A method and system for simulating and analyzing a rainfall-induced landslide geological disaster scene

Abstract

This invention relates to a method for simulating and analyzing rainfall-induced landslide geological disaster scenarios, belonging to the field of image synthesis, and more specifically to the field of 3D geographic modeling for computer mapping. The method includes: collecting customized model data of a given mountain at the current moment based on a 3D graphical model of the mountain; and using a landslide intelligent prediction model to intelligently predict whether a landslide geological disaster will occur on each side of the given mountain within the next time interval, starting from the current moment, based on the customized model data. This invention also relates to a rainfall-induced landslide geological disaster scenario simulation and analysis system. Through this invention, the technical problem of difficulty in predicting future landslide data for landslide-prone physical entities such as mountains is addressed by intelligently predicting whether a landslide will occur on each side of the given mountain within a future time interval based on a 3D graphical model of the mountain at the current moment, thereby solving the aforementioned technical problem.

Description

Technical Field

[0001] The image synthesis of the present invention relates more specifically to the field of 3D geographic modeling for computer mapping, and particularly to a method and system for simulating and analyzing rainfall-induced landslide geological disaster scenarios. Background Technology

[0002] A 3D model is a polygonal representation of an object, typically displayed using a computer or other video equipment. The displayed object can be a real-world entity or a fictional one. Anything existing in the physical world can be represented by a 3D model. 3D models are often generated using specialized software such as 3D modeling tools, but they can also be generated using other methods. As a set of points and other information, a 3D model can be generated manually or according to a specific algorithm. Although it usually exists virtually in a computer or computer file, a similar model described on paper can also be considered a 3D model. A popular method for 3D modeling involves taking multiple images of different sides of an object, synthesizing and fusing these images to obtain a 3D model of the object. If geographic data of the object is then incorporated, a 3D geographic model of the object can be constructed. 3D geographic models have a wide range of applications.

[0003] For example, Chinese invention patent publication CN106018038A discloses a method and apparatus for manufacturing a landslide physical model. The method includes: configuring a first printing slurry, a second printing slurry, and a third printing slurry; establishing a virtual landslide physical model; injecting the first printing slurry into a 3D printer, and controlling the 3D printer to print the slide of the landslide physical model according to the shape and size information of the slide; injecting the second printing slurry into the 3D printer, and controlling the 3D printer to print the slide of the landslide physical model according to the shape and size information of the slide belt; injecting the third printing slurry into the 3D printer, and controlling the 3D printer to print the sliding body of the landslide physical model according to the shape and size information of the slide. The apparatus consists of a modeling unit, a 3D printer, and a control unit. The landslide physical model manufacturing method and apparatus provided by this invention can realistically simulate the main physical and mechanical parameters of an actual landslide zone, facilitating scientific research on landslides.

[0004] For example, Chinese invention patent publication CN114674630A proposes a method for preparing a geomechanical model of a water-related landslide with controllable stress reduction simulation. The method includes the following preparation steps: S1, material preparation: preparing a test chamber, test water, model soil, PVA material, and a bedrock layer; S2, reinforcement layer fabrication: using 3D-printed PVA material to create a reinforcement layer; S3, bedrock layer simulation: arranging the bedrock layer in the test chamber according to the topographic data of the landslide area; S4, stratum shape simulation: arranging the model soil and reinforcement layer on the bedrock layer according to the topographic data of the landslide area, and compacting the model soil; S5, seepage simulation: slowly adding test water into the test chamber through the inner wall of the test chamber. This provides an artificial preparation scheme for a geomechanical model of a water-related landslide suitable for simulating different failure modes, and offers a controllable and easily similar design-compliant experimental method for simulating landslide instability induced by heavy rainfall and slope instability caused by reservoir water level changes.

[0005] However, the aforementioned existing technologies only involve the design and research of 3D simulated physical models, lacking a modeling mechanism for constructing corresponding 3D geographic models that can be displayed in a computer for each mountain. Furthermore, the aforementioned existing technologies only involve real-time landslide detection and analysis of simulated physical entities, and cannot reliably predict the occurrence of landslide geological disasters for real physical entities, such as real mountains, in future time intervals. They also cannot effectively predict the side of the mountain where a landslide geological disaster will occur. As a result, the analysis technology of 3D physical entities in the existing technologies is limited, and it is difficult to provide predictive reference information for disaster emergency response departments based on landslide geological disaster simulation analysis. Summary of the Invention

[0006] To address the technical problems in existing technologies, this invention provides a method and system for simulating and analyzing rainfall-induced landslide geological disaster scenarios. For a given mountain that may experience landslides under rainfall-induced geological disasters, a 3D geographic model of the given mountain is constructed through image synthesis and fusion of various sides of the mountain and the fusion of geographic data. Based on this, intelligent prediction of whether landslides will occur on various sides of the given mountain in the next time interval (starting from the current time) is achieved using the 3D graphic model of the given mountain at the current moment, the total rainfall in the most recent time interval (ending from the current time), and multiple related information of the given mountain. This provides valuable reference information for early warning and response to landslide geological disasters.

[0007] According to a first aspect of the present invention, a method for simulating and analyzing rainfall-induced landslide geological disaster scenarios is provided, the method comprising:

[0008] Receive visual images corresponding to each side of the set mountain at the current moment, and construct a 3D graphic model of the set mountain at the current moment based on each visual image and the closest distance of each camera mechanism that captured each visual image relative to the set mountain.

[0009] Based on the current moment of the 3D graphic model of the designated mountain, collect various customized model data of the designated mountain at the current moment;

[0010] The system outputs multiple related information about the mountain, including its area, peak height, width, length, soil type number, and main crop type number.

[0011] The landslide intelligent prediction model uses the total rainfall, duration of the time interval, customized model data of the set mountain at the current moment, and multiple related information of the set mountain to intelligently predict whether landslide geological disasters will occur on each side of the set mountain in the next time interval with the current moment as the starting moment.

[0012] The decision to trigger a landslide warning for the next time interval of the mountain is based on the results of intelligent prediction.

[0013] The landslide intelligent prediction model is a Hofit neural network after completing each learning action.

[0014] According to a second aspect of the present invention, a simulation and analysis system for rainfall-induced landslide geological disaster scenarios is provided, the system comprising:

[0015] A mountain modeling device is used to receive visual images corresponding to each side of a set mountain at the current moment, and to construct a 3D graphic model of the set mountain at the current moment based on each visual image and the closest distance of each camera mechanism that captured each visual image relative to the set mountain.

[0016] The data acquisition device is connected to the mountain modeling device and is used to acquire various customized model data of the mountain at the current moment based on the 3D graphic model of the mountain at the current moment.

[0017] The target analysis device is used to output multiple related information of the set mountain, including its area, peak height, width, length, soil type number, and main crop type number;

[0018] The intelligent prediction device is connected to the data acquisition device and the target analysis device respectively. It is used to use the landslide intelligent prediction model to intelligently predict whether landslide geological disasters will occur on each side of the set mountain in the next time interval with the current time as the start time. The prediction data is based on the total rainfall in the most recent time interval with the current time as the end time, the duration of the time interval, the customized model data of the set mountain at the current time, and multiple related information of the set mountain.

[0019] The early warning triggering device is connected to the intelligent prediction device and is used to determine whether to trigger an early warning of landslide occurrence in the next time interval of the set mountain based on the intelligent prediction results;

[0020] The landslide intelligent prediction model is a Hofit neural network after completing each learning action.

[0021] Compared with the prior art, the present invention has at least the following four key inventive points:

[0022] Invention Point A: Based on the 3D graphic model of the mountain at the current moment, the total rainfall in the most recent time interval with the current moment as the end moment, and multiple related information of the mountain, intelligent prediction of whether landslide geological disasters will occur on each side of the mountain in the next time interval with the current moment as the start moment is completed. This provides valuable reference information for early warning and response to landslide geological disasters. The 3D graphic model of the mountain at the current moment is completed by image synthesis of each visual image corresponding to each side of the mountain at the current moment and the integration of other auxiliary information.

[0023] Invention Point B: To perform intelligent prediction of the mountain side where a landslide occurs in the next time interval, a landslide intelligent prediction model with a customized structure is adopted. The landslide intelligent prediction model is a Hoffert neural network after completing each learning action, and the number of learning actions of the Hoffert neural network is monotonically positively correlated with the number of each side. Thus, landslide intelligent prediction models with different structures are customized for 3D graphic models of mountains with different accuracies, thereby ensuring the effectiveness and stability of the intelligent prediction results of the mountain side where a landslide occurs.

[0024] Invention Point C: To perform intelligent prediction of the side of a mountain where a landslide is likely to occur in the next time interval, several basic information items are specifically selected. These basic information items include the total rainfall in the most recent time interval ending at the current time, the duration of the time interval, various customized model data of the mountain at the current time, and various related information of the mountain. The various customized model data of the mountain at the current time are the 3D coordinate values, curvature values, gray values, and gray gradient values ​​of each surface pixel of the 3D geographic model of the mountain at the current time. The various related information of the mountain includes the area, peak height, width, length, soil type number, and main crop type number of the mountain. The targeted selection of the above-mentioned basic information items further ensures the effectiveness and stability of the intelligent prediction results of the side of a mountain where a landslide is likely to occur.

[0025] Invention Point D: In each learning action performed on the Hofit Neural Network, the landslide prediction data of whether landslides have occurred on each side of a mountain within a certain historical time interval are used as the output of the intelligent landslide prediction model. The total rainfall, duration of the time interval, customized model data of the mountain at the start time of the certain historical time interval, and multiple related information of the mountain are used as the input of the intelligent landslide prediction model to complete the learning action, thereby ensuring the learning effect of each learning action of the Hofit Neural Network. Attached Figure Description

[0026] The embodiments of the present invention will now be described with reference to the accompanying drawings, wherein:

[0027] Figure 1 This is a schematic diagram of the working scenario of the simulation analysis method and system for rainfall-induced landslide geological disasters according to the present invention.

[0028] Figure 2 The present invention is illustrated in Embodiment 1 of the present invention as a flowchart of a method for simulating and analyzing a landslide geological disaster scenario induced by rainfall.

[0029] Figure 3 The present invention is illustrated in Embodiment 2 of the present invention as a flowchart of a method for simulating and analyzing a landslide geological disaster scenario induced by rainfall.

[0030] Figure 4 The present invention is illustrated in Embodiment 3 of the present invention as a flowchart of a method for simulating and analyzing a landslide geological disaster scenario induced by rainfall.

[0031] Figure 5 The present invention is illustrated in Embodiment 4 of the present invention as a flowchart of a method for simulating and analyzing a landslide geological disaster scenario induced by rainfall.

[0032] Figure 6 The present invention is illustrated in Embodiment 5 of the present invention as a flowchart of a method for simulating and analyzing a landslide geological disaster scenario induced by rainfall.

[0033] Figure 7 This is a schematic diagram of a simulation and analysis system for a rainfall-induced landslide geological disaster scenario, as shown in Embodiment 6 of the present invention. Detailed Implementation

[0034] like Figure 1 The diagram illustrates the working scenario of a simulation analysis method and system for rainfall-induced landslide geological disasters according to the present invention. The image synthesis of the present invention more specifically relates to the field of 3D geographic modeling for computer mapping.

[0035] The specific technical process of this invention is as follows:

[0036] Technical Process 1: A 3D graphic model of the mountain at the current moment is created by combining image composites of various visual images corresponding to different sides of the mountain at the current moment, along with other auxiliary information. Figure 1 As shown;

[0037] In this way, as a physical entity prone to landslide geological disasters, the modeling and processing of the 3D graphic model of the mountain at the current moment provides a model basis for the simulation analysis of landslide geological disasters in the future time interval of the mountain.

[0038] Technical Process 2: To perform intelligent prediction of the side of the mountain where a landslide is expected to occur in the next time interval, a landslide intelligent prediction model with a customized structural design was introduced.

[0039] Specifically, the customized structural design of the landslide intelligent prediction model is mainly reflected in the following aspects:

[0040] First: The landslide intelligent prediction model is a Hoffert neural network after completing each learning action, that is, the model architecture of the landslide intelligent prediction model is based on the network structure of the Hoffert neural network.

[0041] Second: The number of learning actions of the Hofit neural network used is monotonically positively correlated with the number of each side, thereby customizing intelligent landslide prediction models with different structures for 3D graphic models of mountains with different precision.

[0042] For example, if 6 sides are selected, the number of actions learned by the Hofit neural network is 600; if 8 sides are selected, the number of actions learned by the Hofit neural network is 700; if 10 sides are selected, the number of actions learned by the Hofit neural network is 800; if 12 sides are selected, the number of actions learned by the Hofit neural network is 900, and so on.

[0043] Third: In each learning action performed on the Hofit Neural Network, the landslide prediction data of whether landslides have occurred on the sides of a certain mountain within a certain historical time interval are used as the output of the landslide intelligent prediction model. The total rainfall, duration of the time interval, customized model data of the mountain at the start time of the certain historical time interval, and multiple related information of the mountain are used as the input of the landslide intelligent prediction model to complete the learning action, thereby ensuring the learning effect of each learning action of the Hofit Neural Network.

[0044] In this way, through the above-mentioned customized structural designs, the effectiveness and stability of the intelligent prediction results for the side of the mountain where a landslide occurs are ensured.

[0045] Technical Process 3: To perform intelligent prediction of the mountainside where landslides are expected in the next time interval, several basic information items were specifically selected;

[0046] Specifically, such as Figure 1 As shown, the various basic information includes the total rainfall in the most recent time interval ending at the current time, the duration of the time interval, the various customized model data of the set mountain at the current time, and various related information of the set mountain;

[0047] Therefore, the introduction of the total rainfall within the most recent time interval, with the current time as the end time, provides some basic data for the simulation analysis of rainfall-induced landslide geological disaster scenarios;

[0048] More specifically, the customized model data of the mountain at the current moment is set as the 3D coordinate value, curvature value, gray value and gray gradient value of each surface pixel of the 3D geographic model of the mountain at the current moment. The multiple related information of the mountain is set as the area, peak height, width, length, soil type number and main crop type number of the mountain.

[0049] In this way, the targeted selection of the above-mentioned basic information further ensures the effectiveness and stability of the intelligent prediction results for the side of the mountain where a landslide occurs.

[0050] Technical Process 4: The landslide intelligent prediction model, which adopts the customized structural design of Technical Process 2, intelligently predicts the side of the mountain where a landslide will occur in the next time interval based on multiple basic information selected in Technical Process 3.

[0051] Specifically, such as Figure 1 As shown, the intelligent prediction result is a set of landslide prediction data for each side of a set mountain, within the next time interval starting from the current time, indicating whether a landslide geological disaster will occur. The landslide prediction data for each side of a set mountain consists of the side code of the mountain side and an occurrence identifier indicating whether a landslide geological disaster will occur on the mountain side.

[0052] Technical Process 5: Based on the intelligent prediction results of Technical Process 4, determine whether to trigger a landslide warning for the next time interval of the set mountain.

[0053] For example, when the occurrence marker in the landslide prediction data for setting whether a landslide geological disaster will occur on a certain side of a mountain indicates that a landslide geological disaster has occurred on that certain side of the mountain, a landslide occurrence warning is executed for the next time interval of the set mountain side.

[0054] Therefore, through the collaboration of the above-mentioned multiple technical processes, it is possible to intelligently predict whether landslide geological hazards will occur on each side of the mountain within the next time interval starting from the current time, based on the 3D graphic model of the mountain at the current moment, the total rainfall in the most recent time interval with the current moment as the end time, and multiple related information of the mountain. This provides valuable reference information for early warning and response to landslide geological hazards.

[0055] The key points of this invention are: a targeted modeling mechanism for a 3D graphic model of a mountain at the current moment, which is based on the image synthesis of each visual image corresponding to each side of the mountain at the current moment and the integration of other auxiliary information; customization of intelligent landslide prediction models with different structures of 3D graphic models of mountains with different precision; targeted selection of multiple basic information for intelligent prediction; and synchronous simulation analysis of whether landslide geological disasters will occur on each side of the mountain in future time intervals.

[0056] The following will provide a detailed description of the simulation analysis method and system for rainfall-induced landslide geological disaster scenarios of the present invention through examples.

[0057] Example 1

[0058] Figure 2 The present invention is illustrated in Embodiment 1 of the present invention as a flowchart of a method for simulating and analyzing a landslide geological disaster scenario induced by rainfall.

[0059] like Figure 2 As shown, the simulation and analysis method for rainfall-induced landslide geological disaster scenarios includes the following specific steps:

[0060] Step S201: Receive each visual image corresponding to each side of the set mountain at the current moment, and construct a 3D graphic model of the set mountain at the current moment based on each visual image and the closest distance of each camera mechanism that captured each visual image relative to the set mountain.

[0061] For example, receiving visual images corresponding to each side of a set mountain at the current moment, and constructing a 3D graphic model of the set mountain at the current moment based on each visual image and the closest distance of each camera mechanism that captured each visual image relative to the set mountain, includes: using the combination and fusion of each image of the set object from each side to obtain a 3D graphic model of the set object that can be displayed on a computer; if the fusion of some geographic data, such as the fusion of location information, is added, a 3D graphic model of the set object with geographic data that can be displayed on a computer can be further obtained.

[0062] In this way, the 3D graphic model of the object not only has a three-dimensional shape that is a combination and fusion of various sides, but each side of this three-dimensional shape is composed of pixels of the image of each side. These pixels of the side have three-dimensional coordinates, that is, coordinate values ​​in three directions of the three-dimensional coordinate system. Moreover, these side pixels of the 3D graphic model of the object also have geographic data, such as location information.

[0063] Step S202: Collect various customized model data of the set mountain at the current moment based on the 3D graphic model of the set mountain at the current moment;

[0064] Specifically, since the 3D graphic model of the mountain at the current moment can be displayed on the computer, various customized model data of the mountain at the current moment can be directly called and obtained from the computer, such as the coordinate values ​​of the three directions of these side pixels of the 3D graphic model of the object and its geographical data.

[0065] Step S203: Output the area, peak height, width, length, soil type number, and main crop type number of the mountain as multiple related information of the mountain.

[0066] Specifically, the output of multiple related information of the mountain is set, including the area, peak height, width, length, soil type number, and main crop type number. Different soil types have different soil type numbers. For example, red soil and black soil have different soil type numbers, and different main crop types have different main crop type numbers.

[0067] Step S204: Using the landslide intelligent prediction model, based on the total rainfall, duration of the time interval, customized model data of the set mountain at the current moment, and multiple related information of the set mountain, the landslide intelligent prediction model is used to predict whether landslide geological disasters will occur on each side of the set mountain in the next time interval starting from the current moment.

[0068] For example, if the current time is 5:30 PM and the duration of the time interval is 15 minutes, then the nearest time interval ending at the current time is the time interval from 5:15 PM to 5:30 PM, and the next time interval starting at the current time is the time interval from 5:30 PM to 5:45 PM.

[0069] Step S205: Based on the intelligent prediction results, decide whether to trigger a landslide warning for the next time interval of the set mountain.

[0070] In this way, by predicting in advance whether a landslide will occur in a designated mountain in the future, intelligent disaster management of the designated mountain can be achieved, providing valuable reference information for the evacuation of nearby residents or the closure of highways;

[0071] The landslide intelligent prediction model is a Hofit neural network after completing each learning action, that is, the architecture of the landslide intelligent prediction model is the network architecture of the Hofit neural network.

[0072] Among them, the collection of various customized model data of the current moment of the 3D graphic model of the set mountain includes: collecting the 3D coordinate values, curvature values, gray values ​​and gray gradient values ​​of each surface pixel of the 3D geographic model of the set mountain at the current moment as various customized model data of the set mountain at the current moment.

[0073] Specifically, the 3D coordinates, curvature values, gray values, and gray gradient values ​​of each surface pixel of the 3D geographic model of the mountain at the current moment are collected as customized model data of the mountain at the current moment. The gray value of each surface pixel is between 0 and 255.

[0074] The landslide prediction data for determining whether a landslide geological disaster will occur on each side of the mountain consists of a side code of the mountain side and an occurrence identifier indicating whether a landslide geological disaster will occur on the mountain side.

[0075] Specifically, the landslide prediction data for each side of the mountain, which sets the conditions for whether landslides will occur, provides the specific sides of the mountain where landslides are likely to occur.

[0076] Among them, determining whether to trigger a landslide occurrence warning for the next time interval of a set mountain based on intelligent prediction results includes: when the occurrence identifier in the landslide prediction data of whether a landslide geological disaster has occurred on a certain side of a set mountain indicates that a landslide geological disaster has occurred on that certain side of the mountain, executing a landslide occurrence warning for the next time interval of that certain side of the set mountain.

[0077] The process of receiving visual images corresponding to each side of the set mountain at the current moment, and constructing a 3D graphic model of the set mountain at the current moment based on each visual image and the closest distance of each camera that captured each visual image relative to the set mountain, includes: receiving visual images corresponding to each side of the set mountain at the current moment, and constructing a 3D graphic model of the set mountain at the current moment based on each visual image, the closest distance of each camera that captured each visual image relative to the set mountain, and the positioning data corresponding to each camera.

[0078] The landslide intelligent prediction model is a Hoffert neural network after completing each learning action, which includes: the number of learning actions of the Hoffert neural network is monotonically positively correlated with the number of each side.

[0079] For example, the number of learning actions of the Hofit neural network is monotonically positively correlated with the number of sides selected, including: when 6 sides are selected, the number of learning actions of the Hofit neural network is 600; when 8 sides are selected, the number of learning actions of the Hofit neural network is 700; when 10 sides are selected, the number of learning actions of the Hofit neural network is 800; when 12 sides are selected, the number of learning actions of the Hofit neural network is 900, and so on.

[0080] It is evident that the more aspects are selected, the more times the Hofit neural network learns actions, and the more refined and reliable the landslide intelligent prediction model becomes.

[0081] In each learning action performed on the Hofit neural network, the landslide prediction data, which indicates whether landslides have occurred on the various sides of a mountain within a certain historical time interval, are used as the output of the landslide intelligent prediction model. The total rainfall, duration of the time interval, customized model data of the mountain at the start time of the historical time interval, and various related information of the mountain are used as the input of the landslide intelligent prediction model to complete the learning action.

[0082] Example 2

[0083] Figure 3 The present invention is illustrated in Embodiment 2 of the present invention as a flowchart of a method for simulating and analyzing a landslide geological disaster scenario induced by rainfall.

[0084] like Figure 3 As shown, with Figure 2 Unlike the previous embodiment, in the rainfall-induced landslide geological disaster scenario simulation and analysis method, after using a landslide intelligent prediction model to intelligently predict whether a landslide geological disaster will occur on each side of the set mountain in the next time interval starting from the current time, based on the total rainfall, duration of the time interval, customized model data of the set mountain at the current time, and multiple related information of the set mountain, the method further includes, after step S204:

[0085] Step S206: Receive landslide prediction data for each side of the mountain set within the next time interval starting from the current time, indicating whether a landslide geological disaster has occurred, and display in real time the landslide prediction data for each side of the mountain set within the next time interval starting from the current time.

[0086] Specifically, receiving landslide prediction data on whether landslides have occurred on each side of a designated mountain within the next time interval starting from the current time, and displaying such data in real time includes: selecting an LCD screen or a large-screen display device to display the landslide prediction data in real time on each side of a designated mountain within the next time interval starting from the current time. Example

[0087] Figure 4 The present invention is illustrated in Embodiment 3 of the present invention as a flowchart of a method for simulating and analyzing a landslide geological disaster scenario induced by rainfall.

[0088] like Figure 4 As shown, with Figure 2 Unlike the previous embodiment, in the rainfall-induced landslide geological disaster scenario simulation and analysis method, after using a landslide intelligent prediction model to intelligently predict whether a landslide geological disaster will occur on each side of the set mountain in the next time interval starting from the current time, based on the total rainfall, duration of the time interval, customized model data of the set mountain at the current time, and multiple related information of the set mountain, the method further includes, after step S204:

[0089] Step S207: Receive landslide prediction data for each side of the mountain within the next time interval starting from the current time, indicating whether a landslide geological disaster has occurred, and wirelessly transmit the landslide prediction data for each side of the mountain within the next time interval starting from the current time to the remote geological disaster management server via a network communication link.

[0090] For example, receiving landslide prediction data on whether landslides have occurred on each side of a designated mountain within the next time interval starting from the current time, and wirelessly transmitting this landslide prediction data to a remote geological disaster management server via a network communication link includes: the remote geological disaster management server being a big data service node, a cloud computing service node, or a blockchain service node.

[0091] Example 4

[0092] Figure 5 The present invention is illustrated in Embodiment 4 of the present invention as a flowchart of a method for simulating and analyzing a landslide geological disaster scenario induced by rainfall.

[0093] like Figure 5 As shown, with Figure 2 Unlike the previous embodiment, in the simulation and analysis method for rainfall-induced landslide geological disaster scenarios, after outputting the area, peak height, width, length, soil type number, and main crop type number of the set mountain as multiple related information of the set mountain, that is, after step S203, the method further includes:

[0094] Step S208: Perform each learning action on the Hofit neural network to obtain the Hofit neural network after completing each learning action and output it as the landslide intelligent prediction model.

[0095] Specifically, performing each learning action on the Hofit neural network to obtain the Hofit neural network after completing each learning action and outputting it as the landslide intelligent prediction model includes: simulation and testing of the modeling process of performing each learning action on the Hofit neural network to obtain the Hofit neural network after completing each learning action and outputting it as the landslide intelligent prediction model, which can be completed using the MATLAB toolbox.

[0096] Example 5

[0097] Figure 6 The present invention is illustrated in Embodiment 5 of the present invention as a flowchart of a method for simulating and analyzing a landslide geological disaster scenario induced by rainfall.

[0098] like Figure 6 As shown, with Figure 2 Unlike the previous embodiment, in the simulation and analysis method for rainfall-induced landslide geological disaster scenarios, before outputting the area, peak height, width, length, soil type number, and main crop type number of the set mountain as multiple related information of the set mountain, that is, before step S203, the method further includes:

[0099] Step S209: Use multiple different physical measurement components to measure the area, peak height, width, length, soil type, and main crop type of the designated mountain. Based on the measured soil type and main crop type of the designated mountain, determine the soil type number and main crop type number of the designated mountain.

[0100] For example, multiple different physical measurement components are used to measure the area, peak height, width, length, soil type, and main crop type of the designated mountain. Based on the measured soil type and main crop type of the designated mountain, the soil type number and main crop type number of the designated mountain are determined. This includes: the area of ​​the designated mountain can be measured using a satellite image visual analysis device, and the peak height of the designated mountain can be measured using a height measurement device.

[0101] In addition, since setting the area, peak height, width, length, soil type, and main crop type of the mountain is not time-sensitive, these data can be measured in advance and are applicable to multiple times when landslide geological disaster scenario simulation analysis is initiated.

[0102] Next, the various method embodiments of the present invention will be described in detail.

[0103] In the simulation and analysis method for rainfall-induced landslide geological disaster scenarios according to various method embodiments of the present invention:

[0104] The monotonically positive correlation between the number of learning actions and the number of each side in the Hofit neural network includes: using a number mapping function to represent the monotonically positive correlation between the number of learning actions and the number of each side in the Hofit neural network.

[0105] For example, an FPGA chip can be used to represent the monotonically positive correlation between the number of learning actions of the Hoffett neural network and the number of each side using a number mapping function.

[0106] The method of using a frequency mapping function to represent the monotonically positive correlation between the number of learning actions of the Hoffett neural network and the number of each side includes: in the frequency mapping function, the number of each side is the input data of the frequency mapping function, and the number of learning actions of the Hoffett neural network corresponding to the number of each side is the output data of the frequency mapping function.

[0107] And in the simulation and analysis method for rainfall-induced landslide geological disaster scenarios according to various method embodiments of the present invention:

[0108] The landslide intelligent prediction model uses the total rainfall, duration of the time interval, customized model data of the set mountain at the current time, and multiple related information of the set mountain to intelligently predict whether landslide geological disasters will occur on each side of the set mountain in the next time interval starting from the current time. The prediction data includes: the total rainfall, duration of the time interval, customized model data of the set mountain at the current time, and multiple related information of the set mountain are synchronously input into the landslide intelligent prediction model.

[0109] The intelligent landslide prediction model uses the total rainfall, duration of the time interval, customized model data of the set mountain at the current moment, and multiple related information of the set mountain to intelligently predict whether landslide geological disasters will occur on each side of the set mountain in the next time interval starting from the current moment. This prediction data also includes: executing the intelligent landslide prediction model to obtain the output of the intelligent landslide prediction model, which shows whether landslide geological disasters will occur on each side of the set mountain in the next time interval starting from the current moment.

[0110] For example, if the current time is 5:30 PM and the duration of the time interval is 15 minutes, then the nearest time interval ending at the current time is the time interval from 5:15 PM to 5:30 PM, and the next time interval starting at the current time is the time interval from 5:30 PM to 5:45 PM.

[0111] The process of synchronously inputting the total rainfall, duration of the time interval, customized model data of the set mountain at the current moment, and multiple related information of the set mountain into the landslide intelligent prediction model includes: performing octal value conversion processing on the total rainfall, duration of the time interval, customized model data of the set mountain at the current moment, and multiple related information of the set mountain before synchronously inputting them into the landslide intelligent prediction model.

[0112] Among them, the execution of the landslide intelligent prediction model to obtain the landslide prediction data of each side of the mountain set in the next time interval with the current time as the start time, which is a representation of whether a landslide geological disaster will occur on each side of the mountain set in the next time interval with the current time as the start time, includes: the landslide prediction data of each side of the mountain set in the next time interval with the current time as the start time, which is a representation of octal values;

[0113] Specifically, before synchronously inputting the total rainfall, duration of the time interval, various customized model data of the set mountain at the current moment, and various related information of the set mountain into the landslide intelligent prediction model, octal value conversion processing is performed on the total rainfall, duration of the time interval, various customized model data of the set mountain at the current moment, and various related information of the set mountain, respectively. This includes: using a first programmable logic chip to implement octal value conversion processing, and using a second programmable logic chip to implement synchronous input.

[0114] Specifically, the use of a first programmable logic chip to implement octal value conversion processing and the use of a second programmable logic chip to implement synchronous input include: both the first and second programmable logic chips are CPLD chips but of different models, and both the programming design of the first and second programmable logic chips uses VHDL language;

[0115] Furthermore, before synchronously inputting the total rainfall, duration of the time interval, various customized model data of the set mountain at the current moment, and various related information of the set mountain into the landslide intelligent prediction model, the octal value conversion processing of the total rainfall, duration of the time interval, various customized model data of the set mountain at the current moment, and various related information of the set mountain at the current moment also includes: the first programmable logic chip and the second programmable logic chip are connected and share the same clock generator and the same parameter configuration interface.

[0116] Example 6

[0117] Figure 7 This is a schematic diagram of a simulation and analysis system for a rainfall-induced landslide geological disaster scenario, as shown in Embodiment 6 of the present invention.

[0118] like Figure 7 As shown, the rainfall-induced landslide geological hazard scenario simulation and analysis system includes the following components:

[0119] A mountain modeling device is used to receive visual images corresponding to each side of a set mountain at the current moment, and to construct a 3D graphic model of the set mountain at the current moment based on each visual image and the closest distance of each camera mechanism that captured each visual image relative to the set mountain.

[0120] For example, receiving visual images corresponding to each side of a set mountain at the current moment, and constructing a 3D graphic model of the set mountain at the current moment based on each visual image and the closest distance of each camera mechanism that captured each visual image relative to the set mountain, includes: using the combination and fusion of each image of the set object from each side to obtain a 3D graphic model of the set object that can be displayed on a computer; if the fusion of some geographic data, such as the fusion of location information, is added, a 3D graphic model of the set object with geographic data that can be displayed on a computer can be further obtained.

[0121] In this way, the 3D graphic model of the object not only has a three-dimensional shape that is a combination and fusion of various sides, but each side of this three-dimensional shape is composed of pixels of the image of each side. These pixels of the side have three-dimensional coordinates, that is, coordinate values ​​in three directions of the three-dimensional coordinate system. Moreover, these side pixels of the 3D graphic model of the object also have geographic data, such as location information.

[0122] The data acquisition device is connected to the mountain modeling device and is used to acquire various customized model data of the mountain at the current moment based on the 3D graphic model of the mountain at the current moment.

[0123] Specifically, since the 3D graphic model of the mountain at the current moment can be displayed on the computer, various customized model data of the mountain at the current moment can be directly called and obtained from the computer, such as the coordinate values ​​of the three directions of these side pixels of the 3D graphic model of the object and its geographical data.

[0124] The target analysis device is used to output multiple related information of the set mountain, including its area, peak height, width, length, soil type number, and main crop type number;

[0125] Specifically, the output of multiple related information of the mountain is set, including the area, peak height, width, length, soil type number, and main crop type number. Different soil types have different soil type numbers. For example, red soil and black soil have different soil type numbers, and different main crop types have different main crop type numbers.

[0126] The intelligent prediction device is connected to the data acquisition device and the target analysis device respectively. It is used to use the landslide intelligent prediction model to intelligently predict whether landslide geological disasters will occur on each side of the set mountain in the next time interval with the current time as the start time. The prediction data is based on the total rainfall in the most recent time interval with the current time as the end time, the duration of the time interval, the customized model data of the set mountain at the current time, and multiple related information of the set mountain.

[0127] For example, if the current time is 5:30 PM and the duration of the time interval is 15 minutes, then the nearest time interval ending at the current time is the time interval from 5:15 PM to 5:30 PM, and the next time interval starting at the current time is the time interval from 5:30 PM to 5:45 PM.

[0128] The early warning triggering device is connected to the intelligent prediction device and is used to determine whether to trigger an early warning of landslide occurrence in the next time interval of the set mountain based on the intelligent prediction results;

[0129] In this way, by predicting in advance whether a landslide will occur in a designated mountain in the future, intelligent disaster management of the designated mountain can be achieved, providing valuable reference information for the evacuation of nearby residents or the closure of highways;

[0130] The landslide intelligent prediction model is a Hofit neural network after completing each learning action, that is, the architecture of the landslide intelligent prediction model is the network architecture of the Hofit neural network.

[0131] Among them, the collection of various customized model data of the current moment of the 3D graphic model of the set mountain includes: collecting the 3D coordinate values, curvature values, gray values ​​and gray gradient values ​​of each surface pixel of the 3D geographic model of the set mountain at the current moment as various customized model data of the set mountain at the current moment.

[0132] Specifically, the 3D coordinates, curvature values, gray values, and gray gradient values ​​of each surface pixel of the 3D geographic model of the mountain at the current moment are collected as customized model data of the mountain at the current moment. The gray value of each surface pixel is between 0 and 255.

[0133] The landslide prediction data for determining whether a landslide geological disaster will occur on each side of the mountain consists of a side code of the mountain side and an occurrence identifier indicating whether a landslide geological disaster will occur on the mountain side.

[0134] Specifically, the landslide prediction data for each side of the mountain, which sets the conditions for whether landslides will occur, provides the specific sides of the mountain where landslides are likely to occur.

[0135] Among them, determining whether to trigger a landslide occurrence warning for the next time interval of a set mountain based on intelligent prediction results includes: when the occurrence identifier in the landslide prediction data of whether a landslide geological disaster has occurred on a certain side of a set mountain indicates that a landslide geological disaster has occurred on that certain side of the mountain, executing a landslide occurrence warning for the next time interval of that certain side of the set mountain.

[0136] The process of receiving visual images corresponding to each side of the set mountain at the current moment, and constructing a 3D graphic model of the set mountain at the current moment based on each visual image and the closest distance of each camera that captured each visual image relative to the set mountain, includes: receiving visual images corresponding to each side of the set mountain at the current moment, and constructing a 3D graphic model of the set mountain at the current moment based on each visual image, the closest distance of each camera that captured each visual image relative to the set mountain, and the positioning data corresponding to each camera.

[0137] The landslide intelligent prediction model is a Hoffert neural network after completing each learning action, which includes: the number of learning actions of the Hoffert neural network is monotonically positively correlated with the number of each side.

[0138] For example, the number of learning actions of the Hofit neural network is monotonically positively correlated with the number of sides selected, including: when 6 sides are selected, the number of learning actions of the Hofit neural network is 600; when 8 sides are selected, the number of learning actions of the Hofit neural network is 700; when 10 sides are selected, the number of learning actions of the Hofit neural network is 800; when 12 sides are selected, the number of learning actions of the Hofit neural network is 900, and so on.

[0139] It is evident that the more aspects are selected, the more times the Hofit neural network learns actions, and the more refined and reliable the landslide intelligent prediction model becomes.

[0140] In each learning action performed on the Hofit neural network, the landslide prediction data, which indicates whether landslides have occurred on the various sides of a mountain within a certain historical time interval, are used as the output of the landslide intelligent prediction model. The total rainfall, duration of the time interval, customized model data of the mountain at the start time of the historical time interval, and various related information of the mountain are used as the input of the landslide intelligent prediction model to complete the learning action.

[0141] In addition, the present invention may also cite the following technical contents to highlight the significant technical advancements of the present invention:

[0142] The 3D coordinates, curvature values, gray values, and gray gradient values ​​of each surface pixel of the 3D geographic model of the set mountain at the current moment are collected as various customized model data of the set mountain at the current moment. The gray gradient value of each surface pixel is the standard deviation of the gray value set composed of the gray values ​​of each pixel around the surface pixel and the gray value of the surface pixel itself.

[0143] Specifically, the grayscale gradient value of each surface pixel is the standard deviation of the grayscale value set composed of the grayscale values ​​of each pixel surrounding the surface pixel and the grayscale value of the surface pixel itself. The grayscale values ​​of each pixel surrounding the surface pixel are the pixels adjacent to the surface pixel.

[0144] In addition, the process of collecting the 3D coordinate values, curvature values, gray values, and gray gradient values ​​of each surface pixel of the 3D geographic model of the set mountain at the current moment as customized model data of the set mountain at the current moment also includes: the 3D coordinate values ​​of each surface pixel are the coordinate values ​​of the surface pixel in three different directions, such as the coordinate values ​​of the X-axis direction, the Y-axis direction, and the Z-axis direction.

[0145] The foregoing description of exemplary embodiments of the invention is provided for illustrative and descriptive purposes. It is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Many modifications and variations will obviously be apparent to those skilled in the art. Exemplary embodiments were chosen and described in order to best illustrate the principles of the invention and its practical application, thereby enabling others skilled in the art to understand the various embodiments and variations of the invention suitable for the contemplated particular purpose. The scope of the invention is intended to be defined by the appended claims and their equivalents.

Claims

1. A method for simulating and analyzing rainfall-induced landslide geological disaster scenarios, characterized in that, The method includes: Receive visual images corresponding to each side of the set mountain at the current moment, and construct a 3D graphic model of the set mountain at the current moment based on each visual image and the closest distance of each camera mechanism that captured each visual image relative to the set mountain. Based on the current moment of the 3D graphic model of the designated mountain, collect various customized model data of the designated mountain at the current moment; The system outputs multiple related information about the mountain, including its area, peak height, width, length, soil type number, and main crop type number. The landslide intelligent prediction model uses the total rainfall, duration of the time interval, customized model data of the set mountain at the current moment, and multiple related information of the set mountain to intelligently predict whether landslide geological disasters will occur on each side of the set mountain in the next time interval with the current moment as the starting moment. The decision to trigger a landslide warning for the next time interval of the mountain is based on the results of intelligent prediction. The landslide intelligent prediction model is a Hoffert neural network after completing each learning action; Among them, the 3D coordinates, curvature values, gray values ​​and gray gradient values ​​of each surface pixel of the 3D geographic model of the set mountain at the current time are collected as various customized model data of the set mountain at the current time. Among them, the number of learning actions in the Hofit neural network is monotonically positively correlated with the number of each side. In each learning action performed on the Hofit neural network, the landslide prediction data of whether landslides have occurred on each side of a mountain within a certain historical time interval are used as the output of the landslide intelligent prediction model. The total rainfall, duration of the time interval, customized model data of the mountain at the start time of the certain historical time interval, and multiple related information of the mountain are used as the input of the landslide intelligent prediction model to complete the learning action. Wherein, the grayscale gradient value of each surface pixel is the standard deviation of the grayscale value set composed of the grayscale values ​​of each pixel surrounding the surface pixel and the grayscale value of the surface pixel itself, each pixel surrounding the surface pixel is each pixel adjacent to the surface pixel, and the 3D coordinate value of each surface pixel is the coordinate value of the surface pixel in three different directions.

2. The simulation and analysis method for rainfall-induced landslide geological disaster scenarios as described in claim 1, characterized in that: The landslide prediction data for determining whether a landslide geological hazard will occur on each side of a mountain consists of a side code for the mountain side and an occurrence identifier indicating whether a landslide geological hazard will occur on the mountain side. Among them, determining whether to trigger a landslide occurrence warning for the next time interval of a set mountain based on intelligent prediction results includes: when the occurrence identifier in the landslide prediction data of whether a landslide geological disaster has occurred on a certain side of a set mountain indicates that a landslide geological disaster has occurred on that certain side of the mountain, executing a landslide occurrence warning for the next time interval of that certain side of the set mountain. The process of receiving visual images corresponding to each side of a set mountain at the current moment, and constructing a 3D graphic model of the set mountain at the current moment based on each visual image and the closest distance of each camera that captured each visual image relative to the set mountain, includes: receiving visual images corresponding to each side of a set mountain at the current moment, and constructing a 3D graphic model of the set mountain at the current moment based on each visual image, the closest distance of each camera that captured each visual image relative to the set mountain, and the positioning data corresponding to each camera.

3. The simulation and analysis method for rainfall-induced landslide geological disaster scenarios as described in claim 2, characterized in that, After employing a landslide intelligent prediction model to intelligently predict whether landslide geological hazards will occur on each side of the set mountain in the next time interval (starting from the current time), based on the total rainfall, duration of the time interval, customized model data of the set mountain at the current time, and multiple related information of the set mountain, the method further includes: It receives landslide prediction data for each side of a set mountain within the next time interval starting from the current time, indicating whether a landslide geological disaster has occurred, and displays the landslide prediction data for each side of a set mountain within the next time interval starting from the current time in real time.

4. The simulation and analysis method for rainfall-induced landslide geological disaster scenarios as described in claim 2, characterized in that, After employing a landslide intelligent prediction model to intelligently predict whether landslide geological hazards will occur on each side of the set mountain in the next time interval (starting from the current time), based on the total rainfall, duration of the time interval, customized model data of the set mountain at the current time, and multiple related information of the set mountain, the method further includes: It receives landslide prediction data for each side of a mountain within the next time interval starting from the current time, indicating whether a landslide has occurred. It then wirelessly transmits this landslide prediction data to a remote geological disaster management server via a network communication link.

5. The simulation and analysis method for rainfall-induced landslide geological disaster scenarios as described in claim 2, characterized in that, After outputting the area, peak height, width, length, soil type number, and main crop type number of the designated mountain as multiple associated information of the designated mountain, the method further includes: Perform each learning action on the Hofit neural network to obtain the Hofit neural network after completing each learning action, and output it as the landslide intelligent prediction model.

6. The simulation and analysis method for rainfall-induced landslide geological disaster scenarios as described in claim 2, characterized in that, Before outputting the area, peak height, width, length, soil type number, and main crop type number of the designated mountain as multiple associated information of the designated mountain, the method further includes: Multiple different physical measurement components are used to measure the area, peak height, width, length, soil type, and main crop type of the designated mountain. Based on the measured soil type and main crop type of the designated mountain, the soil type number and main crop type number of the designated mountain are determined.

7. The simulation and analysis method for rainfall-induced landslide geological disaster scenarios as described in any one of claims 2-6, characterized in that: The monotonically positive correlation between the number of learning actions and the number of each side in the Hofit neural network includes: using a number mapping function to represent the monotonically positive correlation between the number of learning actions and the number of each side in the Hofit neural network. The method of using a frequency mapping function to represent the monotonically positive correlation between the number of learning actions of the Hoffett neural network and the number of each side includes: in the frequency mapping function, the number of each side is the input data of the frequency mapping function, and the number of learning actions of the Hoffett neural network corresponding to the number of each side is the output data of the frequency mapping function.

8. The simulation and analysis method for rainfall-induced landslide geological disaster scenarios as described in any one of claims 2-6, characterized in that: The landslide intelligent prediction model uses the total rainfall, duration of the time interval, customized model data of the set mountain at the current time, and multiple related information of the set mountain to intelligently predict whether landslide geological disasters will occur on each side of the set mountain in the next time interval starting from the current time. The prediction data includes: the total rainfall, duration of the time interval, customized model data of the set mountain at the current time, and multiple related information of the set mountain are synchronously input into the landslide intelligent prediction model. The intelligent landslide prediction model uses the total rainfall, duration of the time interval, customized model data of the set mountain at the current moment, and multiple related information of the set mountain to intelligently predict whether landslide geological disasters will occur on each side of the set mountain in the next time interval starting from the current moment. This prediction data also includes: executing the intelligent landslide prediction model to obtain the output of the intelligent landslide prediction model, which shows whether landslide geological disasters will occur on each side of the set mountain in the next time interval starting from the current moment. The process of synchronously inputting the total rainfall, duration of the time interval, customized model data of the set mountain at the current moment, and multiple related information of the set mountain into the landslide intelligent prediction model includes: performing octal value conversion processing on the total rainfall, duration of the time interval, customized model data of the set mountain at the current moment, and multiple related information of the set mountain before synchronously inputting them into the landslide intelligent prediction model. Among them, the execution of the landslide intelligent prediction model to obtain the landslide prediction data of each side of the mountain set in the next time interval with the current time as the start time, which is a representation of whether a landslide geological disaster will occur on each side of the mountain set in the next time interval with the current time as the start time, includes: the landslide prediction data of each side of the mountain set in the next time interval with the current time as the start time, which is a representation of octal values; Specifically, before synchronously inputting the total rainfall, duration of the time interval, various customized model data of the set mountain at the current moment, and various related information of the set mountain into the landslide intelligent prediction model, octal value conversion processing is performed on the total rainfall, duration of the time interval, various customized model data of the set mountain at the current moment, and various related information of the set mountain, respectively. This includes: using a first programmable logic chip to implement octal value conversion processing, and using a second programmable logic chip to implement synchronous input. Before synchronously inputting the total rainfall, duration of the time interval, various customized model data of the set mountain at the current moment, and various related information of the set mountain into the landslide intelligent prediction model, the octal value conversion processing of the total rainfall, duration of the time interval, various customized model data of the set mountain at the current moment, and various related information of the set mountain at the current moment also includes: the first programmable logic chip and the second programmable logic chip are connected and share the same clock generator and the same parameter configuration interface.

9. A simulation and analysis system for rainfall-induced landslide geological disaster scenarios, characterized in that, The system includes: A mountain modeling device is used to receive visual images corresponding to each side of a set mountain at the current moment, and to construct a 3D graphic model of the set mountain at the current moment based on each visual image and the closest distance of each camera mechanism that captured each visual image relative to the set mountain. The data acquisition device is connected to the mountain modeling device and is used to acquire various customized model data of the mountain at the current moment based on the 3D graphic model of the mountain at the current moment. The target analysis device is used to output multiple related information of the set mountain, including its area, peak height, width, length, soil type number, and main crop type number; The intelligent prediction device is connected to the data acquisition device and the target analysis device respectively. It is used to use the landslide intelligent prediction model to intelligently predict whether landslide geological disasters will occur on each side of the set mountain in the next time interval with the current time as the start time. The prediction data is based on the total rainfall in the most recent time interval with the current time as the end time, the duration of the time interval, the customized model data of the set mountain at the current time, and multiple related information of the set mountain. The early warning triggering device is connected to the intelligent prediction device and is used to determine whether to trigger an early warning of landslide occurrence in the next time interval of the set mountain based on the intelligent prediction results; The landslide intelligent prediction model is a Hoffert neural network after completing each learning action; Among them, the 3D coordinates, curvature values, gray values ​​and gray gradient values ​​of each surface pixel of the 3D geographic model of the set mountain at the current time are collected as various customized model data of the set mountain at the current time. Among them, the number of learning actions in the Hofit neural network is monotonically positively correlated with the number of each side. In each learning action performed on the Hofit neural network, the landslide prediction data of whether landslides have occurred on each side of a mountain within a certain historical time interval are used as the output of the landslide intelligent prediction model. The total rainfall, duration of the time interval, customized model data of the mountain at the start time of the certain historical time interval, and multiple related information of the mountain are used as the input of the landslide intelligent prediction model to complete the learning action. Wherein, the grayscale gradient value of each surface pixel is the standard deviation of the grayscale value set composed of the grayscale values ​​of each pixel surrounding the surface pixel and the grayscale value of the surface pixel itself, each pixel surrounding the surface pixel is each pixel adjacent to the surface pixel, and the 3D coordinate value of each surface pixel is the coordinate value of the surface pixel in three different directions.

Images