample data to obtain landslide development characteristics; the landslide development characteristics include landslide geographical characteristics and landslide geometric state characteristics;

[0136] According to the landslide geographical characteristics and the landslide geometric characteristics, use the landslide sample data and multiple candidate landslide influence indicators to analyze the background rules of landslide generation, and establish a landslide hazard evaluation index system;

[0137] Through correlation analysis of the multiple candidate landslide influence indicators, determine the target landslide influence indicator, and combine the target landslide influence indicator and the landslide hazard evaluation index system to obtain a landslide hazard assessment model constructed based on an automated machine learning framework; the landslide hazard assessment model is used to evaluate the rainfall-induced mass landslide hazard of heavy rainfall events.

[0138] In one embodiment, when the computer program is executed by a processor, it also implements the steps of the rainfall-induced mass landslide hazard assessment method in the above-mentioned other embodiments.

[0139] In one embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the following steps are implemented:

[0140] Obtain the historical rainfall-induced mass landslide data of historical heavy rainfall events and construct landslide sample data;

[0141] Analyze based on the landslide attribute parameters and the distribution information of the rainfall-induced mass landslide data of the landslide sample data to obtain landslide development characteristics; the landslide development characteristics include landslide geographical characteristics and landslide geometric state characteristics;

[0142] According to the landslide geographical characteristics and the landslide geometric characteristics, use the landslide sample data and multiple candidate landslide influence indicators to analyze the background rules of landslide generation, and establish a landslide hazard evaluation index system;

[0143] By performing a correlation analysis on the multiple candidate landslide impact indicators, the target landslide impact indicator is determined. Combining the target landslide impact indicator and the landslide hazard evaluation index system, a landslide hazard assessment model constructed based on an automated machine learning framework is obtained; the landslide hazard assessment model is used to evaluate the landslide hazard of rainfall group events during heavy rainfall events.

[0144] In one embodiment, when the computer program is executed by the processor, it also implements the steps of the rainfall group landslide hazard assessment method in the above other embodiments.

[0145] It should be noted that the collection, use, and processing of relevant data involved in this application need to comply with relevant regulations.

[0146] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in this application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in this application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, artificial intelligence (AI) processors, etc., without limitation.

[0147] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this application.

[0148] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A method for assessing the hazard of landslides caused by rainfall clusters, characterized in that: The method comprises: Obtain historical rainfall cluster landslide data of historical heavy rainfall events and construct landslide sample data; Based on the distribution information of rainfall-induced landslide data of the landslide sample data, the distribution pattern of the rainfall-induced landslide data in geographic space is obtained, and the landslide density parameter and the landslide geometric parameter are determined as landslide attribute parameters. According to the landslide attribute parameters, the landslide fluidity of the landslide sample data is analyzed using the equivalent friction coefficient to obtain the landslide geographical characteristics and landslide geometric characteristics as landslide development characteristics; Acquire multiple candidate landslide impact indicators associated with the regional scope of the historical heavy rainfall event, analyze the landslide sample data and the multiple candidate landslide impact indicators in combination with the landslide geographical features and the landslide geometric features, determine the spatial distribution law of landslides and the landslide disaster-pregnant environment, and construct a landslide hazard evaluation index system based on the spatial distribution law of landslides and the landslide disaster-pregnant environment, as well as the multiple candidate landslide impact indicators; By performing correlation analysis on the multiple candidate landslide impact indicators, the target landslide impact indicator is determined, and a landslide hazard assessment model constructed based on an automatic machine learning framework is obtained by combining the target landslide impact indicator and the landslide hazard evaluation index system; the landslide hazard assessment model is used to conduct rainfall cluster landslide hazard assessment for heavy rainfall events.

2. The method according to claim 1, characterized in that The acquisition of historical rainfall group landslide data of historical heavy rainfall events and construction of landslide sample data includes: According to the satellite images of the historical heavy rainfall events, the historical rainfall cluster landslide data are obtained; By performing data verification and data supplementation processing on the historical rainfall-induced landslide data, landslide sample data including the distribution of rainfall-induced landslide data are obtained.

3. The method according to claim 1, characterized in that The method comprises: determining a target landslide impact index by performing correlation analysis on the plurality of candidate landslide impact indexes, and combining the target landslide impact index with the landslide hazard evaluation index system to obtain a landslide hazard evaluation model based on an automatic machine learning framework, including: By performing correlation analysis on the plurality of candidate landslide impact indicators, the selected candidate landslide impact indicators are used as the target landslide impact indicators; In the landslide sample data, landslide samples and non-landslide samples are selected according to the target landslide impact index as training sample data; Combining the target landslide impact index and the landslide hazard evaluation index system, an initial assessment model is constructed using an automatic machine learning framework; The initial assessment model is trained based on the training sample data to obtain the landslide hazard assessment model.

4. The method according to any one of claims 1 to 3, characterized in that: The method further comprises: Using preset model evaluation indicators to perform model performance testing on the landslide hazard assessment model; Based on the landslide hazard assessment model that has passed the test, a rainfall cluster landslide hazard assessment is performed on the historical heavy rainfall event to obtain a landslide hazard assessment result; According to the landslide hazard assessment result, a rainfall cluster landslide hazard warning image of the historical heavy rainfall event is displayed; The model evaluation indicators include any one or more of the following: Area under the curve indicator, accuracy indicator, precision indicator, and recall rate indicator.

5. A device for assessing the hazard of landslides caused by rainfall, characterized in that: The device comprises: The landslide sample data construction module is used to obtain the historical rainfall group landslide data of historical heavy rainfall events and construct the landslide sample data; A landslide development feature acquisition module is used to obtain the distribution pattern of rainfall-induced landslide data in geographical space based on the distribution information of rainfall-induced landslide data of the landslide sample data, determine landslide density parameters and landslide geometric parameters as landslide attribute parameters, and analyze the landslide fluidity of the landslide sample data using equivalent friction coefficient according to the landslide attribute parameters to obtain landslide geographical characteristics and landslide geometric characteristics as landslide development characteristics; A landslide evaluation system establishment module is used to obtain multiple candidate landslide impact indicators associated with the regional scope of the historical heavy rainfall event, analyze the landslide sample data and the multiple candidate landslide impact indicators in combination with the landslide geographical characteristics and the landslide geometric characteristics, determine the landslide spatial distribution law and the landslide disaster-pregnant environment, and construct a landslide hazard evaluation index system based on the landslide spatial distribution law and the landslide disaster-pregnant environment, as well as the multiple candidate landslide impact indicators; The landslide hazard assessment model obtaining module is used to determine the target landslide impact index by performing correlation analysis on the multiple candidate landslide impact indicators, and to obtain a landslide hazard assessment model constructed based on an automatic machine learning framework by combining the target landslide impact index and the landslide hazard evaluation index system; the landslide hazard assessment model is used to perform rainfall cluster landslide hazard assessment on heavy rainfall events.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images