An intelligent early warning method, device and electronic equipment

Abstract

The embodiment of the present application provides a kind of intelligent early warning method, device and electronic equipment, can select early warning parameter from monitoring parameter, and early warning parameter is preprocessed, clustering, then early warning parameter is combined into different parameter group, as training sample data, finally, early warning model is trained based on training sample data.This embodiment, select multiple early warning parameters, early warning parameter is clustered, so that early warning parameter can be divided into different early warning levels, finally, different early warning parameters are combined into different training samples to train early warning model, so that the early warning model obtained can combine multiple different early warning parameters to warn landslide disaster based on the same standard, realize that multiple parameters are simultaneously warned, improve the early warning efficiency, and there is no different early warning threshold condition, to improve the accuracy of early warning.

Description

Technical Field

[0001] This invention relates to the field of geological disaster monitoring and early warning technology, and in particular to an intelligent early warning method, device and electronic equipment. Background Technology

[0002] my country has vast mountainous areas, and the rainy season is a peak period for natural disasters in these regions. Landslides are characterized by their rapid onset, large destructive range, and wide distribution. In recent years, to reduce casualties and property losses caused by landslides, monitoring equipment has been installed at appropriate locations on the mountain surface. The collected monitoring data is then used to issue early warnings for landslides. Most existing early warning methods involve collecting parameters such as displacement, cracks, and dip angles at different monitoring points, setting reasonable thresholds, establishing a correlation between monitoring parameters and landslide occurrence levels, and issuing corresponding warnings based on the assessment of the warning levels. This allows for early detection and forecasting of landslides. However, this method relies heavily on the relationship between parameters and thresholds, resulting in low accuracy in early warnings.

[0003] Therefore, how to provide accurate early warnings is an urgent problem to be solved. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent early warning method, device, and electronic device to solve the problem of how to improve the accuracy of landslide disaster early warning. The specific technical solution is as follows:

[0005] In a first aspect of this application, an intelligent early warning method is provided, the method comprising:

[0006] Multiple early warning parameters are determined from the monitoring parameters;

[0007] The K-means clustering algorithm is used to cluster the warning parameter data for each of the multiple warning parameters to obtain clustered data clusters, and the warning level of each clustered data cluster is determined; wherein, the number of centroids in the K-means clustering algorithm is 4, and the warning levels of the warning parameters are: no warning, blue warning, yellow warning, orange warning, and red warning.

[0008] Training sample data is determined from the clustered data clusters; wherein, the warning level in each training sample data is determined based on the highest warning level of the data cluster in that training sample data;

[0009] An early warning model for landslide disasters was trained using multiple training sample data.

[0010] In one possible implementation, determining multiple early warning parameters from the monitoring parameters includes:

[0011] Parameters whose values ​​change more than a preset threshold within a unit of time are obtained from the monitoring parameters and used as early warning parameters.

[0012] In one possible implementation, before clustering the multiple warning parameter data based on data similarity to obtain clustered data clusters, the method further includes:

[0013] The warning parameter data is preprocessed to obtain the processed warning parameter data.

[0014] In one possible implementation, the preprocessing of the warning parameter data to obtain processed warning parameter data includes:

[0015] The warning parameter data is filtered to obtain the noise-reduced warning parameters.

[0016] The denoised warning parameter data is normalized to obtain the processed warning parameter data.

[0017] In one possible implementation, determining the training sample data from the clustered data clusters includes:

[0018] The clustered data clusters are combined into parameter groups according to different parameter types, and the parameter groups are used as training sample data.

[0019] In one possible implementation, combining the clustered data clusters into parameter groups according to different parameter types includes:

[0020] Combine data clusters of a single parameter type into parameter groups;

[0021] And / or,

[0022] Combine data clusters of two parameter types into a parameter group;

[0023] And / or,

[0024] Combine data clusters of three parameter types into parameter groups.

[0025] In a second aspect of this application, an intelligent early warning device is provided, the device comprising:

[0026] The determination module is used to determine multiple early warning parameters from the monitoring parameters;

[0027] The clustering module is used to cluster the multiple warning parameter data using the K-means clustering algorithm to obtain clustered data clusters and determine the warning level of each clustered data cluster; wherein, the K-means clustering algorithm has a centroid number of 4, and the warning levels of the warning parameters are: no warning, blue warning, yellow warning, orange warning, and red warning;

[0028] A combination module is used to determine training sample data from the clustered data clusters; wherein, the warning level in each training sample data is determined based on the highest warning level of the data cluster in that training sample data;

[0029] The training module uses multiple training sample data to train an early warning model for landslide disaster warning.

[0030] In one possible implementation, the determining module determines multiple early warning parameters from the monitoring parameters, including:

[0031] Parameters whose values ​​change more than a preset threshold within a unit of time are obtained from the monitoring parameters and used as early warning parameters;

[0032] The device further includes: a preprocessing module, used to preprocess the warning parameter data to obtain processed warning parameter data;

[0033] The preprocessing module preprocesses the warning parameter data to obtain processed warning parameter data, including:

[0034] The warning parameters are filtered to obtain the noise-reduced warning parameters;

[0035] The denoised warning parameters are normalized to obtain the processed warning parameter data.

[0036] The combination module determines training sample data from the clustered data clusters, including:

[0037] The clustered data clusters are combined into parameter groups according to different parameter types, and the parameter groups are used as training sample data.

[0038] The step of combining clustered data clusters into parameter groups according to different parameter types includes:

[0039] Combine data clusters of a single parameter type into parameter groups;

[0040] And / or,

[0041] Combine data clusters of two parameter types into a parameter group;

[0042] And / or,

[0043] Combine data clusters of three parameter types into parameter groups.

[0044] In a third aspect of this application, an electronic device is provided, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0045] Memory, used to store computer programs;

[0046] When a processor executes a program stored in memory, it implements any of the steps described in the first aspect above.

[0047] In a fourth aspect of this application, a computer-readable storage medium is provided, wherein a computer program is stored therein, and when executed by a processor, the computer program implements the steps of any of the methods described in the first aspect above.

[0048] Beneficial effects of the embodiments of the present invention:

[0049] This invention provides an intelligent early warning method, device, and electronic device. It selects early warning parameters from monitoring parameters, preprocesses and clusters these parameters, combines them into different parameter groups as training sample data, and finally trains an early warning model based on this training sample data. In this embodiment, multiple early warning parameters are selected, and the K-means clustering algorithm is used to cluster the early warning parameter data separately. This divides the early warning parameters into different early warning levels. Finally, different early warning parameters are combined into different training samples to train an early warning model. This early warning model can combine multiple different early warning parameters based on the same standard to issue early warnings for landslide disasters, achieving simultaneous early warning for multiple parameters, improving early warning efficiency, and eliminating the possibility of different early warning thresholds. The K-means clustering method can adaptively classify landslide early warning thresholds, obtaining accurate early warning threshold levels, which helps reduce the false alarm rate of landslide disaster threshold early warnings, thereby improving the accuracy of early warnings.

[0050] Of course, implementing any product or method of the present invention does not necessarily require achieving all of the advantages described above at the same time. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these drawings.

[0052] Figure 1This is a schematic flowchart of the first intelligent early warning method provided in the embodiments of this application;

[0053] Figure 2-1 This application provides a schematic diagram of processed early warning parameter data obtained after careful preprocessing of GNSS horizontal displacement, as an embodiment of the present application.

[0054] Figure 2-2 A schematic diagram of the centroid position provided in an embodiment of this application;

[0055] Figure 2-3 This application provides a schematic diagram of the result of clustering GNSS horizontal displacement data using the K-means clustering algorithm, as shown in the embodiment of the present application.

[0056] Figure 3-1 This is a schematic diagram of the second intelligent early warning method provided in the embodiments of this application;

[0057] Figure 3-2 This application provides a schematic diagram of a preprocessing flow for early warning parameter data.

[0058] Figure 3-3 This application provides an embodiment of a filtering method that uses smoothed filtering, linear regression filtering, and wavelet denoising to filter some data, and the schematic diagram shows the processing effect.

[0059] Figure 3-4 This application provides a schematic diagram of the GNSS horizontal displacement of different landslide points before preprocessing, as shown in an embodiment of the present application.

[0060] Figure 3-5 A GNSS horizontal displacement diagram of a corresponding landslide point after preprocessing, provided for an embodiment of this application;

[0061] Figure 4 This is a schematic diagram of the third intelligent early warning method provided in the embodiments of this application;

[0062] Figure 5 A schematic diagram of an intelligent early warning device provided in an embodiment of this application;

[0063] Figure 6 This is a schematic diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0064] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. The described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art based on this application are within the scope of protection of the present invention.

[0065] This application proposes an intelligent early warning method, such as... Figure 1 As shown, the method includes:

[0066] S101. Determine multiple early warning parameters from the monitoring parameters.

[0067] S102. The K-means clustering algorithm is used to cluster the warning parameter data for multiple warning parameters respectively to obtain the clustered data clusters, and the warning level of each clustered data cluster is determined; wherein, the number of centroids of the K-means clustering algorithm is 4, and the warning level of the warning parameter is: no warning, blue warning, yellow warning, orange warning and red warning.

[0068] S103. Determine training sample data from the clustered data clusters; wherein, the warning level in each training sample data is determined based on the highest warning level of the data cluster in that training sample data.

[0069] S104. A warning model for landslide disasters is trained using multiple training sample data.

[0070] In this embodiment, multiple early warning parameters are selected, and the K-means clustering algorithm is used to cluster the early warning parameter data separately. This allows the early warning parameters to be divided into different early warning levels. Finally, the different early warning parameters are combined into different training samples to train the early warning model. The resulting early warning model can combine multiple different early warning parameters to issue early warnings for landslide disasters based on the same standard, realizing simultaneous early warning for multiple parameters, improving early warning efficiency, and eliminating the situation of different early warning thresholds. The K-means clustering method can adaptively divide the landslide early warning threshold to obtain accurate early warning threshold levels, which helps to reduce the false alarm rate of landslide disaster threshold early warnings and thus improve the accuracy of early warnings.

[0071] The above S101-S105 are described in detail below.

[0072] In S101, during the monitoring of landslide disasters, the monitoring parameters are divided into: displacement parameters, tilt parameters, and other influencing parameters. Displacement parameters include: GNSS (Global Navigation Satellite System) horizontal displacement, GNSS vertical displacement, and depth displacement, etc. Tilt parameters include: tilt X component, tilt Y component, tilt Z component, tilt angle (angle) value, tilt AZI (azimuth) value, etc. Other influencing parameters include: acceleration, cracks, water content, and rainfall, etc. The monitoring parameters in this application embodiment include, but are not limited to, the above-mentioned parameters.

[0073] The warning parameters are selected from the monitoring parameters for issuing early warnings about landslides. There must be at least two warning parameters. For example, warning parameters may include: GNSS horizontal displacement, dip angle value, and crack value.

[0074] In the process of landslide disaster monitoring, there are many monitoring parameters. Some of these parameters do not change in a short period of time. In order to improve monitoring efficiency, parameters that can more clearly characterize the landslide occurrence process can be selected.

[0075] In one possible embodiment, S101 above includes:

[0076] S1011. Obtain parameters whose values ​​change more than a preset threshold within a unit of time from the monitoring parameters, and use them as early warning parameters.

[0077] In this step, the preset threshold is determined by those skilled in the art based on the magnitude of parameter changes per unit time during monitoring. When the parameter value changes more than the preset threshold per unit time, it indicates that the parameter can clearly characterize the landslide occurrence process; when the parameter value changes less than the preset threshold per unit time, it indicates that the parameter cannot clearly characterize the landslide occurrence process.

[0078] In this embodiment, by selecting early warning parameters that change more than a preset threshold within a unit time, i.e., selecting monitoring parameters that can clearly characterize the landslide occurrence process as early warning parameters, since the early warning parameters can clearly characterize the landslide occurrence process, it is possible to more clearly determine whether the early warning is accurate, thereby improving the training accuracy of the subsequent early warning model.

[0079] In step S102, the warning parameter data is clustered into different clusters, with each cluster corresponding to a different warning level. The warning level varies depending on the specific warning parameter and the clusters used for classification.

[0080] For example, clustering GNSS horizontal displacement data into clusters 1, 2, 3, and 4, all data within these clusters represent all data corresponding to GNSS horizontal displacement. The GNSS horizontal displacement data included in each cluster differs, and the corresponding warning levels are different. For instance, cluster 1 corresponds to no warning, cluster 2 to a blue warning, cluster 3 to a yellow warning, and cluster 4 to an orange warning. Similarly, clustering tilt angle data into clusters 5, 6, and 7, all data within these clusters represent all data corresponding to tilt angle. The tilt angle data included in each cluster differs, and the corresponding warning levels are different. For instance, cluster 5 corresponds to no warning, cluster 6 to a blue warning, and cluster 7 to a yellow warning.

[0081] The k-means clustering algorithm uses an iterative approach to minimize error. It randomly sets the initial number and location of centroids as cluster centers, iteratively calculates the distance between all cluster objects and their centroid locations, and assigns each cluster object to the nearest centroid. Once all clusters have been completed, the distance to the centroid of each cluster is recalculated based on the clustering results. This iterative process continues until one of the following termination conditions is met:

[0082] (1) No objects were reallocated.

[0083] (2) The cluster centers no longer change.

[0084] (3) The sum of squared errors of the clustered objects reaches a local minimum.

[0085] For the application scenario in this application embodiment, the K-means clustering algorithm achieves the best clustering effect and the most accurate warning level when the number of centroids is 4. The four centroid positions after clustering correspond to four preset warning thresholds, which divide the warning parameters into five different warning levels: no warning, blue warning, yellow warning, orange warning, and red warning. The priority of the warning levels is: no warning < blue warning < yellow warning < orange warning < red warning, meaning the red warning level is the highest.

[0086] For example, taking the aforementioned GNSS horizontal displacement as an example, such as Figure 2-1 The figure shows the preprocessed warning parameter data obtained after preprocessing the GNSS horizontal displacement. The horizontal axis represents the sampling points, and the vertical axis represents the normalized horizontal displacement. To include as many sampling points as possible, the unit of each sampling point is 10. 5 .like Figure 2-2 The diagram shows the correspondence between the number of centroids and the cluster centroid values. Four centroids were selected, with values ​​of [0.1240, 0.3359, 0.5587, 0.7621], which represent the landslide warning threshold. Based on the centroid locations obtained from clustering, all horizontal displacement samples are clustered according to the following formula:

[0087]

[0088] Where K is the position of the centroid.

[0089] The results are as follows Figure 2-3 As shown, the horizontal axis represents the sampling points, and the unit of each sampling point is 10. 5The vertical axis represents the warning level. When the GNSS horizontal displacement sampling value falls into different numerical ranges, the corresponding threshold level warning is triggered, thereby enabling accurate threshold prediction and early warning for landslide monitoring.

[0090] In this embodiment, the K-means clustering algorithm is used to cluster the warning parameter data separately, with a centroid count of 4. This allows for a more reasonable clustering of the various warning parameter data, ensuring that each cluster corresponds to a different warning level. The warning levels are further divided into no warning, blue warning, yellow warning, orange warning, and red warning. Compared to existing technologies that only offer red, orange, and yellow warnings, this provides a more detailed classification of warning levels, enabling subsequent technical personnel to take more detailed and appropriate countermeasures. Using the K-means clustering method allows for adaptive classification of landslide warning thresholds, resulting in accurate warning threshold levels and helping to reduce the false alarm rate of landslide disaster threshold warnings.

[0091] In S103, each training sample data may include warning parameter data corresponding to different parameter types, and the warning parameter data may be data from different clusters.

[0092] When the training sample data includes warning parameter data from different clusters, the warning level of that parameter group is determined based on the highest warning level of the data cluster in the training sample data.

[0093] For example, taking the aforementioned GNSS horizontal displacement as an example, the warning parameters also include: tilt angle. The tilt angle data is clustered, assuming it is clustered into cluster 5, cluster 6, and cluster 7. Cluster 5 corresponds to no warning, cluster 6 corresponds to blue warning, and cluster 7 corresponds to yellow warning. Parameter group 1 includes: cluster 4 and cluster 5. Orange warning > no warning, so the warning level corresponding to the parameter group is orange warning.

[0094] In one possible embodiment, S104 above includes:

[0095] S1041. Combine the clustered data clusters into parameter groups according to different parameter types, and use the parameter groups as training sample data.

[0096] In this step, warning parameter data of different parameter types can be grouped into the same parameter group, and data in different parameter groups can be repeated.

[0097] In this embodiment, by combining the clustered early warning parameter data into different parameter groups according to parameter type, the obtained training sample data includes early warning parameter data of different parameter types, which enriches the training sample data and can improve the accuracy of training the early warning model.

[0098] In one possible embodiment, S1041 above includes:

[0099] S1041a. Combine data clusters of a single parameter type into a parameter group;

[0100] And / or,

[0101] Combine data clusters of two parameter types into a parameter group;

[0102] And / or,

[0103] Combine data clusters of three parameter types into parameter groups.

[0104] Specifically, the parameter groups include the following seven types: (1) a parameter group that includes a data cluster with a single parameter type; (2) a parameter group that includes a data cluster with two parameter types; (3) a parameter group that includes a data cluster with three parameter types; (4) a parameter group that includes a data cluster with a single parameter type and a data cluster with two parameter types; (5) a parameter group that includes a data cluster with a single parameter type and a data cluster with three parameter types; (6) a parameter group that includes a data cluster with two parameter types and a data cluster with three parameter types; and (7) a parameter group that includes a data cluster with a single parameter type, a data cluster with two parameter types, and a data cluster with three parameter types.

[0105] In this step, data clusters of single parameter type, two parameter types, and three parameter types are combined into parameter groups. When combining single parameter type and two parameter types, missing parameter data is padded with zeros.

[0106] In this embodiment, by combining data clusters of a single parameter type, and / or two parameter types, and / or three parameter types into parameter groups, the obtained training sample data is made as rich as possible, including all cases, which can improve the accuracy of the model when training the early warning model.

[0107] In S104, when training the early warning model using training sample data, the early warning model can be any type of neural network model. During training, maximizing the amount of training sample data can improve the accuracy of the early warning model.

[0108] For example, assuming the early warning model is a BP neural network model, the prediction model adopts a three-layer BP network structure, namely an input layer, a hidden layer and an output layer.

[0109] The training process for an early warning model may include:

[0110] (1) Obtain the warning parameter data from the successful landslide forecast and warning as the training sample dataset, and label the sample data into 5 categories according to the landslide warning level, namely no warning, blue warning, yellow warning, orange warning and red warning; construct a BP neural network using MATLAB, with the input being the 3 selected features, the number of neurons in the input layer of the BP is 10, the number of neurons in the hidden layer is 10, and the number of neurons in the output layer is 1. The output data contains 5 categories, namely no warning is "1", blue warning is "2", yellow warning is "3", orange warning is "4" and red warning is "5".

[0111] (2) Randomly select 90% of the samples from the training sample dataset as the training set to train the BP neural network; use the remaining 10% of the training sample data as the test set, input it into the already trained network, and calculate the accuracy between the output value and the true value.

[0112] (3) If the accuracy does not meet the requirements, further adjust the neural network parameters until the requirements are met.

[0113] In one possible embodiment, this application also provides an early warning method, such as... Figure 3-1 As shown, the method includes:

[0114] S301. Determine multiple early warning parameters from the monitoring parameters.

[0115] This step is the same as S101 above, and will not be described in detail here.

[0116] S302. Preprocess the early warning parameter data to obtain the processed early warning parameter data.

[0117] In this step, preprocessing the warning parameter data can be done by removing some invalid data or by standardizing the warning parameters. The specific preprocessing method can be set by those skilled in the art based on the actual application scenario.

[0118] Once the warning parameters are determined, the corresponding warning parameter data is retrieved from the database. All monitoring parameter data are stored in the database and can be retrieved when needed. Therefore, once the warning parameters are determined, the corresponding warning parameter data can be retrieved from the database.

[0119] S303. Use the K-means clustering algorithm to cluster the warning parameter data for multiple warning parameters respectively, obtain the clustered data clusters, and determine the warning level of each clustered data cluster.

[0120] This step is the same as S102 mentioned above, and will not be described in detail here.

[0121] S304. Determine training sample data from the clustered data clusters; wherein, the warning level in each training sample data is determined based on the highest warning level of the data cluster in that training sample data.

[0122] This step is the same as S103 mentioned above, and will not be described in detail here.

[0123] S305. A warning model for landslide disasters is trained using multiple training sample data.

[0124] This step is the same as S104 above, and will not be described in detail here.

[0125] In this embodiment, the warning parameter data is preprocessed to obtain preprocessed data. The preprocessed data can be directly processed without subsequent format conversion, thereby improving the efficiency of data processing.

[0126] Due to the influence of landslide morphology, equipment installation location, and environmental factors, the variation range of horizontal displacement collected by different GNSS devices varies significantly. Therefore, before classifying threshold levels, the data must be normalized to ensure that the horizontal displacement parameters collected by different devices are within the same range of variation, allowing the subsequent training model to converge better. Figure 3-2 As shown, the above S302 may include:

[0127] S3021. Filter the warning parameter data to obtain the noise-reduced warning parameter data.

[0128] In this step, different filtering methods can be used to denoise portions of the data separately. Then, the filtering method with the best processing effect is selected to filter all the early warning parameter data. Filtering methods include, but are not limited to: smoothing filtering, linear regression filtering, and wavelet denoising.

[0129] Taking GNSS horizontal displacement as an example, smoothing filtering, linear regression filtering, and wavelet denoising are used to filter some data. The processing results are as follows: Figure 3-3 As shown in the figure, the horizontal axis represents the sampling points, and the vertical axis represents the horizontal displacement, with the unit of horizontal displacement being mm. After comparative analysis, the smoothed filtering algorithm was selected to filter the early warning parameter data. The smoothed filtering method showed better noise suppression effect on landslide horizontal displacement data than wavelet denoising and linear regression denoising methods.

[0130] S3022. Normalize the noise-reduced warning parameter data to obtain the processed warning parameter data.

[0131] The noise-reduced warning parameter data can be normalized according to the normalization standard calculation formula.

[0132] The standardized calculation formula is:

[0133]

[0134] Where x is the warning parameter, min(x) is the minimum value of the parameter, and max(x) is the maximum value of the parameter.

[0135] The standardized calculation formula can also be:

[0136]

[0137] Where x is the warning parameter, mean(x) is the mean of the warning parameter, and std(x) is the variance of the warning parameter.

[0138] Taking GNSS horizontal displacement as an example, the GNSS horizontal displacements at different landslide points before normalization are as follows: Figure 3-4 As shown, the horizontal axis represents the sampling point, and the vertical axis represents the horizontal displacement, with the unit of horizontal displacement being mm. The GNSS horizontal displacement of the corresponding landslide point after normalization is shown below. Figure 3-5 As shown, all values ​​are between (0, 1), where the horizontal axis represents the sampling point and the vertical axis represents the normalized horizontal displacement.

[0139] In this embodiment, filtering the warning parameter data removes interfering data, resulting in relatively clean data and improving the accuracy of subsequent data processing. Normalizing the denoised warning parameter data maps it to the (0, 1) range, making data processing more convenient and faster. Furthermore, it allows the warning model to converge well during training.

[0140] In one possible embodiment, this application also provides an early warning system, such as... Figure 4 As shown, the method includes:

[0141] S401. Determine multiple early warning parameters from the monitoring parameters.

[0142] This step is the same as S101 above, and will not be described in detail here.

[0143] S402. Use the K-means clustering algorithm to cluster the warning parameter data for multiple warning parameters respectively, obtain the clustered data clusters, and determine the warning level of each clustered data cluster.

[0144] This step is the same as S102 mentioned above, and will not be described in detail here.

[0145] S403. Determine training sample data from the clustered data clusters; wherein, the warning level in each training sample data is determined based on the highest warning level of the data cluster in that training sample data.

[0146] This step is the same as S103 mentioned above, and will not be described in detail here.

[0147] S404. A warning model for landslide disasters is trained using multiple training sample data.

[0148] This step is the same as S104 above, and will not be described in detail here.

[0149] S405. Encapsulate the early warning model and deploy it on the landslide disaster early warning platform so that the platform can call the early warning model to issue early warnings for landslide disasters.

[0150] In this step, after the trained early warning model is packaged, a function call window is created. When early warning parameters are input, the early warning model is automatically called to directly obtain the early warning level under those parameters.

[0151] In this embodiment, by directly deploying the obtained early warning model to the landslide disaster early warning platform, technicians can directly call the early warning model when using the landslide disaster early warning platform, and then directly issue early warnings based on the input multiple parameters, and obtain the early warning level under the early warning parameters, thereby improving the efficiency of early warning.

[0152] This application also provides an intelligent early warning device, such as... Figure 5 As shown, the device includes:

[0153] The determination module 501 is used to determine multiple early warning parameters from the monitoring parameters;

[0154] Clustering module 502 is used to cluster the multiple warning parameter data using the K-means clustering algorithm to obtain clustered data clusters and determine the warning level of each clustered data cluster; wherein, the K-means clustering algorithm has a centroid number of 4, and the warning levels of the warning parameters are: no warning, blue warning, yellow warning, orange warning and red warning.

[0155] The combination module 503 is used to determine training sample data from the clustered data clusters; wherein, the warning level in each training sample data is determined based on the highest warning level of the data cluster in that training sample data;

[0156] Training module 504 uses multiple training sample data to train an early warning model for landslide disaster warning.

[0157] In one possible implementation, the determining module 501 determines multiple early warning parameters from the monitoring parameters, including:

[0158] Parameters whose values ​​change more than a preset threshold within a unit of time are obtained from the monitoring parameters and used as early warning parameters;

[0159] The device further includes: a preprocessing module 505, used to preprocess the warning parameter data to obtain processed warning parameter data;

[0160] The preprocessing module 505 preprocesses the warning parameter data to obtain processed warning parameter data, including:

[0161] The warning parameters are filtered to obtain the noise-reduced warning parameters;

[0162] The denoised warning parameters are normalized to obtain the processed warning parameter data.

[0163] The combination module 503 determines training sample data from the clustered data clusters, including:

[0164] The clustered data clusters are combined into parameter groups according to different parameter types, and the parameter groups are used as training sample data.

[0165] The step of combining clustered data clusters into parameter groups according to different parameter types includes:

[0166] Combine data clusters of a single parameter type into parameter groups;

[0167] And / or,

[0168] Combine data clusters of two parameter types into a parameter group;

[0169] And / or,

[0170] Combine data clusters of three parameter types into parameter groups.

[0171] This invention also provides an electronic device, such as... Figure 6 As shown, it includes a processor 601, a communication interface 602, a memory 603, and a communication bus 604, wherein the processor 601, the communication interface 602, and the memory 603 communicate with each other through the communication bus 604.

[0172] Memory 603 is used to store computer programs;

[0173] When processor 601 executes a program stored in memory 603, it performs the following steps:

[0174] Multiple early warning parameters are determined from the monitoring parameters;

[0175] The multiple warning parameter data are clustered based on data similarity to obtain clustered data clusters, and the warning level of each clustered data cluster is determined.

[0176] Training sample data is determined from the clustered data clusters; wherein, the warning level in each training sample data is determined based on the highest warning level of the data cluster in that training sample data;

[0177] An early warning model for landslide disasters was trained using multiple training sample data.

[0178] The communication bus mentioned in the above electronic devices can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0179] The communication interface is used for communication between the aforementioned electronic devices and other devices.

[0180] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0181] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0182] In another embodiment of the present invention, a computer-readable storage medium is also provided, which stores a computer program that, when executed by a processor, implements the steps of any of the above-described multi-parameter intelligent early warning methods.

[0183] In another embodiment of the present invention, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute any of the multi-parameter intelligent early warning methods described in the above embodiments.

[0184] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid state disk (SSD)).

[0185] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0186] The various embodiments in this specification are described in a related manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0187] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A smart early warning method, characterized in that, The method includes: Multiple early warning parameters are determined from the monitoring parameters, and these early warning parameters are used to provide early warning of landslides; The K-means clustering algorithm is used to cluster the warning parameter data for each of the multiple warning parameters to obtain clustered data clusters, and the warning level of each clustered data cluster is determined; wherein, the number of centroids in the K-means clustering algorithm is 4, and the warning levels of the warning parameters are: no warning, blue warning, yellow warning, orange warning, and red warning. Determining training sample data from the clustered data clusters includes: combining the clustered data clusters into parameter groups according to different parameter types, and using the parameter groups as training sample data; wherein, the warning level in each training sample data is determined based on the highest warning level of the data cluster in that training sample data; An early warning model for landslide disasters was trained using multiple training sample data. The step of determining multiple early warning parameters from the monitoring parameters includes: Parameters whose values ​​change more than a preset threshold within a unit of time are obtained from the monitoring parameters and used as early warning parameters.

2. The method of claim 1, wherein, Before clustering the multiple warning parameter data based on data similarity to obtain clustered data clusters, the method further includes: The warning parameter data is preprocessed to obtain the processed warning parameter data.

3. The method of claim 2, wherein, The preprocessing of the warning parameter data to obtain processed warning parameter data includes: The warning parameter data is filtered to obtain the noise-reduced warning parameter data; The denoised warning parameter data is normalized to obtain the processed warning parameter data.

4. The method according to claim 1, characterized in that, The step of combining clustered data clusters into parameter groups according to different parameter types includes: Combine data clusters of a single parameter type into parameter groups; And / or, Combine data clusters of two parameter types into a parameter group; And / or, Combine data clusters of three parameter types into parameter groups.

5. An intelligent early warning device, characterized in that, The device includes: A determination module is used to determine multiple early warning parameters from monitoring parameters, wherein the early warning parameters are parameters used to issue early warnings for landslides; The clustering module is used to cluster the multiple warning parameter data using the K-means clustering algorithm to obtain clustered data clusters and determine the warning level of each clustered data cluster; wherein, the K-means clustering algorithm has a centroid number of 4, and the warning levels of the warning parameters are: no warning, blue warning, yellow warning, orange warning, and red warning; A combination module is used to determine training sample data from the clustered data clusters; wherein, the warning level in each training sample data is determined based on the highest warning level of the data cluster in that training sample data; The training module uses multiple training sample data to train an early warning model for landslide disaster warning; The combination module determines training sample data from the clustered data clusters by: combining the clustered data clusters into parameter groups according to different parameter types, and using the parameter groups as training sample data. The determining module determines multiple early warning parameters from the monitoring parameters, including: Parameters whose values ​​change more than a preset threshold within a unit of time are obtained from the monitoring parameters and used as early warning parameters.

6. The apparatus according to claim 5, characterized in that, The device further includes: a preprocessing module, used to preprocess the warning parameter data to obtain processed warning parameter data; The preprocessing module preprocesses the warning parameter data to obtain processed warning parameter data, including: The warning parameters are filtered to obtain the noise-reduced warning parameters; The denoised warning parameters are normalized to obtain the processed warning parameter data. The step of combining clustered data clusters into parameter groups according to different parameter types includes: Combine data clusters of a single parameter type into parameter groups; And / or, Combine data clusters of two parameter types into a parameter group; And / or, Combine data clusters of three parameter types into parameter groups.

7. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, when executing a program stored in memory, implements the steps of the method described in any one of claims 1-4.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method described in any one of claims 1-4.

Images