Earthquake-triggered landslide risk detection method and device, computer equipment, readable storage medium and program product

Abstract

The invention relates to an earthquake-triggered landslide risk detection method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: respectively carrying out earthquake landslide risk detection on each target slope sample according to rock-soil characteristic data of each target slope sample to obtain an earthquake landslide risk index of each target slope sample; according to the earthquake landslide risk index of each target slope sample, negative samples are selected from each target slope sample to generate a target training sample set, and the earthquake landslide risk index of the negative samples is smaller than a preset risk index, and no earthquake landslide occurs; acquiring seismic landslide influence factor data corresponding to the target training sample set, and constructing a seismic landslide detection model by taking the seismic landslide influence factor data as a training data set; and according to the earthquake landslide detection model, earthquake-triggered landslide risk detection is carried out on a to-be-detected slope sample. By adopting the method, the accuracy of landslide risk detection can be improved.

Description

Technical Field

[0001] The present application relates to the field of big data technology, and in particular to a method, device, computer equipment, computer-readable storage medium and computer program product for detecting the hazard of earthquake-triggered landslides. Background Art

[0002] With the development of big data technology, a method of earthquake landslide risk detection based on remote sensing satellite images has emerged. Specifically, an earthquake landslide detection model is constructed using data samples extracted from remote sensing satellite images, and then the earthquake landslide detection model is used to perform earthquake landslide risk detection.

[0003] At present, areas where earthquake landslides have not occurred are usually randomly selected as non-landslide samples based on remote sensing satellite images, so that the non-landslide samples can be used as negative samples for model training.

[0004] However, since earthquake-landslide events are accidental events, even areas where earthquake landslides have not occurred may be samples with high earthquake landslide risks. Therefore, the method of randomly selecting areas where earthquake landslides have not occurred as non-landslide samples to construct earthquake landslide detection models will lead to lower accuracy of earthquake landslide detection models, thereby affecting the accuracy of earthquake-triggered landslide hazard detection. Summary of the invention

[0005] Based on this, it is necessary to provide an earthquake-triggered landslide hazard detection method, device, computer equipment, computer-readable storage medium and computer program product that can improve the accuracy of earthquake-triggered landslide hazard detection in response to the above technical problems.

[0006] In a first aspect, the present application provides a method for detecting the hazard of earthquake-triggered landslides, comprising:

[0007] According to the geotechnical property data of each target slope sample, respectively, an earthquake landslide risk detection is performed on each target slope sample to obtain an earthquake landslide risk index of each target slope sample;

[0008] According to the earthquake landslide risk index of each target slope sample, a negative sample is selected from each target slope sample, and a target training sample set corresponding to the negative sample is generated, wherein the negative sample is a sample whose earthquake landslide risk index is less than a preset risk index and no earthquake landslide occurs;

[0009] Acquire earthquake landslide impact factor data corresponding to the target training sample set, and use the earthquake landslide impact factor data as a training data set to construct an earthquake landslide detection model;

[0010] According to the earthquake landslide detection model, earthquake-triggered landslide hazard detection is performed on the slope sample to be detected.

[0011] In one embodiment, the preset risk index includes a first preset risk index; the selecting a negative sample from each target slope sample according to the earthquake landslide risk index of each target slope sample, and generating a target training sample set corresponding to the negative sample, includes:

[0012] Selecting a plurality of first preselected slope samples whose earthquake landslide risk index is less than a first preset risk index from each of the target slope samples;

[0013] Selecting samples without earthquake landslide from the plurality of first preselected slope samples as negative samples;

[0014] From each of the target slope samples, samples in which earthquake landslides have occurred are selected as positive samples, and all the selected negative samples and positive samples are taken together as a target training sample set.

[0015] In one embodiment, the preset risk index includes a first preset risk index and a second preset risk index, and the second preset risk index is greater than the first preset risk index; the selecting a negative sample from each target slope sample according to the earthquake landslide risk index of each target slope sample, and generating a target training sample set corresponding to the negative sample, comprises:

[0016] Selecting a plurality of second preselected slope samples whose earthquake landslide risk index is less than a second preset risk index from each of the target slope samples;

[0017] Taking slope samples whose earthquake landslide risk index is less than the first preset risk index among the plurality of second preselected slope samples as a first interval sample set, and taking slope samples other than the first interval sample set among the plurality of second preselected slope samples as a second interval sample set;

[0018] Collecting samples without earthquake landslide from the first interval sample set and the second interval sample set as negative samples;

[0019] From each of the target slope samples, samples in which earthquake landslides have occurred are selected as positive samples, and all the selected negative samples and positive samples are taken together as a target training sample set.

[0020] In one embodiment, collecting samples without earthquake landslide from the first interval sample set and the second interval sample set as negative samples includes:

[0021] Obtaining a first negative sample collection ratio of the first interval sample set and a second negative sample collection ratio of the second interval sample set, wherein the first negative sample collection ratio is greater than the second negative sample collection ratio;

[0022] According to the first negative sample collection ratio and the second negative sample collection ratio, samples in which no earthquake landslide occurred are collected in proportion from the first interval sample set and the second interval sample set as negative samples.

[0023] In one embodiment, the earthquake landslide risk detection is performed on each target slope sample according to the geotechnical characteristic data of each target slope sample to obtain the earthquake landslide risk index of each target slope sample, including:

[0024] According to the geotechnical characteristic data of the target slope sample, the static safety factor corresponding to the target slope sample is detected, and according to the static safety factor, the critical acceleration of the earthquake landslide of the target slope sample is detected;

[0025] According to the critical acceleration, the cumulative displacement of the target slope sample causing an earthquake landslide is detected, and according to the cumulative displacement, the probability of an earthquake landslide occurring at the location of the target slope sample is detected to obtain an earthquake landslide risk index for the target slope sample, wherein the earthquake landslide risk index is used to characterize the probability of an earthquake-triggered landslide.

[0026] In one embodiment, the obtaining of earthquake landslide impact factor data corresponding to the target training sample set includes:

[0027] Acquire a slope pixel image corresponding to the target training sample set, and extract a plurality of slope image features from the slope pixel image by performing image detection on the slope pixel image;

[0028] Respectively detecting variance inflation factors and feature importances of the plurality of slope image features, wherein the variance inflation factor is used to characterize the multicollinearity between the plurality of slope image features, and the feature importance is used to characterize the importance of the slope image features for earthquake-triggered landslide hazard detection;

[0029] According to the variance inflation factors and feature importances of the plurality of slope image features, earthquake landslide influencing factor data are extracted from the plurality of slope image features.

[0030] In a second aspect, the present application also provides an earthquake-triggered landslide hazard detection device, comprising:

[0031] A first detection module is used to perform earthquake landslide risk detection on each target slope sample according to the geotechnical characteristic data of each target slope sample, and obtain an earthquake landslide risk index of each target slope sample;

[0032] A sample selection module, for selecting negative samples from each of the target slope samples according to the earthquake landslide risk index of each of the target slope samples, and generating a target training sample set corresponding to the negative samples, wherein the negative samples are samples whose earthquake landslide risk index is less than a preset risk index and where no earthquake landslide occurs;

[0033] A model building module, used to obtain earthquake landslide impact factor data corresponding to the target training sample set, and use the earthquake landslide impact factor data as a training data set to build an earthquake landslide detection model;

[0034] The second detection module is used to perform earthquake-triggered landslide hazard detection on the slope sample to be detected according to the earthquake landslide detection model.

[0035] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0036] According to the geotechnical property data of each target slope sample, earthquake landslide risk detection is performed on each target slope sample to obtain the earthquake landslide risk index of each target slope sample; according to the earthquake landslide risk index of each target slope sample, negative samples are selected from each target slope sample, and a target training sample set corresponding to the negative sample is generated, wherein the negative sample is a sample whose earthquake landslide risk index is less than a preset risk index and no earthquake landslide has occurred; earthquake landslide impact factor data corresponding to the target training sample set is obtained, and an earthquake landslide detection model is constructed using the earthquake landslide impact factor data as a training data set; according to the earthquake landslide detection model, earthquake-triggered landslide hazard detection is performed on the slope samples to be detected.

[0037] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0038] According to the geotechnical property data of each target slope sample, earthquake landslide risk detection is performed on each target slope sample to obtain the earthquake landslide risk index of each target slope sample; according to the earthquake landslide risk index of each target slope sample, negative samples are selected from each target slope sample, and a target training sample set corresponding to the negative sample is generated, wherein the negative sample is a sample whose earthquake landslide risk index is less than a preset risk index and no earthquake landslide has occurred; earthquake landslide impact factor data corresponding to the target training sample set is obtained, and an earthquake landslide detection model is constructed using the earthquake landslide impact factor data as a training data set; according to the earthquake landslide detection model, earthquake-triggered landslide hazard detection is performed on the slope samples to be detected.

[0039] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0040] According to the geotechnical property data of each target slope sample, earthquake landslide risk detection is performed on each target slope sample to obtain the earthquake landslide risk index of each target slope sample; according to the earthquake landslide risk index of each target slope sample, negative samples are selected from each target slope sample, and a target training sample set corresponding to the negative sample is generated, wherein the negative sample is a sample whose earthquake landslide risk index is less than a preset risk index and no earthquake landslide has occurred; earthquake landslide impact factor data corresponding to the target training sample set is obtained, and an earthquake landslide detection model is constructed using the earthquake landslide impact factor data as a training data set; according to the earthquake landslide detection model, earthquake-triggered landslide hazard detection is performed on the slope samples to be detected.

[0041] The above-mentioned earthquake-triggered landslide hazard detection method, device, computer equipment, computer-readable storage medium and computer program product first obtain geotechnical property data of multiple target slope samples, and then perform earthquake landslide risk detection on each target slope sample according to the geotechnical property data to obtain the earthquake landslide risk index of each target slope sample, so that according to each earthquake landslide risk index, non-landslide samples with lower landslide risk can be preliminarily identified in each target slope sample, that is, non-landslide samples with earthquake landslide risk index less than a preset risk index and without earthquake landslide can be selected as negative samples, avoiding the selection of target slope samples with high landslide risk as negative samples for model construction, improving the selection accuracy of target training sample sets, and then obtaining earthquake landslide impact factor data of target training sample sets with higher accuracy, using the earthquake landslide impact factor data as a training data set to construct an earthquake landslide detection model, which can improve the detection accuracy of the earthquake landslide detection model, and using the earthquake landslide detection model to perform earthquake-triggered landslide hazard detection on the slope samples to be detected, which can improve the accuracy of earthquake-triggered landslide hazard detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0043] Figure 1 A schematic flow chart of a method for detecting landslide hazard triggered by an earthquake in one embodiment of the present application;

[0044] Figure 2 A schematic diagram of a process for selecting a target training sample set in one embodiment of the present application;

[0045] Figure 3 This is a schematic diagram of the process of extracting earthquake landslide impact factor data in one embodiment of the present application;

[0046] Figure 4 This is a structural block diagram of an earthquake-triggered landslide hazard detection device in one embodiment of the present application;

[0047] Figure 5 This is a diagram of the internal structure of a computer device in one embodiment of the present application. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying 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.

[0049] In an exemplary embodiment, Figure 1 As shown, a method for detecting the danger of earthquake-triggered landslides is provided, comprising the following steps 202 to 208. Among them:

[0050] Step 202: Perform earthquake landslide risk detection on each target slope sample according to the geotechnical characteristic data of each target slope sample to obtain an earthquake landslide risk index of each target slope sample.

[0051] Among them, in this embodiment, the target slope sample can be extracted from a remote sensing satellite image or an on-site shot image. For example, the target slope sample can be an image pixel in the remote sensing satellite image, and the image pixel can be a grid unit in a grid map, or an image area with a preset range size; the geotechnical property data is data characterizing the geotechnical physical properties of the actual location area of ​​the target slope sample, and the geotechnical property data can be parameters such as the effective cohesion of the slope area, the weight of the rock and soil, the thickness of the rock and soil body, the slope angle, the degree of saturation of the rock and soil body, the weight of groundwater, the effective internal friction angle, and the peak acceleration of the seismic motion.

[0052] Earthquake landslide risk detection based on geotechnical property data can detect the probability of earthquake-triggered landslide in the target slope sample. This probability can be characterized by an earthquake landslide risk index. The higher the earthquake landslide risk index, the higher the probability of earthquake-triggered landslide in the target slope sample.

[0053] As an example, based on the geotechnical property data, the Newmark model can be used to detect the earthquake landslide risk of each target slope sample.

[0054] As an example, according to the geotechnical property data of each target slope sample, each target slope sample is respectively tested for earthquake landslide risk, and the earthquake landslide risk index of each target slope sample is obtained, including:

[0055] According to the geotechnical property data of the target slope sample, the static safety factor corresponding to the target slope sample is detected, and based on the static safety factor, the critical acceleration of the target slope sample for earthquake landslide is detected; based on the critical acceleration, the cumulative displacement of the target slope sample for earthquake landslide is detected, and based on the cumulative displacement, the probability of earthquake landslide occurring at the location of the target slope sample is detected to obtain the earthquake landslide risk index of the target slope sample, wherein the earthquake landslide risk index is used to characterize the probability of earthquake-triggered landslide.

[0056] The calculation process of the above static safety factor is as follows:

[0057]

[0058] in, is the static safety factor, For effective cohesion, is the mass of rock and soil, is the thickness of the rock mass, is the slope angle, The degree of saturation of rock and soil, is the groundwater density, is the effective internal friction angle.

[0059] The above critical acceleration may be the acceleration at which the slope is about to slide, and the calculation process of the critical acceleration is as follows:

[0060]

[0061] in, is the critical acceleration, is the static safety factor, is the acceleration due to gravity, is the slope angle.

[0062] The above cumulative displacement can be calculated using an empirical formula based on global earthquake record fitting, as follows:

[0063] log D n = 0 . 215 + log [ ( 1 − a c PGA ) 2 . 341 * ( a c PGA ) − 1 . 438 ]

[0064] in, is the cumulative displacement, is the critical acceleration, is the peak acceleration of the earthquake.

[0065] Furthermore, based on the static safety factor, critical acceleration and cumulative displacement, the probability of earthquake landslide occurring at the location of the target slope sample can be calculated using the probability formula of earthquake-induced landslide, which is as follows:

[0066] P = 0 . 335 [ 1 − exp ( − 0 . 048 D n 1 . 565 ) ]

[0067] in, is the probability of earthquake-induced slope landslide, that is, the earthquake landslide risk index, is the cumulative displacement.

[0068] Step 204, according to the earthquake landslide risk index of each target slope sample, a negative sample is selected from each target slope sample, and a target training sample set corresponding to the negative sample is generated. The negative sample is a sample whose earthquake landslide risk index is less than a preset risk index and no earthquake landslide has occurred.

[0069] Among them, the target training sample set consists of positive samples and negative samples. The positive sample is a target slope sample where an earthquake landslide has occurred. The target slope sample can be determined by comparing satellite remote sensing images before and after the earthquake. The negative sample is a target slope sample whose earthquake landslide risk index is less than a preset risk index and where no earthquake landslide has occurred. Among them, if the earthquake landslide risk index of the target slope sample is less than the preset risk index, it is considered that the risk of an earthquake-triggered landslide in the target slope sample is low (the probability of occurrence is low), and the target slope sample is suitable for selection as a negative sample. If the earthquake landslide risk index of the target slope sample is not less than the preset risk index, it is considered that the risk of an earthquake-triggered landslide in the target slope sample is high (the probability of occurrence is high), and it is not suitable as a negative sample.

[0070] As an example, step 204 includes: selecting samples whose earthquake landslide risk index is less than a preset risk index and in which no earthquake landslide has occurred as negative samples from each target slope sample, and selecting positive samples of samples in which earthquake landslide has occurred from each target slope sample; and using the positive samples and negative samples together as the target training sample set.

[0071] As an example, step 204 includes: according to each earthquake landslide risk index, each target slope sample is divided into an extremely low risk sample interval, a low risk sample interval, a medium risk sample interval, a high risk sample interval and an extremely high risk sample interval, wherein the earthquake landslide risk index of the target slope samples in the five intervals of the extremely low risk sample interval, the low risk sample interval, the medium risk sample interval, the high risk sample interval and the extremely high risk sample interval increases in sequence; in the extremely low risk sample interval, samples where no earthquake landslide has occurred are selected as negative samples, and in each target slope sample, samples where an earthquake landslide has occurred are selected as positive samples; and the positive samples and the negative samples are used together as the target training sample set.

[0072] As an example, the earthquake landslide risk index corresponding to the extremely low risk sample interval can range from 0 to 0.2, the earthquake landslide risk index corresponding to the low risk sample interval can range from 0.2 to 0.4, the earthquake landslide risk index corresponding to the medium risk sample interval can range from 0.4 to 0.6, the earthquake landslide risk index corresponding to the high risk sample interval can range from 0.6 to 0.8, and the earthquake landslide risk index corresponding to the extremely high risk sample interval can range from 0.8 to 1.0.

[0073] As an example, the preset risk index includes a first preset risk index; according to the earthquake landslide risk index of each target slope sample, a negative sample is selected from each target slope sample, and a target training sample set corresponding to the negative sample is generated, including:

[0074] A plurality of first pre-selected slope samples whose earthquake landslide risk index is less than a first preset risk index are selected from each target slope sample; samples where no earthquake landslide has occurred are selected from the plurality of first pre-selected slope samples as negative samples; samples where earthquake landslide has occurred are selected from each target slope sample as positive samples, and all the selected negative samples and positive samples are taken together as a target training sample set.

[0075] Specifically, a target earthquake landslide risk index that is less than a first preset risk index is screened out from each earthquake landslide risk index, and each target slope sample corresponding to each target earthquake landslide risk index is used as a first pre-selected slope sample, and samples in which no earthquake landslide has occurred are selected from multiple first pre-selected slope samples as negative samples; samples in which earthquake landslide has occurred are selected from each target slope sample as positive samples, and all selected negative samples and positive samples are combined together into a target training sample set.

[0076] Step 206, obtaining earthquake landslide impact factor data of the target training sample set, and using the earthquake landslide impact factor data as the training data set to construct an earthquake landslide detection model.

[0077] Among them, the slope pixel image can be a slope pixel image in a remote sensing satellite image before and / or after an earthquake occurs, the earthquake landslide influencing factor data includes at least one earthquake landslide influencing factor, and the earthquake landslide influencing factor can be an image feature extracted from the slope pixel image, and the image feature is helpful to construct an earthquake landslide detection model; the above-mentioned earthquake landslide influencing factor can be the earthquake intensity, elevation, slope, slope direction, plane curvature, slope position, terrain undulation, terrain humidity index, land use type, vegetation coverage, stratum lithology and distance from active faults at the slope location; and in order to enrich the features for constructing the earthquake landslide detection model, in addition to the image features, rainfall can also be used as one of the earthquake landslide influencing factors.

[0078] It should be noted that the above earthquake landslide detection model can be a random forest model.

[0079] Step 208: Perform earthquake-triggered landslide hazard detection on the slope sample to be detected according to the earthquake landslide detection model.

[0080] As an example, step 208 includes: obtaining a target pixel image corresponding to the slope sample to be detected, and extracting target earthquake landslide impact factor data from the target pixel image; and detecting the danger of earthquake-triggered landslides in the slope sample to be detected by inputting the target earthquake landslide impact factor data into an earthquake landslide detection model.

[0081] In the above earthquake-triggered landslide hazard detection method, geotechnical property data of multiple target slope samples are first obtained, and then, based on the geotechnical property data, earthquake landslide risk detection is performed on each target slope sample respectively to obtain an earthquake landslide risk index for each target slope sample. Therefore, according to each earthquake landslide risk index, non-landslide samples with lower landslide risk can be preliminarily identified in each target slope sample, that is, non-landslide samples with earthquake landslide risk index less than a preset risk index and without earthquake landslide can be selected as negative samples, thereby avoiding selecting target slope samples with high landslide risk as negative samples for model construction, improving the selection accuracy of the target training sample set, and then obtaining earthquake landslide impact factor data of the target training sample set with higher accuracy. Using the earthquake landslide impact factor data as the training data set, an earthquake landslide detection model is constructed, which can improve the detection accuracy of the earthquake landslide detection model. The earthquake landslide detection model is used to perform earthquake-triggered landslide hazard detection on the slope samples to be detected, which can improve the accuracy of earthquake-triggered landslide hazard detection.

[0082] In an exemplary embodiment, Figure 2As shown, the preset risk index includes a first preset risk index and a second preset risk index, and the second preset risk index is greater than the first preset risk index; according to the earthquake landslide risk index of each target slope sample, a negative sample is selected from each target slope sample, and a target training sample set corresponding to the negative sample is generated, including:

[0083] Step 302: Select a plurality of second pre-selected slope samples whose earthquake landslide risk index is less than a second preset risk index from each target slope sample.

[0084] Step 304: taking slope samples whose earthquake landslide risk index is less than the first preset risk index among the plurality of second preselected slope samples as a first interval sample set, and taking slope samples other than the first interval sample set among the plurality of second preselected slope samples as a second interval sample set.

[0085] Among them, the second preset risk index is greater than the first preset risk index. Therefore, the earthquake landslide risk index corresponding to the slope samples in the first interval sample set is smaller than the earthquake landslide risk index corresponding to the slope samples in the first interval sample set. Therefore, the landslide risk of the slope samples in the first interval sample set is lower than the landslide risk of the slope samples in the second interval sample set.

[0086] Step 306: Collect samples without earthquake landslide from the first interval sample set and the second interval sample set as negative samples.

[0087] Step 308: Select samples of earthquake landslides from target slope samples as positive samples, and use all selected negative samples and positive samples as target training sample sets.

[0088] As an example, steps 306 to 308 include: collecting a first preset number of slope samples in which no earthquake landslide has occurred from a first interval sample set as negative samples, and collecting a second preset number of slope samples in which no earthquake landslide has occurred from a second interval sample set as negative samples, wherein the first preset number is greater than the second preset number; selecting samples in which earthquake landslide has occurred from each target slope sample as positive samples, and taking all selected negative samples and positive samples together as a target training sample set.

[0089] As an example, samples without earthquake landslide are collected from the first interval sample set and the second interval sample set as negative samples, including:

[0090] A first negative sample collection ratio of the first interval sample set and a second negative sample collection ratio of the second interval sample set are obtained, wherein the first negative sample collection ratio is greater than the second negative sample collection ratio; and samples in which no earthquake landslides have occurred are proportionally collected from the first interval sample set and the second interval sample set according to the first negative sample collection ratio and the second negative sample collection ratio as negative samples.

[0091] Specifically, a first negative sample collection ratio of the first interval sample set and a second negative sample collection ratio of the second interval sample set are obtained, wherein the first negative sample collection ratio is greater than the second negative sample collection ratio; samples in which no earthquake landslides occurred are collected in the first interval sample set with the first negative sample collection ratio as negative samples, and samples in which no earthquake landslides occurred are collected in the second interval sample set with the second negative sample collection ratio as negative samples.

[0092] In this embodiment, firstly, a plurality of second pre-selected slope samples with lower earthquake landslide risk index are selected from each target slope sample through the second preset risk index, and then the plurality of second pre-selected slope samples are subdivided into a first interval sample set and a second interval sample set according to the first preset risk index which is smaller than the second preset risk index, and then a larger number of negative samples are selected from the first interval sample set and a small number of negative samples are selected from the second interval sample set. In this way, on the one hand, most of the negative samples can be selected from the extremely low risk samples (samples in the first interval sample set) to ensure the accuracy of the selection of negative samples, which is helpful for the accuracy of the earthquake landslide detection model. On the other hand, a small number of negative samples can be selected from the low risk samples (samples in the second interval sample set). In this way, the richness of negative samples can be ensured, and the local convergence of the earthquake landslide detection model due to the construction of the earthquake landslide detection model only based on the extremely low risk samples can be prevented, which is helpful for further improving the accuracy of the earthquake landslide detection model.

[0093] In an exemplary embodiment, Figure 3 As shown, the earthquake landslide impact factor data corresponding to the target training sample set is obtained, including:

[0094] Step 402: Obtain a slope pixel image corresponding to the target training sample set, and extract a plurality of slope image features from the slope pixel image by performing image detection on the slope pixel image.

[0095] Step 404, respectively detect the variance inflation factor and feature importance of multiple slope image features, wherein the variance inflation factor is used to characterize the multicollinearity between multiple slope image features, and the feature importance is used to characterize the importance of slope image features for earthquake-triggered landslide hazard detection.

[0096] Step 406: extract earthquake landslide influencing factor data from the plurality of slope image features according to the variance inflation factors and feature importances of the plurality of slope image features.

[0097] Steps 402 to 406 include: respectively detecting the variance inflation factor and feature importance of multiple slope image features, wherein the variance inflation factor is used to characterize the multicollinearity between multiple slope image features, and the feature importance is used to characterize the importance of slope image features for earthquake-triggered landslide hazard detection; and selecting slope image features whose variance inflation factor is less than a preset inflation factor threshold and whose feature importance is greater than a preset importance threshold as earthquake landslide influencing factor data.

[0098] As an example, the preset expansion factor threshold may be set to 10, and the preset importance threshold may be set to 0; the preset expansion factor threshold may be set to 11, and the preset importance threshold may be set to 0.1.

[0099] The calculation formula of variance inflation factor is as follows:

[0100]

[0101] in, For the The determination coefficient between the earthquake-landslide influencing factor and other earthquake-landslide influencing factors is For the Variance inflation factor of the earthquake-landslide influencing factor.

[0102] The landslide detection module may be a random forest model, and the feature importance may be the Gini importance corresponding to the random forest model.

[0103] In the above embodiment, the slope pixel image corresponding to the target training sample set is first obtained, and image detection is performed on the slope pixel image to extract multiple slope image features from the slope pixel image; then, based on the variance expansion factor and feature importance of the multiple slope image features, image features that are more suitable for model construction are selected from the multiple slope image features as earthquake landslide influencing factor data, which helps to improve the accuracy of the constructed earthquake landslide detection model and helps to improve the accuracy of earthquake-triggered landslide hazard detection.

[0104] In a complete embodiment, first, the static safety factor corresponding to the target slope sample is detected based on the geotechnical property data of the target slope sample, and based on the static safety factor, the critical acceleration of the target slope sample for an earthquake landslide is detected; based on the critical acceleration, the cumulative displacement of the target slope sample for an earthquake landslide is detected, and based on the cumulative displacement, the probability of an earthquake landslide occurring at the location of the target slope sample is detected to obtain an earthquake landslide risk index for the target slope sample, wherein the earthquake landslide risk index is used to characterize the probability of an earthquake-triggered landslide.

[0105] Furthermore, the preset risk index includes a first preset risk index; a target earthquake landslide risk index that is less than the first preset risk index is screened from each earthquake landslide risk index, and each target slope sample corresponding to each target earthquake landslide risk index is used as a first pre-selected slope sample, and samples in which no earthquake landslide has occurred are selected from multiple first pre-selected slope samples as negative samples; samples in which earthquake landslide has occurred are selected from each target slope sample as positive samples, and all selected negative samples and positive samples are combined together into a target training sample set.

[0106] Furthermore, a slope pixel image corresponding to a target training sample set is obtained, and multiple slope image features are extracted from the slope pixel image by performing image detection on the slope pixel image; variance inflation factors and feature importances of multiple slope image features are detected respectively, wherein the variance inflation factor is used to characterize the multicollinearity between multiple slope image features, and the feature importance is used to characterize the importance of slope image features for earthquake-triggered landslide hazard detection; slope image features whose variance inflation factor is less than a preset inflation factor threshold and whose feature importance is greater than a preset importance threshold are selected as earthquake landslide influencing factor data; a random forest module is constructed as an earthquake landslide detection model using the earthquake landslide influencing factor data as a training data set; a target pixel image corresponding to a slope sample to be detected is obtained, and target earthquake landslide influencing factor data is extracted from the target pixel image; and the danger level of an earthquake landslide occurring in the slope sample to be detected is detected by inputting the target earthquake landslide influencing factor data into the earthquake landslide detection model.

[0107] In this embodiment, non-landslide samples whose earthquake landslide risk index is less than a preset risk index and in which earthquake landslides occur are selected as negative samples, thereby avoiding selecting target slope samples with high earthquake landslide risks as negative samples for model construction, thereby improving the selection accuracy of the target training sample set, and then extracting earthquake landslide influencing factor data from the slope pixel image corresponding to the target training sample set with higher selection accuracy. Using the earthquake landslide influencing factor data as a training data set, an earthquake landslide detection model is constructed, which can improve the detection accuracy of the earthquake landslide detection model. Using the earthquake landslide detection model to perform earthquake-triggered landslide hazard detection on the slope samples to be detected can improve the accuracy of earthquake-triggered landslide hazard detection.

[0108] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0109] Based on the same inventive concept, the embodiment of the present application also provides an earthquake-triggered landslide hazard detection device for implementing the earthquake-triggered landslide hazard detection method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more earthquake-triggered landslide hazard detection device embodiments provided below can refer to the limitations of the earthquake-triggered landslide hazard detection method above, and will not be repeated here.

[0110] In an exemplary embodiment, Figure 4 As shown, a device for detecting landslide hazard triggered by earthquake is provided, comprising: a first detection module 502, a sample selection module 504, a model building module 506 and a second detection module 508, wherein:

[0111] The first detection module 502 is used to perform earthquake landslide risk detection on each target slope sample according to the geotechnical characteristic data of each target slope sample, and obtain the earthquake landslide risk index of each target slope sample;

[0112] The sample selection module 504 is used to select negative samples from each target slope sample according to the earthquake landslide risk index of each target slope sample, and generate a target training sample set corresponding to the negative sample, wherein the negative sample is a sample whose earthquake landslide risk index is less than a preset risk index and no earthquake landslide occurs;

[0113] A model building module 506 is used to obtain earthquake landslide impact factor data corresponding to the target training sample set, and use the earthquake landslide impact factor data as a training data set to build an earthquake landslide detection model;

[0114] The second detection module 508 is used to perform earthquake-triggered landslide hazard detection on the slope sample to be detected according to the earthquake landslide detection model.

[0115] In one embodiment, the preset risk index includes a first preset risk index; the sample selection module 504 is further used to:

[0116] A plurality of first preselected slope samples whose earthquake landslide risk index is less than a first preset risk index are selected from each of the target slope samples; samples where no earthquake landslide has occurred are selected from the plurality of first preselected slope samples as negative samples; samples where earthquake landslide has occurred are selected from each of the target slope samples as positive samples, and all the selected negative samples and positive samples are used together as a target training sample set.

[0117] In one embodiment, the preset risk index includes a first preset risk index and a second preset risk index, and the second preset risk index is greater than the first preset risk index; the sample selection module 504 is further used to:

[0118] A plurality of second pre-selected slope samples whose earthquake landslide risk index is less than a second preset risk index are selected from each of the target slope samples; slope samples whose earthquake landslide risk index is less than the first preset risk index in the plurality of second pre-selected slope samples are used as a first interval sample set, and slope samples other than the first interval sample set in the plurality of second pre-selected slope samples are used as a second interval sample set; samples where no earthquake landslide has occurred are collected from the first interval sample set and the second interval sample set as negative samples; samples where earthquake landslide has occurred are selected from each of the target slope samples as positive samples, and all the selected negative samples and positive samples are used together as a target training sample set.

[0119] In one embodiment, the sample selection module 504 is further used to:

[0120] Obtain a first negative sample collection ratio of the first interval sample set and a second negative sample collection ratio of the second interval sample set, wherein the first negative sample collection ratio is greater than the second negative sample collection ratio; and according to the first negative sample collection ratio and the second negative sample collection ratio, collect samples in which no earthquake landslides have occurred from the first interval sample set and the second interval sample set in proportion as negative samples.

[0121] In one embodiment, the first detection module 502 is further configured to:

[0122] According to the geotechnical property data of the target slope sample, the static safety factor corresponding to the target slope sample is detected, and based on the static safety factor, the critical acceleration of the target slope sample for earthquake landslide is detected; according to the critical acceleration, the cumulative displacement of the target slope sample for earthquake landslide is detected, and based on the cumulative displacement, the probability of earthquake landslide occurring at the location of the target slope sample is detected to obtain the earthquake landslide risk index of the target slope sample, wherein the earthquake landslide risk index is used to characterize the probability of earthquake-triggered landslide.

[0123] In one embodiment, the model building module 506 is further configured to:

[0124] Obtain a slope pixel image corresponding to the target training sample set, and extract a plurality of slope image features from the slope pixel image by performing image detection on the slope pixel image; respectively detect variance inflation factors and feature importances of the plurality of slope image features, wherein the variance inflation factor is used to characterize the multicollinearity between the plurality of slope image features, and the feature importance is used to characterize the importance of the slope image features for earthquake-triggered landslide hazard detection; extract earthquake landslide influencing factor data from the plurality of slope image features according to the variance inflation factors and feature importances of the plurality of slope image features.

[0125] Each module in the above-mentioned earthquake-triggered landslide hazard detection device can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each of the above modules.

[0126] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 5As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. 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, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store landslide risk detection data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for detecting the hazard of earthquake-triggered landslides is implemented.

[0127] Those skilled in the art will understand that Figure 5 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0128] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0129] According to the geotechnical property data of each target slope sample, earthquake landslide risk detection is performed on each target slope sample to obtain the earthquake landslide risk index of each target slope sample; according to the earthquake landslide risk index of each target slope sample, negative samples are selected from each target slope sample, and a target training sample set corresponding to the negative sample is generated, wherein the negative sample is a sample whose earthquake landslide risk index is less than a preset risk index and no earthquake landslide has occurred; earthquake landslide impact factor data corresponding to the target training sample set is obtained, and an earthquake landslide detection model is constructed using the earthquake landslide impact factor data as a training data set; according to the earthquake landslide detection model, earthquake-triggered landslide hazard detection is performed on the slope samples to be detected.

[0130] In one embodiment, the preset risk index includes a first preset risk index; when the processor executes the computer program, the following steps are also implemented:

[0131] A plurality of first preselected slope samples whose earthquake landslide risk index is less than a first preset risk index are selected from each of the target slope samples; samples where no earthquake landslide has occurred are selected from the plurality of first preselected slope samples as negative samples; samples where earthquake landslide has occurred are selected from each of the target slope samples as positive samples, and all the selected negative samples and positive samples are used together as a target training sample set.

[0132] In one embodiment, the preset risk index includes a first preset risk index and a second preset risk index, and the second preset risk index is greater than the first preset risk index; when the processor executes the computer program, the following steps are also implemented:

[0133] A plurality of second pre-selected slope samples whose earthquake landslide risk index is less than a second preset risk index are selected from each of the target slope samples; slope samples whose earthquake landslide risk index is less than the first preset risk index in the plurality of second pre-selected slope samples are used as a first interval sample set, and slope samples other than the first interval sample set in the plurality of second pre-selected slope samples are used as a second interval sample set; samples where no earthquake landslide has occurred are collected from the first interval sample set and the second interval sample set as negative samples; samples where earthquake landslide has occurred are selected from each of the target slope samples as positive samples, and all the selected negative samples and positive samples are used together as a target training sample set.

[0134] In one embodiment, when the processor executes the computer program, the following steps are also implemented:

[0135] Obtain a first negative sample collection ratio of the first interval sample set and a second negative sample collection ratio of the second interval sample set, wherein the first negative sample collection ratio is greater than the second negative sample collection ratio; and according to the first negative sample collection ratio and the second negative sample collection ratio, collect samples in which no earthquake landslides have occurred from the first interval sample set and the second interval sample set in proportion as negative samples.

[0136] In one embodiment, when the processor executes the computer program, the following steps are also implemented:

[0137] According to the geotechnical property data of the target slope sample, the static safety factor corresponding to the target slope sample is detected, and based on the static safety factor, the critical acceleration of the target slope sample for earthquake landslide is detected; according to the critical acceleration, the cumulative displacement of the target slope sample for earthquake landslide is detected, and based on the cumulative displacement, the probability of earthquake landslide occurring at the location of the target slope sample is detected to obtain the earthquake landslide risk index of the target slope sample, wherein the earthquake landslide risk index is used to characterize the probability of earthquake-triggered landslide.

[0138] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0139] Obtain a slope pixel image corresponding to the target training sample set, and extract a plurality of slope image features from the slope pixel image by performing image detection on the slope pixel image; respectively detect variance inflation factors and feature importances of the plurality of slope image features, wherein the variance inflation factor is used to characterize the multicollinearity between the plurality of slope image features, and the feature importance is used to characterize the importance of the slope image features for earthquake-triggered landslide hazard detection; extract earthquake landslide influencing factor data from the plurality of slope image features according to the variance inflation factors and feature importances of the plurality of slope image features.

[0140] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0141] According to the geotechnical property data of each target slope sample, earthquake landslide risk detection is performed on each target slope sample to obtain the earthquake landslide risk index of each target slope sample; according to the earthquake landslide risk index of each target slope sample, negative samples are selected from each target slope sample, and a target training sample set corresponding to the negative sample is generated, wherein the negative sample is a sample whose earthquake landslide risk index is less than a preset risk index and no earthquake landslide has occurred; earthquake landslide impact factor data corresponding to the target training sample set is obtained, and an earthquake landslide detection model is constructed using the earthquake landslide impact factor data as a training data set; according to the earthquake landslide detection model, earthquake-triggered landslide hazard detection is performed on the slope samples to be detected.

[0142] In one embodiment, the preset risk index includes a first preset risk index; when the computer program is executed by the processor, the following steps are also implemented:

[0143] A plurality of first preselected slope samples whose earthquake landslide risk index is less than a first preset risk index are selected from each of the target slope samples; samples where no earthquake landslide has occurred are selected from the plurality of first preselected slope samples as negative samples; samples where earthquake landslide has occurred are selected from each of the target slope samples as positive samples, and all the selected negative samples and positive samples are used together as a target training sample set.

[0144] In one embodiment, the preset risk index includes a first preset risk index and a second preset risk index, and the second preset risk index is greater than the first preset risk index; when the computer program is executed by the processor, the following steps are also implemented:

[0145] A plurality of second pre-selected slope samples whose earthquake landslide risk index is less than a second preset risk index are selected from each of the target slope samples; slope samples whose earthquake landslide risk index is less than the first preset risk index in the plurality of second pre-selected slope samples are used as a first interval sample set, and slope samples other than the first interval sample set in the plurality of second pre-selected slope samples are used as a second interval sample set; samples where no earthquake landslide has occurred are collected from the first interval sample set and the second interval sample set as negative samples; samples where earthquake landslide has occurred are selected from each of the target slope samples as positive samples, and all the selected negative samples and positive samples are used together as a target training sample set.

[0146] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0147] Obtain a first negative sample collection ratio of the first interval sample set and a second negative sample collection ratio of the second interval sample set, wherein the first negative sample collection ratio is greater than the second negative sample collection ratio; and according to the first negative sample collection ratio and the second negative sample collection ratio, collect samples in which no earthquake landslides have occurred from the first interval sample set and the second interval sample set in proportion as negative samples.

[0148] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0149] According to the geotechnical property data of the target slope sample, the static safety factor corresponding to the target slope sample is detected, and based on the static safety factor, the critical acceleration of the target slope sample for earthquake landslide is detected; according to the critical acceleration, the cumulative displacement of the target slope sample for earthquake landslide is detected, and based on the cumulative displacement, the probability of earthquake landslide occurring at the location of the target slope sample is detected to obtain the earthquake landslide risk index of the target slope sample, wherein the earthquake landslide risk index is used to characterize the probability of earthquake-triggered landslide.

[0150] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0151] Obtain a slope pixel image corresponding to the target training sample set, and extract a plurality of slope image features from the slope pixel image by performing image detection on the slope pixel image; respectively detect variance inflation factors and feature importances of the plurality of slope image features, wherein the variance inflation factor is used to characterize the multicollinearity between the plurality of slope image features, and the feature importance is used to characterize the importance of the slope image features for earthquake-triggered landslide hazard detection; extract earthquake landslide influencing factor data from the plurality of slope image features according to the variance inflation factors and feature importances of the plurality of slope image features.

[0152] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0153] According to the geotechnical property data of each target slope sample, earthquake landslide risk detection is performed on each target slope sample to obtain the earthquake landslide risk index of each target slope sample; according to the earthquake landslide risk index of each target slope sample, negative samples are selected from each target slope sample, and a target training sample set corresponding to the negative sample is generated, wherein the negative sample is a sample whose earthquake landslide risk index is less than a preset risk index and no earthquake landslide has occurred; earthquake landslide impact factor data corresponding to the target training sample set is obtained, and an earthquake landslide detection model is constructed using the earthquake landslide impact factor data as a training data set; according to the earthquake landslide detection model, earthquake-triggered landslide hazard detection is performed on the slope samples to be detected.

[0154] In one embodiment, the preset risk index includes a first preset risk index; when the computer program is executed by the processor, the following steps are also implemented:

[0155] A plurality of first preselected slope samples whose earthquake landslide risk index is less than a first preset risk index are selected from each of the target slope samples; samples where no earthquake landslide has occurred are selected from the plurality of first preselected slope samples as negative samples; samples where earthquake landslide has occurred are selected from each of the target slope samples as positive samples, and all the selected negative samples and positive samples are used together as a target training sample set.

[0156] In one embodiment, the preset risk index includes a first preset risk index and a second preset risk index, and the second preset risk index is greater than the first preset risk index; when the computer program is executed by the processor, the following steps are also implemented:

[0157] A plurality of second pre-selected slope samples whose earthquake landslide risk index is less than a second preset risk index are selected from each of the target slope samples; slope samples whose earthquake landslide risk index is less than the first preset risk index in the plurality of second pre-selected slope samples are used as a first interval sample set, and slope samples other than the first interval sample set in the plurality of second pre-selected slope samples are used as a second interval sample set; samples where no earthquake landslide has occurred are collected from the first interval sample set and the second interval sample set as negative samples; samples where earthquake landslide has occurred are selected from each of the target slope samples as positive samples, and all the selected negative samples and positive samples are used together as a target training sample set.

[0158] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0159] Obtain a first negative sample collection ratio of the first interval sample set and a second negative sample collection ratio of the second interval sample set, wherein the first negative sample collection ratio is greater than the second negative sample collection ratio; and according to the first negative sample collection ratio and the second negative sample collection ratio, collect samples in which no earthquake landslides have occurred from the first interval sample set and the second interval sample set in proportion as negative samples.

[0160] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0161] According to the geotechnical property data of the target slope sample, the static safety factor corresponding to the target slope sample is detected, and based on the static safety factor, the critical acceleration of the target slope sample for earthquake landslide is detected; according to the critical acceleration, the cumulative displacement of the target slope sample for earthquake landslide is detected, and based on the cumulative displacement, the probability of earthquake landslide occurring at the location of the target slope sample is detected to obtain the earthquake landslide risk index of the target slope sample, wherein the earthquake landslide risk index is used to characterize the probability of earthquake-triggered landslide.

[0162] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0163] Obtain a slope pixel image corresponding to the target training sample set, and extract a plurality of slope image features from the slope pixel image by performing image detection on the slope pixel image; respectively detect variance inflation factors and feature importances of the plurality of slope image features, wherein the variance inflation factor is used to characterize the multicollinearity between the plurality of slope image features, and the feature importance is used to characterize the importance of the slope image features for earthquake-triggered landslide hazard detection; extract earthquake landslide influencing factor data from the plurality of slope image features according to the variance inflation factors and feature importances of the plurality of slope image features.

[0164] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and 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-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present 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), magnetic 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. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0165] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, 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, they should be considered to be within the scope of this application.

[0166] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for detecting the danger of earthquake-triggered landslides, characterized in that: The method comprises: According to the geotechnical property data of each target slope sample, respectively, an earthquake landslide risk detection is performed on each target slope sample to obtain an earthquake landslide risk index of each target slope sample; According to the earthquake landslide risk index of each target slope sample, a negative sample is selected from each target slope sample, and a target training sample set corresponding to the negative sample is generated, wherein the negative sample is a sample whose earthquake landslide risk index is less than a preset risk index and no earthquake landslide occurs; Acquire earthquake landslide impact factor data corresponding to the target training sample set, and use the earthquake landslide impact factor data as a training data set to construct an earthquake landslide detection model; According to the earthquake landslide detection model, earthquake-triggered landslide hazard detection is performed on the slope sample to be detected.

2. The method according to claim 1, characterized in that The step of selecting negative samples from each target slope sample according to the earthquake landslide risk index of each target slope sample and generating a target training sample set corresponding to the negative sample comprises: Selecting a plurality of first preselected slope samples whose earthquake landslide risk index is less than a first preset risk index from each of the target slope samples; Selecting samples without earthquake landslide from the plurality of first preselected slope samples as negative samples; From each of the target slope samples, samples in which earthquake landslides have occurred are selected as positive samples, and all the selected negative samples and positive samples are taken together as a target training sample set.

3. The method according to claim 1, characterized in that The preset risk index includes a first preset risk index and a second preset risk index, and the second preset risk index is greater than the first preset risk index; The step of selecting negative samples from each target slope sample according to the earthquake landslide risk index of each target slope sample and generating a target training sample set corresponding to the negative sample comprises: Selecting a plurality of second preselected slope samples whose earthquake landslide risk index is less than a second preset risk index from each of the target slope samples; Taking slope samples whose earthquake landslide risk index is less than the first preset risk index among the plurality of second preselected slope samples as a first interval sample set, and taking slope samples other than the first interval sample set among the plurality of second preselected slope samples as a second interval sample set; Collecting samples without earthquake landslide from the first interval sample set and the second interval sample set as negative samples; From each of the target slope samples, samples in which earthquake landslides have occurred are selected as positive samples, and all the selected negative samples and positive samples are taken together as a target training sample set.

4. The method according to claim 3, characterized in that: The collecting samples without earthquake landslide from the first interval sample set and the second interval sample set as negative samples includes: Obtaining a first negative sample collection ratio of the first interval sample set and a second negative sample collection ratio of the second interval sample set, wherein the first negative sample collection ratio is greater than the second negative sample collection ratio; According to the first negative sample collection ratio and the second negative sample collection ratio, samples in which no earthquake landslide occurred are collected in proportion from the first interval sample set and the second interval sample set as negative samples.

5. The method according to claim 1, characterized in that According to the geotechnical characteristic data of each target slope sample, each target slope sample is respectively subjected to earthquake landslide risk detection to obtain an earthquake landslide risk index of each target slope sample, including: According to the geotechnical characteristic data of the target slope sample, the static safety factor corresponding to the target slope sample is detected, and according to the static safety factor, the critical acceleration of the earthquake landslide of the target slope sample is detected; According to the critical acceleration, the cumulative displacement of the target slope sample causing an earthquake landslide is detected, and according to the cumulative displacement, the probability of an earthquake landslide occurring at the location of the target slope sample is detected to obtain an earthquake landslide risk index for the target slope sample, wherein the earthquake landslide risk index is used to characterize the probability of an earthquake-triggered landslide.

6. The method according to claim 1, characterized in that The step of obtaining earthquake landslide impact factor data corresponding to the target training sample set includes: Acquire a slope pixel image corresponding to the target training sample set, and extract a plurality of slope image features from the slope pixel image by performing image detection on the slope pixel image; Respectively detecting variance inflation factors and feature importances of the plurality of slope image features, wherein the variance inflation factor is used to characterize the multicollinearity between the plurality of slope image features, and the feature importance is used to characterize the importance of the slope image features for earthquake-triggered landslide hazard detection; According to the variance inflation factors and feature importances of the plurality of slope image features, earthquake landslide influencing factor data are extracted from the plurality of slope image features.

7. An earthquake-triggered landslide hazard detection device, characterized in that: The device comprises: A first detection module is used to perform earthquake landslide risk detection on each target slope sample according to the geotechnical characteristic data of each target slope sample, and obtain an earthquake landslide risk index of each target slope sample; A sample selection module, for selecting negative samples from each of the target slope samples according to the earthquake landslide risk index of each of the target slope samples, and generating a target training sample set corresponding to the negative samples, wherein the negative samples are samples whose earthquake landslide risk index is less than a preset risk index and where no earthquake landslide occurs; A model building module, used to obtain earthquake landslide impact factor data corresponding to the target training sample set, and use the earthquake landslide impact factor data as a training data set to build an earthquake landslide detection model; The second detection module is used to perform earthquake-triggered landslide hazard detection on the slope sample to be detected according to the earthquake landslide detection model.

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

9. 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 6 are implemented.

10. 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 6 are implemented.

Images