Pylon foundation area slope risk prediction model training method, pylon foundation area slope risk prediction method, device and computer equipment

Abstract

The application relates to a tower foundation region slope risk prediction model training method, a tower foundation region slope risk prediction method, a device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: obtaining sample target data of a sample tower foundation region slope, inputting the sample target data into a plurality of to-be-trained slope risk prediction models, and determining prediction risk evaluation results of the sample tower foundation region slope output by the to-be-trained slope risk prediction models; according to the prediction risk evaluation results, iteratively training the to-be-trained slope risk prediction models to obtain a plurality of trained slope risk prediction models; and from the trained slope risk prediction models, screening a slope risk prediction model corresponding to a model performance parameter satisfying a preset parameter condition as a pre-trained slope risk prediction model. By using the method, the evaluation accuracy of the tower foundation region slope risk can be improved.

Description

Technical Field

[0001] This application relates to the field of power grid technology, and in particular to a method for training a tower base area slope risk prediction model, a method for predicting tower base area slope risk, an apparatus, a computer device, a computer-readable storage medium, and a computer program product. Background Technology

[0002] In the power grid sector, risk assessment of the slopes in the tower base area is crucial for ensuring the stability and safety of the tower base.

[0003] In traditional techniques, risk assessment of the slope in the tower base area is generally conducted through manual surveys. However, this manual survey method is subject to subjective factors and prone to errors, resulting in a low accuracy rate in assessing the risk of the slope in the tower base area. Summary of the Invention

[0004] Based on this, it is necessary to provide a training method, apparatus, computer equipment, computer-readable storage medium, and computer program product for a tower base area slope risk prediction model that can improve the accuracy of the assessment of tower base area slope risk.

[0005] Firstly, this application provides a method for training a slope risk prediction model for the base area of ​​a tower, including:

[0006] Obtain sample target data for the slope of the sample tower base area; the sample target data includes at least slope data and rainfall data for the slope of the sample tower base area;

[0007] The target data of the sample is input into multiple slope risk prediction models to be trained, and the slope risk assessment results and rainfall risk assessment results of the sample tower base area slope are obtained from the output of each slope risk prediction model to be trained.

[0008] Based on the slope risk assessment results and rainfall risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained, the predicted risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained are determined.

[0009] Based on the predicted risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained, and the actual risk assessment results of the sample tower base area slope, the slope risk prediction models to be trained are iteratively trained to obtain multiple trained slope risk prediction models, and the model performance parameters corresponding to each trained slope risk prediction model.

[0010] From the slope risk prediction models that have been trained, select the slope risk prediction models whose performance parameters meet the preset parameter conditions, and use them as pre-trained slope risk prediction models.

[0011] In one embodiment, the step of inputting the sample target data into multiple slope risk prediction models to be trained, and obtaining the slope risk assessment results and rainfall risk assessment results of the sample base area slope output by each of the slope risk prediction models to be trained, includes:

[0012] The slope data of the sample tower base area slope is input into multiple slope risk prediction models to be trained. Through each slope risk prediction model to be trained, the slope data of the sample tower base area slope is processed for risk prediction, and the first prediction probability corresponding to each preset slope risk assessment result output by each slope risk prediction model to be trained is obtained.

[0013] The rainfall data of the slope in the sample tower base area is input into each of the slope risk prediction models to be trained. The rainfall data of the slope in the sample tower base area is processed by the slope risk prediction models to be trained to obtain the second prediction probability corresponding to each preset rainfall risk assessment result output by each slope risk prediction model to be trained.

[0014] The preset slope risk assessment result corresponding to the maximum value of the first predicted probability output by each of the slope risk prediction models to be trained is taken as the slope risk assessment result of the sample tower base area slope output by each of the slope risk prediction models to be trained, and the preset rainfall risk assessment result corresponding to the maximum value of the second predicted probability output by each of the slope risk prediction models to be trained is taken as the rainfall risk assessment result of the sample tower base area slope output by each of the slope risk prediction models to be trained.

[0015] In one embodiment, the step of inputting the sample target data into multiple slope risk prediction models to be trained, and obtaining the slope risk assessment results and rainfall risk assessment results of the sample tower base area slope output by each of the slope risk prediction models to be trained, further includes:

[0016] Feature extraction processing is performed on the slope data of the sample tower base area slope to obtain the first feature vector corresponding to the slope data of the sample tower base area slope;

[0017] Feature extraction processing is performed on the rainfall data of the slope in the sample tower base area to obtain the second feature vector corresponding to the rainfall data of the slope in the sample tower base area;

[0018] The first feature vector is input into multiple slope risk prediction models to be trained, respectively, to obtain the slope risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained, and the second feature vector is input into each slope risk prediction model to be trained, respectively, to obtain the rainfall risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained.

[0019] In one embodiment, determining the predicted risk assessment result of the sample base area slope output by each of the slope risk prediction models to be trained, based on the slope risk assessment result and rainfall risk assessment result of the sample base area slope output by each of the slope risk prediction models to be trained, includes:

[0020] Obtain the target correspondence; the target correspondence is used to characterize the correspondence between slope risk assessment results, rainfall risk assessment results and predicted risk assessment results;

[0021] Based on the target correspondence and the slope risk assessment results and rainfall risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained, the predicted risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained are determined.

[0022] In one embodiment, acquiring sample target data of the slope in the sample tower base area includes:

[0023] Obtain the slope identifier of the base area of ​​the sample tower;

[0024] Obtain the slope data and rainfall data corresponding to the slope markers in the tower base area;

[0025] The slope data and rainfall data are preprocessed to obtain preprocessed slope data and preprocessed rainfall data;

[0026] Both the preprocessed slope data and the preprocessed rainfall data are used as the target sample data for the slope in the sample tower base area.

[0027] Secondly, this application provides a method for predicting slope risk in the base area of ​​a tower, including:

[0028] Obtain target data for the slope of the tower base area to be analyzed; the target data includes at least slope data and rainfall data for the slope of the tower base area to be analyzed.

[0029] The target data is input into a pre-trained slope risk prediction model to obtain the slope risk assessment results and rainfall risk assessment results of the slope in the base area to be analyzed; the pre-trained slope risk prediction model is trained using the base area slope risk prediction model training method.

[0030] Based on the slope risk assessment results and rainfall risk assessment results of the slope in the area to be analyzed, the target risk assessment results of the slope in the area to be analyzed are determined.

[0031] Thirdly, this application also provides a training device for a slope risk prediction model in the base area, comprising:

[0032] The data acquisition module is used to acquire sample target data of the slope in the sample tower base area; the sample target data includes at least slope data and rainfall data of the slope in the sample tower base area.

[0033] The result prediction module is used to input the sample target data into multiple slope risk prediction models to be trained, and to obtain the slope risk assessment results and rainfall risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained.

[0034] The result determination module is used to determine the predicted risk assessment results of the sample tower base area slopes output by each of the slope risk prediction models to be trained, based on the slope risk assessment results and rainfall risk assessment results of the sample tower base area slopes output by each of the slope risk prediction models to be trained.

[0035] The model training module is used to iteratively train each slope risk prediction model to be trained based on the predicted risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained, and the actual risk assessment results of the sample tower base area slope, to obtain multiple trained slope risk prediction models, and the model performance parameters corresponding to each trained slope risk prediction model.

[0036] The model determination module is used to select slope risk prediction models whose performance parameters meet the preset parameter conditions from the slope risk prediction models that have been trained, and use them as pre-trained slope risk prediction models.

[0037] Fourthly, this application also provides a device for predicting slope risk in the base area of ​​a tower, comprising:

[0038] The target acquisition module is used to acquire target data of the slope of the tower base area to be analyzed; the target data includes at least the slope data and rainfall data of the slope of the tower base area to be analyzed.

[0039] The model prediction module is used to input the target data into a pre-trained slope risk prediction model to obtain the slope risk assessment results and rainfall risk assessment results of the slope in the base area to be analyzed; the pre-trained slope risk prediction model is trained by the base area slope risk prediction model training method.

[0040] The result generation module is used to determine the target risk assessment result of the slope in the area to be analyzed, based on the slope risk assessment result and the rainfall risk assessment result of the slope in the area to be analyzed.

[0041] Fifthly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0042] Obtain sample target data for the slope of the sample tower base area; the sample target data includes at least slope data and rainfall data for the slope of the sample tower base area;

[0043] The target data of the sample is input into multiple slope risk prediction models to be trained, and the slope risk assessment results and rainfall risk assessment results of the sample tower base area slope are obtained from the output of each slope risk prediction model to be trained.

[0044] Based on the slope risk assessment results and rainfall risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained, the predicted risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained are determined.

[0045] Based on the predicted risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained, and the actual risk assessment results of the sample tower base area slope, the slope risk prediction models to be trained are iteratively trained to obtain multiple trained slope risk prediction models, and the model performance parameters corresponding to each trained slope risk prediction model.

[0046] From the slope risk prediction models that have been trained, select the slope risk prediction models whose performance parameters meet the preset parameter conditions, and use them as pre-trained slope risk prediction models.

[0047] Sixthly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0048] Obtain target data for the slope of the tower base area to be analyzed; the target data includes at least slope data and rainfall data for the slope of the tower base area to be analyzed.

[0049] The target data is input into a pre-trained slope risk prediction model to obtain the slope risk assessment results and rainfall risk assessment results of the slope in the base area to be analyzed; the pre-trained slope risk prediction model is trained using the base area slope risk prediction model training method.

[0050] Based on the slope risk assessment results and rainfall risk assessment results of the slope in the area to be analyzed, the target risk assessment results of the slope in the area to be analyzed are determined.

[0051] Seventhly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0052] Obtain sample target data for the slope of the sample tower base area; the sample target data includes at least slope data and rainfall data for the slope of the sample tower base area;

[0053] The target data of the sample is input into multiple slope risk prediction models to be trained, and the slope risk assessment results and rainfall risk assessment results of the sample tower base area slope are obtained from the output of each slope risk prediction model to be trained.

[0054] Based on the slope risk assessment results and rainfall risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained, the predicted risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained are determined.

[0055] Based on the predicted risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained, and the actual risk assessment results of the sample tower base area slope, the slope risk prediction models to be trained are iteratively trained to obtain multiple trained slope risk prediction models, and the model performance parameters corresponding to each trained slope risk prediction model.

[0056] From the slope risk prediction models that have been trained, select the slope risk prediction models whose performance parameters meet the preset parameter conditions, and use them as pre-trained slope risk prediction models.

[0057] Eighthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0058] Obtain target data for the slope of the tower base area to be analyzed; the target data includes at least slope data and rainfall data for the slope of the tower base area to be analyzed.

[0059] The target data is input into a pre-trained slope risk prediction model to obtain the slope risk assessment results and rainfall risk assessment results of the slope in the base area to be analyzed; the pre-trained slope risk prediction model is trained using the base area slope risk prediction model training method.

[0060] Based on the slope risk assessment results and rainfall risk assessment results of the slope in the area to be analyzed, the target risk assessment results of the slope in the area to be analyzed are determined.

[0061] Ninthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0062] Obtain sample target data for the slope of the sample tower base area; the sample target data includes at least slope data and rainfall data for the slope of the sample tower base area;

[0063] The target data of the sample is input into multiple slope risk prediction models to be trained, and the slope risk assessment results and rainfall risk assessment results of the sample base area slope are obtained from the output of each slope risk prediction model to be trained.

[0064] Based on the slope risk assessment results and rainfall risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained, the predicted risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained are determined.

[0065] Based on the predicted risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained, and the actual risk assessment results of the sample tower base area slope, the slope risk prediction models to be trained are iteratively trained to obtain multiple trained slope risk prediction models, and the model performance parameters corresponding to each trained slope risk prediction model.

[0066] From the slope risk prediction models that have been trained, select the slope risk prediction models whose performance parameters meet the preset parameter conditions, and use them as pre-trained slope risk prediction models.

[0067] In a tenth aspect, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0068] Obtain target data for the slope of the tower base area to be analyzed; the target data includes at least slope data and rainfall data for the slope of the tower base area to be analyzed.

[0069] The target data is input into a pre-trained slope risk prediction model to obtain the slope risk assessment results and rainfall risk assessment results of the slope in the base area to be analyzed; the pre-trained slope risk prediction model is trained using the base area slope risk prediction model training method.

[0070] Based on the slope risk assessment results and rainfall risk assessment results of the slope in the area to be analyzed, the target risk assessment results of the slope in the area to be analyzed are determined.

[0071] The aforementioned training method, apparatus, computer equipment, storage medium, and computer program product for slope risk prediction models in the tower base area first acquires slope data and rainfall data of the sample tower base area slopes as sample target data. Then, the sample target data is input into multiple slope risk prediction models to be trained, resulting in slope risk assessment results and rainfall risk assessment results for the sample tower base area slopes output by each model. Finally, based on the slope risk assessment results and rainfall risk assessment results output by each model, the risk assessment results for each model to be trained are determined. The slope risk prediction model outputs the predicted risk assessment results of the sample tower base area slope. Then, based on the predicted risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained, and the actual risk assessment results of the sample tower base area slope, the slope risk prediction model to be trained is iteratively trained to obtain multiple trained slope risk prediction models and the corresponding model performance parameters of each trained slope risk prediction model. Finally, from the trained slope risk prediction models, the slope risk prediction models whose corresponding model performance parameters meet the preset parameter conditions are selected as the pre-trained slope risk prediction models. In this way, when conducting risk assessment of the slope in the base area, the slope risk prediction model is iteratively trained. This allows the model to receive new data in each iteration, enabling internal improvements and optimizations, thus facilitating more effective predictions and improving its accuracy. Furthermore, by pre-training the slope risk prediction model, in practical applications, once the target data for the slope to be analyzed is obtained, the risk assessment results can be predicted directly without human intervention. This avoids the subjective factors and errors inherent in manual surveys, which can lead to lower accuracy in slope risk assessment, thereby improving the overall accuracy of slope risk assessment. Attached Figure Description

[0072] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0073] Figure 1 This is a flowchart illustrating the training method for a slope risk prediction model in the base area of ​​a tower in one embodiment.

[0074] Figure 2 This is a flowchart illustrating a method for predicting slope risk in the base area of ​​a tower in one embodiment;

[0075] Figure 3 This is a flowchart illustrating the training method for the slope risk prediction model in the tower base area in another embodiment;

[0076] Figure 4 This is a flowchart illustrating the training method for the slope risk prediction model in the tower base area in another embodiment;

[0077] Figure 5 Here is a monthly rainfall distribution map for City A in one embodiment;

[0078] Figure 6 Here is a map showing the annual rainfall distribution of a pole in one embodiment;

[0079] Figure 7 This is a statistical chart of annual rainfall at different stages for a tower in one embodiment;

[0080] Figure 8 This is a distribution map of the maximum 7-day cumulative rainfall on a pole in one embodiment;

[0081] Figure 9 This is a distribution map of the maximum monthly cumulative rainfall on a pole in one embodiment;

[0082] Figure 10 This is a cumulative rainfall level distribution map for a power pole in one embodiment;

[0083] Figure 11 This is a schematic diagram illustrating the distribution of hazard levels for slope elements in one embodiment;

[0084] Figure 12 This is a schematic diagram of the receiver operation characteristic curves of each model in one embodiment (Class I slope is the positive class);

[0085] Figure 13 This is a schematic diagram of the receiver operation characteristic curves of each model in one embodiment (Class II slope is the positive class);

[0086] Figure 14This is a schematic diagram of the prediction results of a linear support vector machine in one embodiment;

[0087] Figure 15 This is a schematic diagram of the prediction results of a quadratic support vector machine in one embodiment;

[0088] Figure 16 This is a schematic diagram of the prediction results of a Gaussian support vector machine in one embodiment;

[0089] Figure 17 This is a structural block diagram of a training device for a slope risk prediction model in the base area of ​​a tower in one embodiment;

[0090] Figure 18 This is a structural block diagram of a slope risk prediction device for the tower base area in one embodiment;

[0091] Figure 19 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0092] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0093] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0094] In one exemplary embodiment, such as Figure 1 As shown, a method for training a slope risk prediction model for a tower base area is provided. This embodiment illustrates the application of this method to a server; it is understood that this method can also be applied to a terminal, or to a system including a terminal and a server, and is implemented through interaction between the terminal and the server. The terminal can be, but is not limited to, various personal computers, laptops, smartphones, and tablets; the server can be a standalone server or a server cluster composed of multiple servers. In this embodiment, the method includes the following steps:

[0095] Step S101: Obtain sample target data of the slope in the sample tower base area; the sample target data includes at least slope data and rainfall data of the slope in the sample tower base area.

[0096] Among them, the sample base area slope refers to the base area slope used to train the base area slope risk prediction model.

[0097] Among them, the slope of the tower base area refers to the sloping area around the foundation (tower base) of the transmission line tower.

[0098] Among them, the target data of the sample refers to the data associated with the slope of the sample tower base area, including slope data and rainfall data.

[0099] The slope data includes external disturbance data, hazard investigation data, and slope parameter data.

[0100] The classification of slope risk characterization indicators corresponding to the slope data is shown in Table 1 below:

[0101] Table 1. Classification of Slope Risk Characterization Indicators

[0102]

[0103] Among them, the external disturbance data is used to characterize the impact of natural climate change and human production and life on the slope of the tower base area, including four indicators: public transportation construction disturbance, machinery disturbance, excavation disturbance, and water erosion. The impact of natural climate on the slope of the tower base area is mainly characterized by the degree of water erosion. The more severe the erosion, the greater the danger of the slope of the tower base area. The other disturbance indicators reflect the construction situation of the slope of the tower base area. Different construction methods and severity will have different impacts on the slope of the tower base area.

[0104] Among them, the hidden danger investigation data is used to characterize the safety hazards existing on the slope of the tower base area and the detection and protection measures for the hazards, including three indicators: landslide traces, slope protection measures, and monitoring and early warning. Slope protection measures and monitoring and early warning are the protection and monitoring of the slope of the tower base area, which can effectively prevent landslides and increase the stability of the slope of the tower base area. The more complete the measures, the higher the safety factor of the slope of the tower base area. Landslide traces are used to indicate the hidden dangers existing on the slope of the tower base area. The more serious the landslide traces, the worse the stability of the slope of the tower base area.

[0105] Among them, the slope parameter data is used to characterize the structure and properties of the slope in the tower base area. It consists of four slope parameters: soil type, slope, cover layer thickness, and vegetation cover degree. These slope parameters directly affect the stability of the slope in the tower base area. The more solid the soil type and the greater the vegetation cover degree, the more stable the slope in the tower base area.

[0106] Among them, rainfall data is used to characterize the rainfall situation at the location of the slope in the base area, including annual rainfall and cumulative rainfall.

[0107] The classification of rainfall impact indicators corresponding to the rainfall data is shown in Table 2 below:

[0108] Table 2 Classification of Rainfall Impact Indicators

[0109]

[0110] For example, in response to a model training instruction for a slope risk prediction model for a tower base area, the server obtains the tower base area slope identifier for the sample tower base area slope; then, based on the tower base area slope identifier, the server obtains the slope data and rainfall data corresponding to the tower base area slope identifier, and uses them as the slope data and rainfall data for the sample tower base area slope; then, the server uses both the slope data and rainfall data of the sample tower base area slope as the sample target data for the sample tower base area slope.

[0111] Step S102: Input the sample target data into multiple slope risk prediction models to be trained, and obtain the slope risk assessment results and rainfall risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained.

[0112] Among them, the slope risk prediction model refers to a network model that can use sample target data to obtain the slope risk assessment results and rainfall risk assessment results corresponding to the sample tower base area slope, such as the SVM (Support Vector Machine) model. It should be noted that the multiple slope risk prediction models to be trained include linear SVM models, quadratic SVM models, and Gaussian SVM models.

[0113] Among them, the slope risk assessment result refers to the slope risk level of the slope in the base area.

[0114] Among them, the rainfall risk assessment result refers to the rainfall risk level of the slope in the base area of ​​the tower.

[0115] For example, the server inputs the sample target data into multiple slope risk prediction models to be trained. Through the multiple slope risk prediction models to be trained, the sample target data is processed for risk prediction, and the slope risk assessment results and rainfall risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained are obtained.

[0116] Step S103: Based on the slope risk assessment results and rainfall risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained, determine the predicted risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained.

[0117] Among them, the predicted risk assessment result refers to the predicted value corresponding to the comprehensive risk assessment result of the slope in the sample base area output by each slope risk prediction model to be trained.

[0118] For example, the server determines the predicted value corresponding to the comprehensive risk assessment result of the sample base area slope output by each slope risk prediction model to be trained, based on the slope risk assessment results and rainfall risk assessment results of the sample base area slope output by each slope risk prediction model to be trained, and uses it as the predicted risk assessment result of the sample base area slope output by each slope risk prediction model to be trained.

[0119] Step S104: Based on the predicted risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained, and the actual risk assessment results of the sample tower base area slopes, iteratively train each slope risk prediction model to be trained to obtain multiple trained slope risk prediction models, and the model performance parameters corresponding to each trained slope risk prediction model.

[0120] The actual risk assessment result refers to the actual value corresponding to the comprehensive risk assessment result of the sample base area slope output by each slope risk prediction model to be trained.

[0121] Among them, model performance parameters refer to parameter values ​​related to the model performance of the slope risk prediction model, such as accuracy, F1 score, ROC (Receiver Operating Characteristic Curve) curve and AUC (Area Under the Curve).

[0122] The recognition accuracy of each model is shown in Table 3 below:

[0123] Table 3. Identification accuracy of each model

[0124]

[0125] The F1 scores for each model are shown in Table 4 below:

[0126] Table 4 F1 scores for each model

[0127]

[0128] For example, the server obtains the actual risk assessment results of the slope in the sample tower base area, and obtains the loss value based on the predicted risk assessment results of the slope in the sample tower base area output by each slope risk prediction model to be trained, and the actual risk assessment results of the slope in the sample tower base area. Then, the server adjusts the model parameters of each slope risk prediction model to be trained based on the loss value. Then, the server retrains each slope risk prediction model after the model parameters are adjusted until the loss value obtained by each slope risk prediction model after training is less than the loss value threshold, then training stops, and the trained tower slope risk prediction model is used as multiple trained slope risk prediction models. Then, the server obtains the accuracy, F1 score, ROC curve and AUC corresponding to each trained slope risk prediction model as the model performance parameters corresponding to each trained slope risk prediction model.

[0129] Step S105: Select slope risk prediction models whose performance parameters meet the preset parameter conditions from the slope risk prediction models that have been trained, and use them as pre-trained slope risk prediction models.

[0130] Among them, the preset parameter conditions refer to the pre-set parameter conditions used to judge the model performance parameters corresponding to the slope risk prediction model. It should be noted that the preset parameter conditions can be determined as needed.

[0131] For example, the server selects slope risk prediction models whose performance parameters meet preset parameter conditions from the slope risk prediction models based on the model performance parameters corresponding to each trained slope risk prediction model; then, the server uses the slope risk prediction models whose performance parameters meet preset parameter conditions as pre-trained slope risk prediction models.

[0132] For example, as shown in Tables 3 and 4, the server can integrate multiple model performance parameters to screen pre-trained slope risk prediction models. For instance, in terms of accuracy, the linear SVM model has an overall accuracy of 87.61%, the quadratic SVM model has an overall accuracy of 93.24%, and the Gaussian SVM model has an overall accuracy of 73.94%. In terms of F1 score, the linear SVM model has an overall F1 score of 0.872, the quadratic SVM model has an overall F1 score of 0.931, and the Gaussian SVM model has an overall F1 score of 0.738. Overall, the quadratic SVM model is optimal; therefore, it is selected as the pre-trained slope risk prediction model.

[0133] In the above-mentioned training method for slope risk prediction models in the tower base area, slope data and rainfall data of the sample tower base area slopes are first acquired as sample target data. Then, the sample target data are input into multiple slope risk prediction models to be trained, resulting in slope risk assessment results and rainfall risk assessment results for the sample tower base area slopes output by each model. Next, based on the slope risk assessment results and rainfall risk assessment results of the sample tower base area slopes output by each model, the sample tower base area slopes output by each model are determined. The predicted risk assessment results of the regional slope are obtained. Then, based on the predicted risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained, and the actual risk assessment results of the sample tower base area slope, the slope risk prediction models to be trained are iteratively trained to obtain multiple trained slope risk prediction models and the corresponding model performance parameters of each trained slope risk prediction model. Finally, from the trained slope risk prediction models, the slope risk prediction models whose corresponding model performance parameters meet the preset parameter conditions are selected as the pre-trained slope risk prediction models. In this way, when conducting risk assessment of the slope in the base area, the slope risk prediction model is iteratively trained. This allows the model to receive new data in each iteration, enabling internal improvements and optimizations, thus facilitating more effective predictions and improving its accuracy. Furthermore, by pre-training the slope risk prediction model, in practical applications, once the target data for the slope to be analyzed is obtained, the risk assessment results can be predicted directly without human intervention. This avoids the subjective factors and errors inherent in manual surveys, which can lead to lower accuracy in slope risk assessment, thereby improving the overall accuracy of slope risk assessment.

[0134] In an exemplary embodiment, step S102, which involves inputting sample target data into multiple slope risk prediction models to be trained, and obtaining slope risk assessment results and rainfall risk assessment results for the sample base area slopes output by each slope risk prediction model to be trained, further includes: inputting slope data of the sample base area slopes into multiple slope risk prediction models to be trained, performing risk prediction processing on the slope data of the sample base area slopes through each slope risk prediction model to obtain the first prediction probability corresponding to each preset slope risk assessment result output by each slope risk prediction model to be trained; and inputting rainfall data of the sample base area slopes into each slope risk prediction model to be trained. In this process, the rainfall data of the sample tower base area slope is processed by each slope risk prediction model to be trained to obtain the second prediction probability corresponding to each preset rainfall risk assessment result output by each slope risk prediction model to be trained. The preset slope risk assessment result corresponding to the maximum value of the first prediction probability output by each slope risk prediction model to be trained is taken as the slope risk assessment result of the sample tower base area slope output by each slope risk prediction model to be trained, and the preset rainfall risk assessment result corresponding to the maximum value of the second prediction probability output by each slope risk prediction model to be trained is taken as the rainfall risk assessment result of the sample tower base area slope output by each slope risk prediction model to be trained.

[0135] Among them, risk prediction processing refers to the process of generating the predicted probability corresponding to the preset risk assessment result.

[0136] The preset slope risk assessment result refers to the pre-defined slope risk assessment result, such as Level I, Level II, etc. It should be noted that the preset slope risk assessment result can be determined depending on the circumstances.

[0137] The first predicted probability refers to the predicted probability corresponding to the preset slope risk assessment result.

[0138] The preset rainfall risk assessment result refers to the pre-set rainfall risk assessment result, such as Level I, Level II, etc. It should be noted that the preset rainfall risk assessment result can be determined on a case-by-case basis.

[0139] The second prediction probability refers to the prediction probability corresponding to the preset rainfall risk assessment result.

[0140] For example, the server inputs slope data of the sample tower base area slope into multiple slope risk prediction models to be trained. Each model performs risk prediction processing on the slope data, obtaining the prediction probability corresponding to each preset slope risk assessment result output by the model, which is used as the first prediction probability output by each model. Next, the server inputs rainfall data of the sample tower base area slope into each model, again performing risk prediction processing on the rainfall data, obtaining the prediction probability corresponding to each preset rainfall risk assessment result output by each model, which is used as the first prediction probability output by each model. The server outputs a second predicted probability; then, it determines the preset slope risk assessment result corresponding to the maximum value of the first predicted probability output by each slope risk prediction model to be trained, and the preset rainfall risk assessment result corresponding to the maximum value of the second predicted probability output by each slope risk prediction model to be trained; finally, the server uses the preset slope risk assessment result corresponding to the maximum value of the first predicted probability output by each slope risk prediction model to be trained as the slope risk assessment result of the sample base area slope output by each slope risk prediction model to be trained, and the preset rainfall risk assessment result corresponding to the maximum value of the second predicted probability output by each slope risk prediction model to be trained as the rainfall risk assessment result of the sample base area slope output by each slope risk prediction model to be trained.

[0141] In this embodiment, by comprehensively considering slope data and rainfall data, using multiple models to be trained for prediction and selecting the evaluation result corresponding to the maximum value, the risk status of the slope in the base area can be assessed more comprehensively and accurately. Moreover, the participation and comprehensive judgment of multiple models can help reduce the bias and uncertainty that may exist in a single model and improve the reliability of the risk assessment results.

[0142] In an exemplary embodiment, step S102, which involves inputting sample target data into multiple slope risk prediction models to be trained, and obtaining slope risk assessment results and rainfall risk assessment results for the sample base area slopes output by each slope risk prediction model to be trained, further includes: performing feature extraction processing on the slope data of the sample base area slopes to obtain a first feature vector corresponding to the slope data of the sample base area slopes; performing feature extraction processing on the rainfall data of the sample base area slopes to obtain a second feature vector corresponding to the rainfall data of the sample base area slopes; inputting the first feature vector into multiple slope risk prediction models to be trained to obtain slope risk assessment results for the sample base area slopes output by each slope risk prediction model to be trained; and inputting the second feature vector into each slope risk prediction model to be trained to obtain rainfall risk assessment results for the sample base area slopes output by each slope risk prediction model to be trained.

[0143] Feature extraction refers to the process of determining the feature vectors corresponding to the data.

[0144] The first feature vector refers to the feature vector corresponding to the slope data of the sample tower base area slope.

[0145] The second feature vector refers to the feature vector corresponding to the rainfall data of the slope in the sample tower base area.

[0146] For example, the server uses the slope data of the sample tower base area slope as the primary data and the rainfall data of the sample tower base area slope as the auxiliary data, and inputs them into the feature extraction model for feature extraction processing to obtain the feature vector corresponding to the slope data of the sample tower base area slope, which is used as the first feature vector; then, the server uses the rainfall data of the sample tower base area slope as the primary data and the slope data of the sample tower base area slope as the auxiliary data, and inputs them into the feature extraction model for feature extraction processing to obtain the feature vector corresponding to the rainfall data of the sample tower base area slope, which is used as the second feature vector; Then, the server inputs the first feature vector into multiple slope risk prediction models to be trained, and classifies the first feature vector through multiple slope risk prediction models to obtain the slope risk assessment results of the sample base area slopes output by each slope risk prediction model to be trained. Next, the server inputs the second feature vector into each slope risk prediction model to be trained, and classifies the second feature vector through multiple slope risk prediction models to obtain the rainfall risk assessment results of the sample base area slopes output by each slope risk prediction model to be trained.

[0147] In this embodiment, feature extraction simplifies complex data into feature vectors, facilitating rapid model processing and analysis and improving overall computational efficiency. Furthermore, inputting the extracted feature vectors into the model for prediction enables a more accurate assessment of slope risk and rainfall risk, making the evaluation results more targeted and reliable.

[0148] In an exemplary embodiment, step S103, which determines the predicted risk assessment result of the sample base area slope output by each slope risk prediction model to be trained based on the slope risk assessment result and rainfall risk assessment result of the sample base area slope output by each slope risk prediction model to be trained, further includes: obtaining a target correspondence; the target correspondence is used to characterize the correspondence between the slope risk assessment result, the rainfall risk assessment result and the predicted risk assessment result; and the predicted risk assessment result of the sample base area slope output by each slope risk prediction model to be trained is determined based on the target correspondence and the slope risk assessment result and rainfall risk assessment result of the sample base area slope output by each slope risk prediction model to be trained.

[0149] The target correspondence is used to characterize the correspondence between the slope risk assessment results, rainfall risk assessment results, and predicted risk assessment results. For example, if the slope risk assessment result is Level I and the rainfall risk assessment result is Level II, the predicted risk assessment result is Level I; if the slope risk assessment result is Level I and the rainfall risk assessment result is Level I, the predicted risk assessment result is Level I; if the slope risk assessment result is Level II and the rainfall risk assessment result is Level II, the predicted risk assessment result is Level II.

[0150] For example, the server obtains the correspondence between the slope risk assessment results, rainfall risk assessment results, and predicted risk assessment results as the target correspondence; then, based on the target correspondence and the slope risk assessment results and rainfall risk assessment results of the sample base area slopes output by each slope risk prediction model to be trained, the server determines the predicted risk assessment results corresponding to the slope risk assessment results and rainfall risk assessment results of the sample base area slopes output by each slope risk prediction model to be trained, as the predicted risk assessment results of the sample base area slopes output by each slope risk prediction model to be trained.

[0151] In this embodiment, by establishing a target correspondence, slope risk and rainfall risk can be combined to determine the final predicted risk assessment result. Moreover, by comprehensively considering the correlation of multiple factors, it is beneficial to grasp the overall risk situation more accurately, avoid the limitations of single-factor assessment, and thus improve the accuracy of the predicted risk assessment result.

[0152] In an exemplary embodiment, step S101 above, obtaining sample target data of the sample tower base area slope, further includes: obtaining the tower base area slope identifier of the sample tower base area slope; obtaining the slope data and rainfall data corresponding to the tower base area slope identifier; preprocessing the slope data and rainfall data to obtain preprocessed slope data and preprocessed rainfall data; and using both the preprocessed slope data and preprocessed rainfall data as sample target data of the sample tower base area slope.

[0153] Among them, the slope identification of the tower base area refers to the unique identification information used to distinguish the slopes in the tower base area, such as the name of the slope in the tower base area.

[0154] Preprocessing refers to the process of pre-processing slope data and rainfall data.

[0155] Among them, the pre-processed slope data refers to the slope data after pre-processing.

[0156] Among them, preprocessed rainfall data refers to rainfall data that has undergone preprocessing.

[0157] For example, in response to a model training instruction for a slope risk prediction model for the base area, the server obtains the base area slope identifier of the sample base area slope. Then, based on the base area slope identifier, the server obtains the slope data and rainfall data corresponding to that identifier, using them as the slope data and rainfall data for the sample base area slope. Next, the server preprocesses the slope data and rainfall data to obtain preprocessed slope data and preprocessed rainfall data. For example, the server uses data processing techniques to clean, fuse, and transform the slope data and rainfall data to obtain preprocessed slope data and preprocessed rainfall data. Finally, the server uses both the preprocessed slope data and preprocessed rainfall data as the sample target data for the sample base area slope.

[0158] In this embodiment, by preprocessing the data, noise and outliers in the slope data and rainfall data can be removed, making the data more accurate and reliable, thus providing more valuable information for subsequent analysis. Moreover, through preprocessing, the data becomes more standardized and uniform in terms of format and magnitude, which facilitates the processing and analysis of the subsequent slope risk prediction model.

[0159] In one exemplary embodiment, such as Figure 2 As shown, a method for predicting slope risk in the base area of ​​a tower is provided. Taking the application of this method to a server as an example, the specific steps include:

[0160] Step S201: Obtain target data of the slope in the tower base area to be analyzed; the target data includes at least slope data and rainfall data of the slope in the tower base area to be analyzed.

[0161] Step S202: Input the target data into the pre-trained slope risk prediction model to obtain the slope risk assessment results and rainfall risk assessment results of the slope in the base area to be analyzed; the pre-trained slope risk prediction model is trained by the base area slope risk prediction model training method.

[0162] Step S203: Based on the slope risk assessment results and rainfall risk assessment results of the slope in the tower base area to be analyzed, determine the target risk assessment results of the slope in the tower base area to be analyzed.

[0163] Among them, the slope of the tower base area to be analyzed refers to the slope of the tower base area that needs to be risk predicted.

[0164] The target data refers to the data associated with the slope of the tower base area to be analyzed, including slope data and rainfall data.

[0165] Among them, the target risk assessment result refers to the comprehensive risk assessment result of the slope in the base area to be analyzed, output by the pre-trained slope risk prediction model.

[0166] For example, in response to a risk prediction instruction for the slope of the base area to be analyzed, the server obtains the slope identifier of the base area to be analyzed. Then, based on the slope identifier, the server obtains the corresponding slope data and rainfall data, which are used as the slope data and rainfall data for the base area to be analyzed. Next, the server uses both the slope data and rainfall data as the target data for the base area to be analyzed. Then, the server inputs the target data into a pre-trained slope risk prediction model, performs risk prediction processing on the target data using the pre-trained model, and obtains the slope risk assessment result and rainfall risk assessment result for the base area to be analyzed. Finally, based on the target correspondence, the server determines the risk assessment result corresponding to the slope risk assessment result and rainfall risk assessment result for the base area to be analyzed, and uses this as the target risk assessment result for the base area to be analyzed.

[0167] In this embodiment, a pre-trained slope risk prediction model is used to predict the slope data and rainfall data of the slope in the base area to be analyzed, thereby enabling a more comprehensive and accurate assessment of the risk status of the slope in the base area, which is beneficial to improving the accuracy of the slope risk assessment in the base area.

[0168] In one exemplary embodiment, such as Figure 3As shown, another method for training a slope risk prediction model for the base area is provided. Taking the application of this method to a server as an example, the method includes the following steps:

[0169] Step S301: Obtain the slope identifier of the base area slope of the sample tower base area; obtain the slope data and rainfall data corresponding to the slope identifier of the base area slope; preprocess the slope data and rainfall data to obtain preprocessed slope data and preprocessed rainfall data; use the preprocessed slope data and preprocessed rainfall data as the sample target data of the sample tower base area slope.

[0170] Step S302: Input the slope data of the sample tower base area slope into multiple slope risk prediction models to be trained. Through each slope risk prediction model to be trained, perform risk prediction processing on the slope data of the sample tower base area slope to obtain the first prediction probability corresponding to each preset slope risk assessment result output by each slope risk prediction model to be trained.

[0171] In step S303, the rainfall data of the sample tower base area slope is input into each slope risk prediction model to be trained. Through each slope risk prediction model to be trained, the rainfall data of the sample tower base area slope is processed for risk prediction, and the second prediction probability corresponding to each preset rainfall risk assessment result output by each slope risk prediction model to be trained is obtained.

[0172] Step S304: The preset slope risk assessment result corresponding to the maximum value of the first prediction probability output by each slope risk prediction model to be trained is used as the slope risk assessment result of the sample tower base area slope output by each slope risk prediction model to be trained; and the preset rainfall risk assessment result corresponding to the maximum value of the second prediction probability output by each slope risk prediction model to be trained is used as the rainfall risk assessment result of the sample tower base area slope output by each slope risk prediction model to be trained.

[0173] Step S305: Perform feature extraction processing on the slope data of the sample tower base area slope to obtain the first feature vector corresponding to the slope data of the sample tower base area slope.

[0174] Step S306: Perform feature extraction processing on the rainfall data of the slope in the sample tower base area to obtain the second feature vector corresponding to the rainfall data of the slope in the sample tower base area.

[0175] Step S307: Input the first feature vector into each of the multiple slope risk prediction models to be trained to obtain the slope risk assessment results of the sample base area slopes output by each slope risk prediction model to be trained, and input the second feature vector into each slope risk prediction model to be trained to obtain the rainfall risk assessment results of the sample base area slopes output by each slope risk prediction model to be trained.

[0176] Step S308: Obtain the target correspondence; the target correspondence is used to characterize the correspondence between the slope risk assessment results, rainfall risk assessment results and predicted risk assessment results.

[0177] Step S309: Based on the target correspondence and the slope risk assessment results and rainfall risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained, determine the predicted risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained.

[0178] Step S310: Based on the predicted risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained, and the actual risk assessment results of the sample tower base area slopes, iteratively train each slope risk prediction model to be trained to obtain multiple trained slope risk prediction models and the corresponding model performance parameters of each trained slope risk prediction model.

[0179] Step S311: Select slope risk prediction models whose performance parameters meet the preset parameter conditions from the slope risk prediction models that have been trained, and use them as pre-trained slope risk prediction models.

[0180] In the aforementioned training method for the slope risk prediction model in the tower base area, when conducting risk assessment of the tower base area slope, the slope risk prediction model is iteratively trained. This allows the model to receive new data in each iteration, enabling internal model improvement and optimization, thus facilitating more effective prediction and improving the prediction accuracy. Furthermore, by pre-training the slope risk prediction model, in practical applications, after obtaining the target data of the slope to be analyzed, the risk assessment results of the slope to be analyzed can be predicted. The entire process requires no manual intervention, avoiding the subjective factors and errors inherent in manual surveying, which can lead to low accuracy in assessing the risk of the tower base area slope. This, in turn, improves the accuracy of the assessment of the risk of the tower base area slope.

[0181] In an exemplary embodiment, to more clearly illustrate the training method for the slope risk prediction model of the tower base area provided in this application, the following specific embodiment will be used to describe the training method for the slope risk prediction model of the tower base area. In one embodiment, as follows... Figure 4 As shown, this application also provides a slope hazard assessment model for the base area. When assessing the risk of the base area slope, slope data and rainfall data of the sample base area slope are first obtained as sample target data. Then, the sample target data are input into multiple slope risk prediction models to be trained, obtaining the slope risk assessment results and rainfall risk assessment results of the sample base area slope output by each model. Next, based on the slope risk assessment results and rainfall risk assessment results of the sample base area slope output by each model, the slope risk of each model is determined. The process involves several steps. First, the slope risk prediction model outputs the predicted risk assessment results for the sample tower base area slopes. Then, based on these results and the actual risk assessment results, the model is iteratively trained to obtain multiple trained slope risk prediction models and their corresponding performance parameters. Finally, the model whose performance parameters meet the preset conditions is selected as the pre-trained slope risk prediction model. This process includes the following:

[0182] This application uses the support vector machine algorithm to train and analyze 249 transmission line slopes in City A, establishes a multi-class support vector machine model for the identification of transmission line slope hazards, and applies the model to the practice of transmission line slope hazard identification to analyze its identification accuracy.

[0183] 1. Analysis of factors influencing the risk level of transmission line slopes:

[0184] Many factors affect the stability of transmission line slopes. This application, through extensive review of references and summarizing engineering experience, analyzes the slope from three aspects: its own characteristics, human and environmental disturbances, and slope protection measures. Ultimately, it determines three major indicators—external force disturbance, hidden danger investigation, and slope parameters—to characterize the degree of danger of the slope.

[0185] (1) External disturbances: External disturbances mainly characterize the impact of natural climate change and human production and life on slopes. Four indicators were selected: disturbance from public transportation construction, disturbance from machinery, disturbance from excavation, and disturbance from water flow. The impact of natural climate on slopes is mainly characterized by the degree of water flow erosion. The more severe the erosion, the greater the danger of the slope. The other disturbance indicators reflect the construction situation of power transmission line slopes. Different construction methods and degrees of severity have different impacts on slopes.

[0186] (2) Hazard Investigation: Hazard investigation mainly identifies existing safety hazards on the slope and the detection and protection measures for these hazards, including three indicators: slope collapse traces, slope protection measures, and monitoring and early warning. Slope protection measures and monitoring and early warning are for the protection and monitoring of the slope, which can effectively prevent landslides and increase slope stability. The more comprehensive the measures, the higher the safety factor of the slope. Slope collapse traces are used to indicate existing hidden dangers on the slope. The more severe the slip marks, the worse the slope stability.

[0187] (3) Slope parameters: Slope parameters mainly characterize the structure and properties of the slope itself, and are characterized by soil type, slope gradient, cover thickness, and vegetation cover. These four slope parameters directly affect the stability of the slope. The more solid the soil type and the greater the vegetation cover, the more stable the slope.

[0188] 2. Impact of rainfall on the risk level of power transmission line slopes:

[0189] Rainfall is a significant factor inducing landslides in transmission line tower foundation areas and is a crucial indicator in slope hazard assessment models. To improve the accuracy of model evaluation, considering rainfall as a factor in the support vector machine model for tower foundation slope hazard assessment allows for a more objective and effective evaluation of the actual slope hazard level, thus improving the accuracy of model predictions. The annual rainfall data for City A is as follows: Figure 5 As shown.

[0190] The impact of rainfall is reflected by two indicators: annual rainfall and cumulative rainfall. These indicators respectively characterize the impact of general rainfall level (static intrinsic factor) and accidental rainfall (external triggering factor) on slope hazard of the tower base, thus establishing the correlation between rainfall and slope hazard level.

[0191] like Figure 6 As shown in the diagram, the annual rainfall distribution map for each tower reflects the annual rainfall in its region. The highest annual rainfall is 2799 mm, and the lowest is 1917 mm. There are a total of 52 towers with annual rainfall exceeding 2400 mm, and a total of 100 towers with annual rainfall less than 2100 mm. Figure 7 As shown, the specific distribution for each stage is illustrated. Figure 8 As shown, cumulative rainfall is reflected by two indicators: the maximum monthly cumulative rainfall and the maximum 7-day cumulative rainfall in a year.

[0192] like Figure 9 As shown, based on the maximum monthly cumulative rainfall and the maximum 7-day cumulative rainfall of the tower, the cumulative rainfall index of the tower is divided into 5 levels. Level I: 7-day cumulative rainfall greater than 350mm or monthly cumulative rainfall greater than 700mm; Level II: 7-day cumulative rainfall greater than 300mm or monthly cumulative rainfall greater than 600mm; Level III: 7-day cumulative rainfall greater than 250mm or monthly cumulative rainfall greater than 500mm; Level IV: 7-day cumulative rainfall greater than 200mm or monthly cumulative rainfall greater than 400mm; Level V: 7-day cumulative rainfall less than 200mm or monthly cumulative rainfall less than 400mm. Level I represents the most concentrated rainfall and has the highest probability of causing rainfall-related disasters; Level V represents the most moderate rainfall and has the lowest probability of causing rainfall-related disasters. Figure 10 As shown.

[0193] 3. Establishment of a multi-classification SVM model for identifying the risk level of transmission line slopes:

[0194] (1) Establish learning samples:

[0195] A high-voltage power company conducted on-site inspections of the slopes of its 249 transmission lines. Based on the actual engineering conditions, the company categorized the indicators and graded their deterioration levels. To better reflect the slope conditions, the three indicators—public transportation construction disturbance, slope protection measures, and soil type—were further refined into three levels. Ultimately, based on the slope conditions, they were classified into four hazard levels: 122 slopes were classified as Level I and II hazardous slopes, while 127 slopes were classified as Level III and IV safe slopes.

[0196] (2) SVM model establishment:

[0197] This application employs a classification learner, taking the 249 transmission line slope samples mentioned above as learning samples and using their various indicators as inputs to establish a multi-class support vector machine model for transmission line slope hazard identification.

[0198] First, the data was screened. Since the proportion of different slope types in the sample was uneven, with a larger number of Class II and III slopes and a smaller number of Class I and IV slopes, stratified random sampling was performed according to the slope proportions. The slope samples were divided into 5 groups, each with an equal proportion of each slope type. Then, the model was repeatedly trained using 5-fold cross-validation to verify its accuracy.

[0199] Regarding the selection of kernel functions for Support Vector Machines (SVMs), linear SVMs have a simple structure and low computational cost, but insufficient dimensionality may lead to unsatisfactory classification results. In contrast, quadratic kernels have higher dimensionality and a more complete model structure, but correspondingly, higher computational cost and longer processing time. Gaussian kernels are generally used to solve nonlinear mapping problems and have good generalization ability, with computational cost and processing time not significantly different from quadratic kernels. Therefore, SVM models using three kernel functions were compared to evaluate the prediction results of quadratic, Gaussian, and linear SVMs. By combining practical engineering needs, a comprehensive analysis was conducted on prediction accuracy, F1 score, ROC curve, and AUC to identify the model with the best prediction performance and most suitable for engineering practice.

[0200] 4. Comparison and analysis of recognition effects:

[0201] Analysis of the predicted distribution of hazard levels for various slope types revealed that linear SVM identified 8.43%, 38.96%, 52.2%, and 0.40% of slopes as high, medium, low, and no hazard levels, respectively. In contrast, quadratic SVM and Gaussian SVM identified 11.25%, 36.54%, 51.81%, and 0.40% of slopes as high, medium, low, and no hazard levels, respectively, and 2.41%, 25.30%, 63.05%, and 9.24% of slopes as high, medium, low, and no hazard levels, respectively. Preliminary analysis based on the distribution indicates that Gaussian SVM predicted a significantly higher proportion of Class III and Class IV slopes than linear and quadratic SVMs, while predicting Class I and Class II slopes was lower than that of linear and quadratic SVMs. The following section analyzes specific indicators, such as... Figure 11 As shown.

[0202] (1) Accuracy:

[0203] Accuracy can initially reflect the predictive performance of the model. This application uses 5-fold cross-validation, randomly dividing the training samples into 5 parts, each part serving as the test set in turn, and the remaining samples as the training set. The average accuracy after 5 training iterations is used as the overall accuracy of the sample training. By analyzing the identification accuracy of quadratic kernel, Gaussian kernel, and linear support vector machine, Gaussian SVM has the weakest identification rate. Although linear SVM has a strong ability to identify Class II and III slopes, it is weak in identifying Class I slopes, resulting in a lower overall accuracy. Quadratic SVM has better overall accuracy and a stronger ability to identify Class I and Class II slopes with high risk, making it the most suitable for engineering needs.

[0204] (2) F1 score:

[0205] The F1 score is an evaluation metric used in machine learning for classification models. When the training samples are imbalanced, the accuracy cannot accurately and objectively reflect the quality of the model. This application focuses on analyzing the F1 scores of Class I and Class II slopes, which have a higher degree of danger.

[0206] Analysis shows that Gaussian SVM performs well in identifying Class II slopes, but its F1 score is poor for Class I slopes, failing to effectively identify them. Linear SVM shows good precision for Class I slopes, but its low recall results in a lower F1 score compared to quadratic SVM. Quadratic SVM achieves the best F1 scores for both Class I and Class II slopes among the three models, demonstrating the strongest identification ability.

[0207] (3) ROC curve and AUC:

[0208] The ROC curve is plotted using a series of different classification thresholds, with the true positive rate on the y-axis and the false positive rate on the x-axis. This curve can reflect the model's classification ability to a certain extent; the closer it is to the upper left corner, the better the model's classification performance. AUC refers to the area enclosed by the ROC curve and the lower coordinate axis, ranging from [0,1]. A larger AUC indicates that the ROC curve is closer to the upper left corner, and the better the model's classification performance.

[0209] like Figure 12 and Figure 13 As shown, by plotting the ROC curves with Class I slopes and Class II slopes as the positive class, it can be clearly seen that the quadratic SVM has the highest AUC value among the three models, the largest area enclosed by the ROC curve and the horizontal and vertical axes, and the ROC curve is closer to the upper left corner. Therefore, the quadratic kernel is significantly better than the linear SVM and Gaussian SVM in terms of classification ability.

[0210] (4) Prediction results display

[0211] like Figure 14 , Figure 15 and Figure 16 As shown, the prediction results of the three support vector machine models can provide a reference for engineering practice, and preliminarily screen out Class I and Class II slopes for key detection and investigation. According to the above index analysis, the quadratic SVM has the best performance, so we can focus on the Class I and Class II slopes predicted by the quadratic SVM.

[0212] 5. Conclusion:

[0213] This application uses Support Vector Machines (SVMs) to train and analyze 249 transmission line slopes, establishing a multi-class SVM model for transmission line slope hazard identification. The classification performance of quadratic kernel functions, Gaussian kernel functions, and linear SVMs was compared using three indicators: classification accuracy, F1 score, and ROC curve. The linear SVM model is simple and computationally inexpensive, but its dimensionality is insufficient, resulting in a lower classification performance compared to quadratic SVMs. Gaussian SVMs have moderate computational difficulty but are prone to misclassifying slopes as Class II and Class III slopes, with insufficient accuracy in identifying Class I slopes. Ultimately, the quadratic kernel function was selected as the best performer, achieving an accuracy of 93.24%, with accuracies of 96.83% and 94.38% for Class I and Class II slopes, respectively. This model effectively identifies highly hazardous Class I and Class II transmission line slopes. This model can provide a certain reference standard for subsequent slope hazard investigation and hazard prevention measures.

[0214] In the above embodiments, when conducting risk assessment of the slope in the tower base area, the risk prediction model for the tower base area is iteratively trained. This allows the model to receive new data in each iteration, enabling internal improvements and optimizations, thus facilitating more effective predictions and improving the prediction accuracy. Furthermore, by pre-training the risk prediction model, in practical applications, after obtaining the target data for the slope in the tower base area to be analyzed, the risk assessment results can be predicted without human intervention. This avoids the subjective factors and errors inherent in manual surveys, which can lead to lower accuracy in assessing the risk of the tower base area slope, thereby improving the overall accuracy of the risk assessment. Meanwhile, this application uses external disturbance, hidden danger investigation, and slope parameters as the characterization of the slope hazard level in the tower base area. The maximum monthly cumulative rainfall and the maximum 7-day cumulative rainfall of the tower are used to divide the cumulative rainfall index of the tower into 5 levels, resulting in a function characterizing the slope hazard level in the tower base area. The characterization is then divided into intervals to achieve the classification of the slope hazard level in the tower base area. This provides a more intuitive description of the current hazard level of the tower base slope area, providing an effective evaluation method for the safe maintenance of the tower base slope. It is beneficial to save manpower and financial resources and improve the safety of the tower and ensure power supply.

[0215] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated in this application, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed 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 performed alternately or in turn with other steps or at least some of the steps or stages in other steps.

[0216] Based on the same inventive concept, this application also provides a tower base area slope risk prediction model training device for implementing the above-mentioned tower base area slope risk prediction model training method. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more tower base area slope risk prediction model training device embodiments provided below can be found in the limitations of the tower base area slope risk prediction model training method above, and will not be repeated here.

[0217] In one exemplary embodiment, such as Figure 17 As shown, a training device for a slope risk prediction model in a tower base area is provided, comprising: a data acquisition module 1701, a result prediction module 1702, a result determination module 1703, a model training module 1704, and a model determination module 1705, wherein:

[0218] The data acquisition module 1701 is used to acquire sample target data of the slope in the sample tower base area; the sample target data includes at least slope data and rainfall data of the slope in the sample tower base area.

[0219] The result prediction module 1702 is used to input the sample target data into multiple slope risk prediction models to be trained, and obtain the slope risk assessment results and rainfall risk assessment results of the sample base area slopes output by each slope risk prediction model to be trained.

[0220] The result determination module 1703 is used to determine the predicted risk assessment results of the sample base area slopes output by each slope risk prediction model to be trained, based on the slope risk assessment results and rainfall risk assessment results of the sample base area slopes output by each slope risk prediction model to be trained.

[0221] The model training module 1704 is used to iteratively train each slope risk prediction model to be trained based on the predicted risk assessment results of the sample tower base area slopes output by each slope risk prediction model to be trained, as well as the actual risk assessment results of the sample tower base area slopes, to obtain multiple trained slope risk prediction models and the corresponding model performance parameters of each trained slope risk prediction model.

[0222] The model determination module 1705 is used to select slope risk prediction models whose performance parameters meet the preset parameter conditions from the various trained slope risk prediction models, and use them as pre-trained slope risk prediction models.

[0223] In an exemplary embodiment, the result prediction module 1702 is further configured to input the slope data of the sample tower base area slope into multiple slope risk prediction models to be trained, and to perform risk prediction processing on the slope data of the sample tower base area slope through each slope risk prediction model to be trained, thereby obtaining the first prediction probability corresponding to each preset slope risk assessment result output by each slope risk prediction model to be trained; and to input the rainfall data of the sample tower base area slope into each slope risk prediction model to be trained, and to perform risk prediction processing on the rainfall data of the sample tower base area slope through each slope risk prediction model to be trained. The prediction process yields the second prediction probability corresponding to each preset rainfall risk assessment result output by each slope risk prediction model to be trained. The preset slope risk assessment result corresponding to the maximum value of the first prediction probability output by each slope risk prediction model to be trained is used as the slope risk assessment result of the sample base area slope output by each slope risk prediction model to be trained. The preset rainfall risk assessment result corresponding to the maximum value of the second prediction probability output by each slope risk prediction model to be trained is also used as the rainfall risk assessment result of the sample base area slope output by each slope risk prediction model to be trained.

[0224] In an exemplary embodiment, the result prediction module 1702 is further configured to perform feature extraction processing on the slope data of the sample tower base area slope to obtain a first feature vector corresponding to the slope data of the sample tower base area slope; perform feature extraction processing on the rainfall data of the sample tower base area slope to obtain a second feature vector corresponding to the rainfall data of the sample tower base area slope; input the first feature vector into multiple slope risk prediction models to be trained to obtain the slope risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained; and input the second feature vector into each slope risk prediction model to be trained to obtain the rainfall risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained.

[0225] In an exemplary embodiment, the result determination module 1703 is further configured to obtain a target correspondence; the target correspondence is used to characterize the correspondence between the slope risk assessment results, the rainfall risk assessment results, and the predicted risk assessment results; based on the target correspondence, and the slope risk assessment results and rainfall risk assessment results of the sample base area slopes output by each slope risk prediction model to be trained, the predicted risk assessment results of the sample base area slopes output by each slope risk prediction model to be trained are determined.

[0226] In an exemplary embodiment, the data acquisition module 1701 is further configured to acquire the base area slope identifier of the sample tower base area slope; acquire the slope data and rainfall data corresponding to the base area slope identifier; preprocess the slope data and rainfall data to obtain preprocessed slope data and preprocessed rainfall data; and use both the preprocessed slope data and preprocessed rainfall data as sample target data for the sample tower base area slope.

[0227] Each module in the aforementioned training device for the slope risk prediction model of the tower base area can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0228] Based on the same inventive concept, this application also provides a tower base area slope risk prediction device for implementing the above-mentioned tower base area slope risk prediction method. The solution provided by this device is similar to the solution described in the above-described method. Therefore, the specific limitations of one or more tower base area slope risk prediction device embodiments provided below can be found in the limitations of the tower base area slope risk prediction method described above, and will not be repeated here.

[0229] In one exemplary embodiment, such as Figure 18 As shown, a slope risk prediction device for a tower base area is provided, comprising: a target acquisition module 1801, a model prediction module 1802, and a result generation module 1803, wherein:

[0230] The target acquisition module 1801 is used to acquire target data of the slope of the tower base area to be analyzed; the target data includes at least the slope data and rainfall data of the slope of the tower base area to be analyzed.

[0231] The model prediction module 1802 is used to input the target data into the pre-trained slope risk prediction model to obtain the slope risk assessment results and rainfall risk assessment results of the slope in the base area to be analyzed; the pre-trained slope risk prediction model is trained by the base area slope risk prediction model training method.

[0232] The result generation module 1803 is used to determine the target risk assessment result of the slope in the area to be analyzed based on the slope risk assessment result and the rainfall risk assessment result of the slope in the area to be analyzed.

[0233] Each module in the aforementioned slope risk prediction device for the tower base area can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the computer device's memory as software, so that the processor can call and execute the corresponding operations of each module.

[0234] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 19 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs in the non-volatile storage media to run. The database stores slope data and rainfall data, among other things. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a training method for a slope risk prediction model in a tower base area.

[0235] Those skilled in the art will understand that Figure 19 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0236] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0237] In one exemplary embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above-described method embodiments.

[0238] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0239] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile 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. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0240] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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 specification.

[0241] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A training method for a slope risk prediction model in a tower base area, characterized in that, The method includes: Obtain sample target data for the slope of the sample tower base area; the sample target data includes at least slope data and rainfall data for the slope of the sample tower base area; The sample target data is input into multiple slope risk prediction models to be trained, and the slope risk assessment results and rainfall risk assessment results of the sample tower base area slope are obtained from the output of each slope risk prediction model to be trained. Obtain the target correspondence; the target correspondence is used to characterize the correspondence between the slope risk assessment results, rainfall risk assessment results and predicted risk assessment results; based on the target correspondence, and the slope risk assessment results and rainfall risk assessment results of the sample tower base area slopes output by each of the slope risk prediction models to be trained, determine the predicted risk assessment results of the sample tower base area slopes output by each of the slope risk prediction models to be trained. Based on the predicted risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained, and the actual risk assessment results of the sample tower base area slope, the slope risk prediction models to be trained are iteratively trained to obtain multiple trained slope risk prediction models, and the model performance parameters corresponding to each trained slope risk prediction model. From the slope risk prediction models that have been trained, select the slope risk prediction models whose performance parameters meet the preset parameter conditions and use them as pre-trained slope risk prediction models. The step of inputting the sample target data into multiple slope risk prediction models to obtain slope risk assessment results and rainfall risk assessment results for the sample base area slope output by each of the trained slope risk prediction models includes: inputting the slope data of the sample base area slope into multiple slope risk prediction models to be trained; performing risk prediction processing on the slope data of the sample base area slope through each of the trained slope risk prediction models to obtain a first prediction probability corresponding to each preset slope risk assessment result output by each of the trained slope risk prediction models; and inputting the rainfall data of the sample base area slope into each of the trained slope risk prediction models to obtain a first prediction probability corresponding to each preset slope risk assessment result output by each of the trained slope risk prediction models. The slope risk prediction model performs risk prediction processing on the rainfall data of the sample tower base area slope to obtain the second prediction probability corresponding to each preset rainfall risk assessment result output by each slope risk prediction model to be trained; the preset slope risk assessment result corresponding to the maximum value of the first prediction probability output by each slope risk prediction model to be trained is taken as the slope risk assessment result of the sample tower base area slope output by each slope risk prediction model to be trained, and the preset rainfall risk assessment result corresponding to the maximum value of the second prediction probability output by each slope risk prediction model to be trained is taken as the rainfall risk assessment result of the sample tower base area slope output by each slope risk prediction model to be trained.

2. The method according to claim 1, characterized in that, The step of inputting the sample target data into multiple slope risk prediction models to be trained, and obtaining the slope risk assessment results and rainfall risk assessment results of the sample tower base area slope output by each of the slope risk prediction models to be trained, further includes: Feature extraction processing is performed on the slope data of the sample tower base area slope to obtain the first feature vector corresponding to the slope data of the sample tower base area slope; Feature extraction processing is performed on the rainfall data of the slope in the sample tower base area to obtain the second feature vector corresponding to the rainfall data of the slope in the sample tower base area; The first feature vector is input into multiple slope risk prediction models to be trained, respectively, to obtain the slope risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained, and the second feature vector is input into each slope risk prediction model to be trained, respectively, to obtain the rainfall risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained.

3. The method according to claim 1, characterized in that, The acquisition of sample target data for the slope of the sample tower base area includes: Obtain the slope identifier of the base area of ​​the sample tower; Obtain the slope data and rainfall data corresponding to the slope markers in the tower base area; The slope data and rainfall data are preprocessed to obtain preprocessed slope data and preprocessed rainfall data; Both the preprocessed slope data and the preprocessed rainfall data are used as the target sample data for the slope in the sample tower base area.

4. A method for predicting slope risk in a tower base area, characterized in that, The method includes: Obtain target data for the slope of the tower base area to be analyzed; the target data includes at least slope data and rainfall data for the slope of the tower base area to be analyzed. The target data is input into a pre-trained slope risk prediction model to obtain the slope risk assessment results and rainfall risk assessment results of the slope in the base area to be analyzed; the pre-trained slope risk prediction model is trained by the method described in any one of claims 1 to 3. Based on the slope risk assessment results and rainfall risk assessment results of the slope in the tower base area to be analyzed, the target risk assessment result of the slope in the tower base area to be analyzed is determined; the target risk assessment result refers to the comprehensive risk assessment result of the slope in the tower base area to be analyzed output by the pre-trained slope risk prediction model.

5. A training device for a slope risk prediction model in a tower base area, characterized in that, The device includes: The data acquisition module is used to acquire sample target data of the slope in the sample tower base area; the sample target data includes at least slope data and rainfall data of the slope in the sample tower base area. The result prediction module is used to input the sample target data into multiple slope risk prediction models to be trained, and to obtain the slope risk assessment results and rainfall risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained. The result determination module is used to obtain the target correspondence; the target correspondence is used to characterize the correspondence between the slope risk assessment results, the rainfall risk assessment results and the predicted risk assessment results; based on the target correspondence, and the slope risk assessment results and rainfall risk assessment results of the sample tower base area slopes output by each of the slope risk prediction models to be trained, the predicted risk assessment results of the sample tower base area slopes output by each of the slope risk prediction models to be trained are determined. The model training module is used to iteratively train each slope risk prediction model to be trained based on the predicted risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained, and the actual risk assessment results of the sample tower base area slope, to obtain multiple trained slope risk prediction models, and the model performance parameters corresponding to each trained slope risk prediction model. The model determination module is used to select slope risk prediction models whose performance parameters meet the preset parameter conditions from the slope risk prediction models that have been trained, and use them as pre-trained slope risk prediction models. The result prediction module is further configured to input the slope data of the sample tower base area slope into multiple slope risk prediction models to be trained, and to perform risk prediction processing on the slope data of the sample tower base area slope through each of the slope risk prediction models to be trained, thereby obtaining the first prediction probability corresponding to each preset slope risk assessment result output by each of the slope risk prediction models to be trained; and to input the rainfall data of the sample tower base area slope into each of the slope risk prediction models to be trained, and to perform risk prediction processing on the rainfall data of the sample tower base area slope through each of the slope risk prediction models to be trained, thereby obtaining the first prediction probability corresponding to each preset slope risk assessment result output by ... preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output by each preset slope risk assessment result output The second prediction probability corresponding to each preset rainfall risk assessment result output by each slope risk prediction model to be trained is used as the preset slope risk assessment result corresponding to the maximum value of the first prediction probability output by each slope risk prediction model to be trained, and the preset rainfall risk assessment result corresponding to the maximum value of the second prediction probability output by each slope risk prediction model to be trained is used as the rainfall risk assessment result of the sample tower base area slope output by each slope risk prediction model to be trained.

6. The apparatus according to claim 5, characterized in that, The result prediction module is further configured to perform feature extraction processing on the slope data of the sample tower base area slope to obtain a first feature vector corresponding to the slope data of the sample tower base area slope; perform feature extraction processing on the rainfall data of the sample tower base area slope to obtain a second feature vector corresponding to the rainfall data of the sample tower base area slope; input the first feature vector into multiple slope risk prediction models to be trained to obtain the slope risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained; and input the second feature vector into each slope risk prediction model to be trained to obtain the rainfall risk assessment results of the sample tower base area slope output by each slope risk prediction model to be trained.

7. The apparatus according to claim 5, characterized in that, The data acquisition module is also used to acquire the base area slope identifier of the sample tower base area slope; acquire the slope data and rainfall data corresponding to the base area slope identifier; and preprocess the slope data and rainfall data to obtain preprocessed slope data and preprocessed rainfall data. Both the preprocessed slope data and the preprocessed rainfall data are used as the target sample data for the slope in the sample tower base area.

8. A device for predicting slope risk in a tower base area, characterized in that, The device includes: The target acquisition module is used to acquire target data of the slope of the tower base area to be analyzed; the target data includes at least slope data and rainfall data of the slope of the tower base area to be analyzed. The model prediction module is used to input the target data into a pre-trained slope risk prediction model to obtain the slope risk assessment results and rainfall risk assessment results of the slope in the base area to be analyzed; the pre-trained slope risk prediction model is trained by the method of any one of claims 1 to 3; The result generation module is used to determine the target risk assessment result of the slope in the tower base area to be analyzed based on the slope risk assessment result and the rainfall risk assessment result of the slope in the tower base area to be analyzed; the target risk assessment result refers to the comprehensive risk assessment result of the slope in the tower base area to be analyzed output by the pre-trained slope risk prediction model.

9. 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, it implements the steps of the method according to any one of claims 1 to 4.

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

Images