A monolithic landslip warning system and method

Abstract

This invention relates to the field of geological monitoring technology and provides a single landslide early warning system and method. The system includes a data acquisition and cleaning module, an algorithm model library, a data analysis module, a data verification module, and an early warning module. The data acquisition and cleaning module processes monitoring data from multiple landslides. The algorithm model library stores algorithm models related to equipment, surface displacement, and rainfall. The data analysis module analyzes landslide alarms related to equipment, surface displacement, and rainfall involved in the causes of a single landslide based on the monitoring data and the algorithm model library to generate early warning information. The beneficial effects of this invention are: it can define specific alarm thresholds for different single landslides and predict landslide alarms through classification and rainfall prediction values, thereby achieving the effect of early warning.

Description

Technical Field

[0001] This invention belongs to the field of geological monitoring technology, and in particular relates to a single landslide early warning system and method. Background Technology

[0002] Currently, the main method for individual landslide warnings is real-time warning. Common methods include using expert experience to infer the approximate warning threshold for a region based on historical data, which is usually the size of a province or prefecture-level city; another method is to use rainfall to make coarse-grained divisions, such as the international practice of using rainfall-based landslide warning thresholds divided by country; or, in recent years, some scholars in China have been using the improved tangent angle method to calculate the threshold for individual landslides in Guizhou.

[0003] The applicant discovered through research on the aforementioned existing technologies that the first two methods have the drawback of excessively large alarm areas, and all three methods have the defect of being unable to provide early warnings. Summary of the Invention

[0004] The purpose of this invention is to provide a single landslide early warning system and method, which aims to solve the problems mentioned in the background art.

[0005] The present invention is implemented as follows: a single landslide early warning system, the system including a data acquisition and cleaning module, an algorithm model library, a data analysis module, a data verification module and an early warning module;

[0006] The data acquisition and cleaning module is used to process monitoring data from multiple landslides;

[0007] The algorithm model library is used to store algorithm models related to equipment, surface displacement, and rainfall.

[0008] The data analysis module is used to perform landslide alarm analysis on equipment, surface displacement and rainfall involved in the causes of individual landslides based on monitoring data and algorithm model library, so as to generate early warning information;

[0009] The data verification module is used to verify the algorithm models in the algorithm model library and dynamically adjust the algorithm verification cycle according to categories;

[0010] The early warning module is used to release early warning information to the public and to receive and process feedback information.

[0011] As a further aspect of the present invention, the data acquisition and cleaning module is specifically used for: equipment log acquisition, sensor data acquisition, and meteorological data synchronization; and also for cleaning the acquired and synchronized data.

[0012] The equipment log collection specifically includes: acquiring the operation logs of environmental data acquisition equipment deployed around a single landslide, wherein the operation logs are used to determine the operational health status of the environmental data acquisition equipment;

[0013] Sensor data acquisition specifically includes: acquiring relevant environmental data collected by environmental data acquisition devices deployed around a single landslide, including the ambient temperature, vertical surface displacement, horizontal surface displacement, displacement acceleration, and groundwater level.

[0014] Cleaning the collected and synchronized data includes: performing usability processing on the collected and synchronized data, which includes: removing dirty data, filling in missing data, and correcting unreasonable data.

[0015] As a further aspect of the present invention, the algorithm model library includes: displacement-related algorithm models and rainfall-related algorithm models. The algorithm model library supports continuous data input of displacement-related and rainfall-related data for individual landslides to output alarm thresholds for different parameters of individual landslides.

[0016] Displacement correlation algorithm model: This model is designed for non-rainfall type single landslides. The input data is the historical monitoring displacement correlation data of the single landslide. The first correlation algorithm calculates the variation law between the two sets of data: horizontal displacement, vertical displacement, displacement acceleration, tilt angle and first corresponding displacement. The first correlation algorithm includes classification, clustering, regression and association rule correlation algorithms in data mining. The first corresponding displacement includes: horizontal, vertical and resultant displacement.

[0017] Rainfall-related algorithm model: This model is designed for rainfall-induced single landslides. The input data is the historical monitoring data of water-related data of the single landslide. The model calculates the variation pattern between two sets of data: rainfall, water level change, soil moisture, water level change rate and second corresponding displacement through a second correlation algorithm. The second correlation algorithm includes classification, clustering, regression and association rule correlation algorithms in data mining. The second corresponding displacement includes horizontal, vertical and combined displacement.

[0018] As a further embodiment of the present invention, the data analysis module is used for: equipment reliability analysis, rainfall correlation analysis, surface displacement correlation analysis and hybrid analysis.

[0019] As a further embodiment of the present invention, the data verification module is used for: landslide type identification, landslide stage identification, landslide main influencing factor identification, data acquisition cycle identification, and landslide displacement verification.

[0020] As a further aspect of the present invention, in the data analysis module, the equipment reliability analysis includes: performing confidence assessment analysis on the equipment failure rate and the collected failure data rate by monitoring equipment logs; and performing multiple data cleaning and verification on the data collected by equipment with a confidence level lower than a preset value. Rainfall-related analysis includes: completing deformation prediction for whether a single landslide is a rainfall-type or non-rainfall-type landslide. In this operation, a clustering algorithm is used to determine the correlation between rainfall, adjacent water areas, groundwater, and landslide displacement, thereby determining whether a single landslide is a rainfall-type landslide. A classification model is used to rank the importance of water-related parameters affecting the displacement of a single landslide, and then a regression model is used to predict the change in the displacement of a single landslide as important water-related parameters change. Surface displacement-related analysis includes: completing landslide deformation prediction. In this operation, a classification model is used to rank the importance of displacement parameters affecting the displacement of a single landslide, including displacement acceleration, displacement increment tilt angle, and displacement rate. Then, a regression model is used to predict the change in the displacement of a single landslide as important displacement parameters change. Important displacement parameters include at least one, which is a combination of displacement parameters based on their importance.

[0021] As a further aspect of the present invention, in the data verification module, landslide type discrimination includes determining whether the landslide deformation is a false alarm, a rainfall-induced landslide, or a non-rainfall-induced landslide by calling the calculation results in the data analysis module; landslide stage discrimination includes determining the landslide type by using the time series model in the algorithm model, wherein the landslide type includes abrupt change, gradual change, and stable change; and data acquisition cycle discrimination includes judging the validity of the landslide data by the corresponding landslide data quality, thereby determining the time granularity of the acquired data for landslide calculation.

[0022] As a further aspect of the present invention, in the early warning module, the threshold interface includes: an externally exposed threshold data interface that can be called by other systems; and a threshold log including a scheduling log generated when an external system calls the threshold data of this system.

[0023] On the other hand, a method for early warning of individual landslides, the method comprising:

[0024] S01, through the data cleaning operation of the data acquisition and cleaning module and the data acquisition cycle discrimination operation of the data verification module, the effective range of the acquired landslide data is determined. Here, the effective range of data is when the effective data of the landslide accounts for more than 95% during the data acquisition period, and then the time period of the landslide is evaluated when calculating the landslide threshold.

[0025] S02, in the data analysis module, calculate the improved tangent angle of the landslide;

[0026] S03, In the data analysis module, by improving the correspondence between the tangent angle and the landslide displacement, the surface displacement value of the landslide displacement when the tangent angle threshold is improved at different alarm stages can be calculated, which is the static threshold of the landslide in the current collection period.

[0027] S04, In the data verification module, determine whether a single landslide is a rainfall-type single landslide;

[0028] S05. If it is a rainfall-induced single landslide, firstly, the correspondence between rainfall and landslide displacement is calculated by using a combination of Bayesian network, elastic network regression, and heuristic algorithm. Secondly, the future deformation value of the landslide is inferred from the rainfall value in the future weather forecast, and then the alarm forecast threshold for the current collection period is predicted.

[0029] S06. If it is a non-rainfall-induced single landslide, firstly, a random forest combined with a Bayesian network is used to select the importance of the influencing factors of landslide-induced deformation; secondly, a neural network algorithm is used to obtain the fusion parameters of the deformation influencing factors under multiple sampling periods by combining landslide displacement data, and the correspondence between the fusion parameters and landslide displacement can be obtained; thirdly, combined with the landslide deformation acceleration change rate, the influencing factor parameters with an importance of more than 20% are selected, and the Bayesian network is used to predict the value of the fusion parameter of the influencing factors in the next sampling period by using the influencing factor data of the previous two sampling periods, and then the fusion parameter value is used to set the alarm threshold for the next sampling period.

[0030] Furthermore, in the early warning module, regarding the dynamic changes in alarm thresholds, the threshold interval for rainfall-type single landslides is updated once per sampling cycle, while for non-rainfall-type single landslides, the threshold is updated based on whether the improved tangent angle reaches 45 degrees.

[0031] This invention provides a single-slope landslide early warning system and method, comprising a data acquisition and cleaning module, a data storage module, an algorithm model library, a data analysis module, an alarm module, and a data transmission module. The data acquisition and cleaning module processes monitoring data from multiple landslides; the algorithm model library contains algorithm models related to equipment, surface displacement, and rainfall; the data analysis module analyzes the correlation between equipment, surface displacement, and rainfall involved in the causes of a single landslide; the early warning module is used to publicly release early warning information and receive feedback on response; and the data verification module verifies the algorithm models and dynamically adjusts the algorithm verification cycle according to the classification of the algorithm models. Specific alarm thresholds can be defined for different single landslides, and landslide alarms can be predicted using classification and rainfall forecasts, thereby achieving the effect of early warning. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of the main structure of a single landslide early warning system.

[0033] Figure 2 This is a flowchart of the workflow for a single landslide early warning system. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0035] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0036] The present invention provides a single landslide early warning system and method, which can set specific alarm thresholds for different single landslides and predict landslide alarms by classifying and predicting values ​​such as rainfall, thereby achieving the effect of early warning and solving the technical problems in the background art.

[0037] like Figure 1 and Figure 2 The diagram shown is a schematic diagram of the main structure of a single landslide early warning system according to an embodiment of the present invention. The system includes a data acquisition and cleaning module, an algorithm model library, a data analysis module, a data verification module, and an early warning module.

[0038] The data acquisition and cleaning module is used to process monitoring data from multiple landslides;

[0039] The algorithm model library is used to store algorithm models related to equipment, surface displacement, and rainfall.

[0040] The data analysis module is used to perform landslide alarm analysis on equipment, surface displacement and rainfall involved in the causes of individual landslides based on monitoring data and algorithm model library, so as to generate early warning information;

[0041] The data verification module is used to verify the algorithm models in the algorithm model library and dynamically adjust the algorithm verification cycle according to categories;

[0042] The early warning module is used to release early warning information to the public and receive and process feedback information.

[0043] In a preferred embodiment of the present invention, the data acquisition and cleaning module is specifically used for: acquiring equipment logs, acquiring sensor data, and synchronizing meteorological data; and also for cleaning the acquired and synchronized data.

[0044] The equipment log collection specifically includes: acquiring the operation logs of environmental data acquisition equipment deployed around a single landslide, wherein the operation logs are used to determine the operational health status of the environmental data acquisition equipment;

[0045] Sensor data acquisition specifically includes: acquiring relevant environmental data collected by environmental data acquisition devices deployed around a single landslide, including the ambient temperature, vertical surface displacement, horizontal surface displacement, displacement acceleration, and groundwater level.

[0046] Cleaning the collected and synchronized data includes: performing usability processing on the collected and synchronized data, which includes: removing dirty data, filling in missing data, and correcting unreasonable data.

[0047] The algorithm model library includes: displacement-related algorithm models and rainfall-related algorithm models. The library supports continuous input of displacement-related and rainfall-related data for individual landslides to output alarm thresholds for different parameters of individual landslides.

[0048] Displacement correlation algorithm model: This model is designed for non-rainfall type single landslides. The input data is the historical monitoring displacement correlation data of the single landslide. The first correlation algorithm calculates the variation law between the two sets of data: horizontal displacement, vertical displacement, displacement acceleration, tilt angle and first corresponding displacement. The first correlation algorithm includes classification, clustering, regression and association rule correlation algorithms in data mining. The first corresponding displacement includes: horizontal, vertical and resultant displacement.

[0049] Rainfall-related algorithm model: This model is designed for rainfall-induced isolated landslides. The input data consists of historical monitoring data on water-related factors affecting the landslide. A second correlation algorithm is used to calculate the relationship between two sets of data: rainfall, water level change, soil moisture, water level change rate, and corresponding displacement. This second correlation algorithm includes classification, clustering, regression, and association rule algorithms from data mining. The corresponding displacement includes horizontal, vertical, and combined displacements. A time-series model may also be included.

[0050] In another preferred embodiment of the present invention, the data analysis module is used for: equipment reliability analysis, rainfall correlation analysis, surface displacement correlation analysis and hybrid analysis.

[0051] In another preferred embodiment of the present invention, the data verification module is used for: landslide type identification, landslide stage identification, landslide main influencing factor identification, data acquisition cycle identification, and landslide displacement verification.

[0052] When this invention is applied,

[0053] In another preferred embodiment of the present invention, the equipment reliability analysis in the data analysis module includes: performing confidence assessment analysis on the equipment failure rate and the collected failure data rate by monitoring equipment logs; and performing multiple data cleaning and verification on the data collected by equipment with a confidence level lower than a preset value; the rainfall-related analysis includes: completing the deformation prediction of whether a single landslide is a rainfall-type or non-rainfall-type landslide. In this operation, a clustering algorithm is used to determine the correlation between rainfall, adjacent water areas, groundwater, and landslide displacement, thereby determining whether a single landslide is a rainfall-type landslide; a classification model is used to rank the importance of water-related parameters affecting the displacement of a single landslide, and then a regression model is used to predict the change in the displacement of a single landslide as important water-related parameters change; the surface displacement-related analysis includes: completing landslide deformation prediction. In this operation, a classification model is used to rank the importance of displacement parameters affecting the displacement of a single landslide, including displacement acceleration, displacement increment tilt angle, and displacement rate; then a regression model is used to predict the change in the displacement of a single landslide as important displacement parameters change, wherein at least one important displacement parameter is included, which is determined by a combination of displacement parameters based on their importance.

[0054] In application of this invention, the landslide type determination in the data verification module includes determining whether the landslide deformation is a false alarm, a rainfall-induced landslide, or a non-rainfall-induced landslide by calling the calculation results in the data analysis module; the landslide stage determination includes determining the landslide type by using the time series model in the algorithm model, wherein the landslide type includes abrupt change, gradual change, and stable change; and the data acquisition cycle determination includes determining the validity of the landslide data by judging the data quality of the corresponding landslide data, thereby determining the time granularity of the acquired data for landslide calculation.

[0055] In another preferred embodiment of the present invention, the threshold interface in the early warning module includes: an externally exposed threshold data interface that can be called by other systems; and a threshold log that includes a scheduling log generated when an external system calls the threshold data of this system.

[0056] On the other hand, a method for early warning of individual landslides, the method comprising:

[0057] S01, through the data cleaning operation of the data acquisition and cleaning module and the data acquisition cycle discrimination operation of the data verification module, the effective range of the acquired landslide data is determined (the effective range of data here is that the effective data of the landslide accounts for more than 95% during the data acquisition period), and then the time period of the landslide in calculating the landslide threshold is evaluated (the time period here refers to: day, week, month, quarter; the displacement acquisition data for the landslide threshold determination here is the data of the last landslide stabilization period before deformation occurs).

[0058] S02, In the data analysis module, calculate the improved tangent angle of the landslide (the data selection principle is the same as in step 1; the improved tangent angle is calculated using an industry-standard method).

[0059] S03, In the data analysis module, by improving the correspondence between the tangent angle and the landslide displacement, the surface displacement value of the landslide displacement at different alarm stages (reminder level, warning level, alert level, alarm level) when the tangent angle threshold is improved can be calculated, which is the static threshold of the landslide in the current collection period;

[0060] S04, In the data verification module, determine whether a single landslide is a rainfall-type single landslide;

[0061] S05. If it is a rainfall-induced single landslide, firstly, the correspondence between rainfall and landslide displacement is calculated by using a combination of Bayesian network, elastic network regression, and heuristic algorithm. Secondly, the future deformation value of the landslide is inferred from the rainfall value in the future weather forecast, and then the alarm forecast threshold for the current collection period is predicted.

[0062] S06. If it is a non-rainfall-induced single landslide, firstly, a random forest combined with a Bayesian network is used to select the importance of the influencing factors of landslide-induced deformation; secondly, a neural network algorithm is used to obtain the fusion parameters of the deformation influencing factors under multiple sampling periods (the fusion parameters remain unchanged before the improved tangent angle changes by 45 degrees) by combining landslide displacement data, and the correspondence between the fusion parameters and landslide displacement can be obtained; thirdly, combined with the landslide deformation acceleration change rate, the influencing factor parameters with an importance of more than 20% are selected, and the Bayesian network is used to predict the value of the fusion parameter of the influencing factors in the next sampling period through the influencing factor data of the previous two sampling periods, and then the fusion parameter value is used to set the alarm threshold for the next sampling period.

[0063] In the early warning module, regarding the dynamic changes of alarm thresholds, the threshold interval for rainfall-type single landslides is updated once per sampling cycle, while for non-rainfall-type single landslides, the threshold is updated based on whether the improved tangent angle reaches 45 degrees.

[0064] The present invention provides a single landslide early warning system and method, comprising a data acquisition and cleaning module, a data storage module, an algorithm model library, a data analysis module, an alarm module, and a data transmission module. The data acquisition and cleaning module processes monitoring data from multiple landslides; the algorithm model library contains algorithm models related to equipment, surface displacement, and rainfall; the data analysis module analyzes the correlation between equipment, surface displacement, and rainfall involved in the causes of a single landslide; the early warning module is used to release early warning information and receive feedback information on handling; and the data verification module is used to verify the algorithm models and dynamically adjust the algorithm verification cycle according to the classification of the algorithm models. Specific alarm thresholds can be defined for different single landslides, and landslide alarms can be predicted through classification and rainfall prediction values, thereby achieving the effect of early warning.

[0065] In order for the above methods and systems to operate smoothly, the system may include more or fewer components than those described above, or combine certain components, or different components, in addition to the various modules mentioned above. For example, it may include input / output devices, network access devices, buses, processors, and memory.

[0066] The processor referred to can be a Central Processing Unit (CPU), or other general-purpose processors, 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, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. This processor is the control center of the system, connecting various parts through various interfaces and lines.

[0067] 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.

[0068] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

[0069] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A monolithic landslide warning system, characterized by, The system comprises a data collection and cleaning module, an algorithm model library, a data analysis module, a data verification module and a warning module; The data collection and cleaning module is used for processing monitoring data of multiple landslides; The algorithm model library is used for storing algorithm models related to equipment, ground displacement and rainfall; The data analysis module is used for performing relevant landslide warning analysis of equipment, ground displacement and rainfall involved in single landslide inducement based on monitoring data and the algorithm model library to generate early warning information; The data verification module is used for verifying the algorithm models in the algorithm model library and dynamically adjusting the algorithm verification period; The warning module is used for opening the early warning information to the outside and receiving treatment feedback information; In the data analysis module, the equipment reliability analysis comprises: performing confidence evaluation analysis on the failure rate of the equipment and the collected failure data rate through monitoring of the equipment log, and performing multiple data cleaning tests on the collected data of the equipment with a confidence lower than a preset value; the rainfall related analysis comprises: completing landslide deformation prediction of whether a single landslide is a rainfall type or a non-rainfall type, in which operation, a clustering algorithm is used to judge the correlation between rainfall, adjacent water area, underground water and landslide displacement, and then whether a single landslide is a rainfall type is judged; a classification model is used to sort the importance of water-related parameters affecting the displacement of a single landslide, and then a regression model is used to predict the change of the displacement of a single landslide with the change of important water-related parameters; the ground displacement related analysis comprises: completing landslide deformation prediction, in which operation, a classification model is used to sort the importance of displacement parameters affecting the displacement of a single landslide, and the sorting includes displacement acceleration, displacement increment inclination angle and displacement rate; then a regression model is used to predict the change of the displacement of a single landslide with the change of important displacement parameters, wherein the important displacement parameters at least include one, which is combined according to the importance of the displacement parameters.

2. The monolithic landslide warning system of claim 1, wherein, The data collection and cleaning module is specifically used for collecting equipment logs, collecting sensor data and synchronizing meteorological data, and is also used for cleaning the collected and synchronized data, wherein, The equipment log collection specifically comprises: obtaining the running log of the environmental data collection equipment deployed around the single landslide, and the running log is used to judge the running health status of the environmental data collection equipment; The sensor data collection specifically comprises: obtaining the relevant environmental data collected by the environmental data collection equipment deployed around the single landslide, and the relevant environmental data includes the temperature of the environment, vertical ground displacement, horizontal ground displacement, displacement acceleration and underground water level; The cleaning of the collected and synchronized data comprises: performing availability processing on the collected and synchronized data, and the availability processing comprises: removing dirty data, filling in missing data and changing unreasonable data.

3. The monolithic landslide warning system of claim 1, wherein, The algorithm model library comprises: displacement related algorithm models and rainfall related algorithm models, and the algorithm model library supports continuous data input of single landslide displacement and rainfall to output single landslide different parameter warning thresholds, wherein, The displacement correlation algorithm model is used for a non-rainfall type single landslide, and input data is historical monitoring displacement correlation data of the single landslide; a first correlation algorithm is used to calculate the change law between two groups of data, i.e., a horizontal displacement, a vertical displacement, a displacement acceleration, an inclination angle and a first corresponding displacement; the first correlation algorithm includes classification, clustering, regression and association rule correlation algorithms in data mining, and the first corresponding displacement includes a horizontal displacement, a vertical displacement and a combined displacement. The rainfall correlation algorithm model is used for a rainfall type single landslide, and input data is historical monitoring water-related data of the single landslide; a second correlation algorithm is used to calculate the change law between two groups of data, i.e., a rainfall, a water level change, a soil humidity, a water level change rate and a second corresponding displacement; the second correlation algorithm includes classification, clustering, regression, association rule correlation algorithms in data mining, and the second corresponding displacement includes a horizontal displacement, a vertical displacement and a combined displacement.

4. The monolithic landslide warning system of claim 1, wherein, The data analysis module is used for device reliability analysis, rainfall correlation analysis, surface displacement correlation analysis and mixed analysis.

5. The monolithic landslide warning system of claim 1, wherein, The data verification module is used for landslide type discrimination, landslide stage discrimination, landslide main influence factor discrimination, data acquisition cycle discrimination and landslide displacement verification.

6. The monitory landslides warning system according to any one of claims 1-5, characterized in that, In the data verification module, the landslide type discrimination includes determining, by calling the calculation result in the data analysis module, that the landslide deformation is a device false alarm, a rainfall type or a non-rainfall type; the landslide stage discrimination includes determining, by a time sequence model in the algorithm model, the landslide type, wherein the landslide type includes a sudden change, a gradual change and a stability; and the data acquisition cycle discrimination includes determining, by corresponding landslide data quality, the data effectiveness of the landslide, and then determining the time granularity of the collected data for landslide calculation.

7. The monolithic landslide warning system of claim 1, wherein, In the early warning module, the threshold interface includes a threshold data interface exposed to the outside, which can be called by other systems; and the threshold log includes a scheduling log generated when the external system calls the threshold data of the system.

8. A monomer landslide early warning method applied to the monomer landslide early warning system of any one of claims 1-7, characterized in that, The method comprises: S01, determining the effective interval of the collected landslide data by the data cleaning operation of the data acquisition and cleaning module and the data acquisition cycle discrimination operation of the data verification module, and then evaluating the time period of the landslide in calculating the landslide threshold; S02, calculating the improved tangent angle of the landslide in the data analysis module; S03, calculating the ground surface displacement value of the landslide displacement at the improved tangent angle threshold in different alarm stages, i.e., the static threshold of the landslide in the current acquisition cycle, by the corresponding relationship between the improved tangent angle and the landslide displacement in the data analysis module; S04, determining whether the single landslide is a rainfall type single landslide in the data verification module; S05, if it is a rainfall type single landslide, firstly, calculating the mapping relationship between rainfall and landslide displacement by using a Bayesian network combined with an elastic network regression combined with a heuristic algorithm, and secondly, inferring the future deformation value of the landslide by the future weather forecast rainfall value, and then predicting the alarm prediction threshold of the current acquisition cycle. S06, if it is a non-rainfall type monomer landslide, firstly, the importance of the influence factors of landslide-induced deformation is selected by combining random forest with Bayesian network; secondly, the fusion parameters of the influence factors of deformation in multiple sampling periods are obtained by combining landslide displacement data with a neural network algorithm, wherein the fusion parameters remain unchanged before the improved tangent angle changes by 45 degrees, and the corresponding relationship between the fusion parameters and the landslide displacement can be obtained; thirdly, combined with the acceleration change rate of landslide deformation, the influence factor parameters with an importance of more than 20% are selected, the fusion parameter value of the influence factors in the next collection period is predicted by Bayesian network through the influence factor data in the previous two collection periods, and then the fusion parameter value is used to set the alarm threshold in the next collection period.

9. The monolithic landslide early warning method of claim 8, wherein, In the early warning module, regarding the dynamic change of the alarm threshold, the threshold interval of the rainfall type monomer landslide is updated every sampling period, and the non-rainfall type monomer landslide is updated according to whether the improved tangent angle reaches 45 degrees.

Images