A multi-point remote synchronous geological disaster warning method and system based on rainfall observation

Abstract

The present invention discloses a multi-point remote synchronous geological disaster early warning method and system based on rainfall observation, belonging to the field of disaster early warning. The method includes: dividing the target area into multiple monitoring nodes, arranging rainfall detection stations and soil humidity sensors at each monitoring node, and installing slope sensors in landslide-sensitive areas; coupling the collected data in terms of time, and then sending the collected data to the cloud platform; after receiving the sensor data, the cloud platform first decompresses the data, and then inputs the collected sensor data into a comprehensive prediction model to dynamically update the soil humidity; a geological condition database of each geological disaster-prone and high-incidence area is included in the cloud platform, and based on the change of soil humidity and combined with the data in the geological condition database, the geological disaster risk is evaluated. The present invention can make full use of multi-source heterogeneous data, so as to comprehensively evaluate the geological disaster risk.

Description

Technical Field

[0001] The present invention relates to the field of disaster prediction, and particularly to a multi-point remote synchronous geological disaster warning method and system based on rainfall observation. Background Art

[0002] Geological disaster warning technology aims to predict the potential threats of various natural disasters in advance and reduce the impact of disasters. The existing geological disaster warning methods mainly rely on limited environmental monitoring and relatively simple data processing technologies, and there are obvious deficiencies in the accuracy of disaster prediction, the timeliness of warning signals, and the effectiveness of emergency responses. Traditional methods often rely on a single parameter index, ignoring the comprehensive influence among multiple parameters, making it difficult to make full use of multi-source heterogeneous data and accurately and dynamically predict the probability of geological disasters, resulting in low accuracy and reliability of warning information.

[0003] In addition, the existing technology has limitations in considering the diversity and complexity of scenarios and is difficult to adapt to different types of geological disasters and environmental conditions. At the same time, the existing methods have limitations in prediction models, making it difficult to accurately predict the occurrence time and location of geological disasters, and the emergency response measures are often insufficient, exacerbating the impact of disasters. Therefore, there is an urgent need for a new geological disaster warning method that can integrate real-time rainfall data, geological condition data, and meteorological information from multiple locations to dynamically predict the probability of geological disasters in different regions, improve the accuracy and timeliness of warnings, and provide warnings as early as possible before disasters occur. Summary of the Invention

[0004] One of the objectives of the present invention is to provide a multi-point remote synchronous geological disaster warning method based on rainfall observation to solve the problem in the prior art that it is difficult to accurately predict the occurrence time and location of geological disasters.

[0005] The present invention is achieved through the following technical solutions. A multi-point remote synchronous geological disaster early warning method based on rainfall observation includes the following steps: S100. Divide the target area into multiple monitoring nodes, arrange rainfall detection stations at each monitoring node, and at the same time, arrange soil moisture sensors in the soil layers at different depths at each monitoring node. At the same time, install slope sensors in landslide-sensitive areas to monitor the slope change of the area in real time; S200. Correlate the soil moisture data at different depths, the data of the rainfall detection stations, and the data collected by the slope sensors, and upload them uniformly. By coupling the collected precipitation information, soil moisture data, and the data of the slope sensors in terms of time, and then sending the data after time coupling to the cloud platform; S300. After the cloud platform receives the sensor data, first decompress the data, and then input the collected sensor data into the comprehensive prediction model to dynamically update the soil moisture; S400. The cloud platform contains a geological condition database for each area where geological disasters are prone to occur. According to the dynamic update of the soil moisture calculated in step S300, based on the change of the soil moisture, combine the data in the geological condition database to judge whether the predetermined risk threshold is reached, and evaluate the geological disaster risk.

[0006] Further, the soil moisture sensors are respectively arranged at 1 cm, 15 cm, 30 cm, 45 cm, and 60 cm below the soil surface layer.

[0007] Further, sending the collected data to the cloud platform includes:

[0008] S210. Each rainfall detection station is provided with a corresponding wireless network node. Converge the multiple wireless network nodes in the target area into a wireless network node set, and integrate the received sensor data in this wireless network node set;

[0009] S220. Stamp the collected data with timestamps, check whether there are duplicates in the data according to the timestamps, and delete the duplicate sensor data to complete the cleaning of the data;

[0010] S230. After the cleaning is completed, compress the data, and send the compressed data to the cloud platform through the TCP / IP protocol.

[0011] Further, the comprehensive prediction model is composed of a rainfall-soil saturation model and a state space model. The rainfall-soil saturation model is used to describe the evolution of soil moisture in space and time, and then discretize the rainfall-soil saturation model into the state space model.

[0012] Further, the rainfall-soil saturation model is constructed based on partial differential equations, including the following steps:

[0013] S311. Describe the influence of rainfall on soil moisture at different locations based on partial differential equations. The change of soil moisture is affected by external rainfall and internal water transport in the soil. The diffusion equation is used to describe the spatial propagation of soil moisture:

[0014] ,

[0015] where, is used to represent the rate of change of soil moisture S(x, t) with time; D is used to represent the diffusion coefficient; is the Laplace operator, representing the second-order derivative of soil moisture in space and describing the diffusion behavior of moisture in space;

[0016] S312. At the same time, since soil moisture is also directly affected by external factors, the influence of rainfall is represented by introducing a source term f(R(x,t), S(x,t)). Adding the source term to the diffusion equation gives:

[0017] ,

[0018] where, f(R(x,t), S(x,t)) represents the influence of rainfall R(x,t) on soil moisture S(x,t);

[0019] S313. The slope data θ(x,t) collected by the slope sensor is used to represent the angle of ground inclination. In places with a large slope, water flow is more likely to concentrate, thus triggering geological disasters.

[0020] The influence of the slope is reflected by correcting f(R(x,t), S(x,t)). Taking the slope data θ(x,t) as a regulating factor,

[0021] Finally, the rainfall-soil saturation model based on partial differential equations is:

[0022] .

[0023] Furthermore, the state space model is jointly composed of a state equation and an observation equation. The state space model uses state variables to describe the dynamic behavior of the entire system and describes the change of the state through a state transition equation. Among them,

[0024] The state equation is used to describe the evolution of the system state, and this equation is shown as follows:

[0025] ,

[0026] where, x(t) represents the state of the system, u(t) represents the input, w(t) is noise or system uncertainty, and A and B are system parameters;

[0027] The observation equation is used to describe obtaining observation data from the state of the system, and the equation is as follows:

[0028] y(t)=Cx(t)+v(t)

[0029] y(t) is the observation data, C is the mapping matrix, and v(t) is the observation noise.

[0030] Furthermore, discretizing the rainfall - soil saturation model into the state - space model includes the following steps:

[0031] S321. Discretize the rainfall - soil saturation model by discretizing time t into discrete time points t k , t k =kΔt, where Δt is the time step, and discretize space x into grid points x i to obtain the discrete - form state equation:

[0032] ,

[0033] where the soil moisture S(x i , t k ) becomes the state variable in the state - space model,

[0034] The discretization of the partial differential equation is used to describe the state transition, indicating how the soil moisture transfers from time t k to t k+1 ;

[0035] S322. The observation equation describes obtaining the measurement results of the sensor from the state of the soil moisture. The measurement data of multiple sensors is related to the state variable, so the observation equation can be written as:

[0036] ,

[0037] where y(t k ) is the observation data of the sensor, C is the mapping matrix, and v(t k ) is the observation noise.

[0038] Furthermore, determining whether the disaster - triggering threshold is reached includes: S410. Soil - saturation threshold judgment: Determine the threshold of soil moisture according to the data in the geological - condition database for the situation of the monitoring location; S420. High - risk area identification: Construct the spatial distribution of soil moisture and geological information such as slope through sensor data, and identify high - risk areas in combination with the geological data in the geological - condition database; S430. Geological - disaster prediction and warning trigger: According to the identified high - risk areas, when it is detected that the soil moisture in the high - risk areas exceeds the predetermined threshold, trigger a warning.

[0039] On the other hand, the present invention provides a multi-point remote synchronous geological disaster warning system based on rainfall observation. The warning system includes: a data acquisition module, a data transmission module, and a cloud platform module. Among them, the data acquisition module is configured to divide the target area into multiple monitoring nodes, arrange rainfall detection stations at each monitoring node, and at the same time arrange soil moisture sensors in the soil layers at different depths at each monitoring node, and install slope sensors in landslide-sensitive areas to monitor the slope change of the area in real time; the data transmission module is connected to the data acquisition module and is configured to correlate the soil moisture data at different depths, the data of the rainfall detection stations, and the data collected by the slope sensors, and perform unified upload. By coupling the collected precipitation information, soil moisture data, and the data of the slope sensors in terms of time, and then sending the data after time coupling to the cloud platform; the cloud platform module is connected to the data transmission module and is configured to include: a comprehensive prediction model sub-module and a warning prompt sub-module in the cloud platform module. The comprehensive prediction model sub-module is configured to receive the sensor data collected by the data acquisition module and perform dynamic update of the soil moisture through the comprehensive prediction model; the warning prompt sub-module is connected to the comprehensive prediction model sub-module and is configured to judge whether the predetermined risk threshold is reached based on the dynamic update of the soil moisture obtained by the comprehensive prediction model sub-module, combine the data in the geological condition database according to the change of the soil moisture, evaluate the geological disaster risk and give a warning.

[0040] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0041] 1. By integrating real-time rainfall data, geological condition data, and meteorological information from multiple locations, the present invention can make full use of multi-source heterogeneous data, thereby comprehensively evaluating the geological disaster risk. At the same time, by combining the rainfall and soil saturation model constructed in the present invention and the state space model, it can dynamically predict the time, location, and probability of geological disasters occurring in different regions, thus improving the accuracy of prediction.

[0042] 2. By arranging rainfall detection stations, soil moisture sensors, and slope sensors in the target area, the present invention can adapt to the diversity and complexity of different types of geological disasters and environmental conditions. At the same time, according to the calculated change of soil moisture, it combines the geological condition database to judge whether the predetermined threshold is reached, identifies high-risk areas, triggers warnings at different levels, improves the accuracy and reliability of warning information, and takes emergency response measures in a timely manner, which can effectively reduce the impact of disasters. Description of the Drawings

[0043] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, form a part of this application, and do not constitute a limitation to the embodiments of the present invention. In the drawings:

[0044] Figure 1 This is the flowchart of the method provided in Embodiment 1 of the present invention. Specific implementation manner

[0045] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. The components of the embodiments of the present invention usually described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations.

[0046] Embodiment 1

[0047] In this embodiment, a multi-point remote synchronous geological disaster warning method based on rainfall observation is provided. The warning method disclosed in this embodiment constructs a rainfall and soil saturation model by integrating real-time rainfall data, geological condition data, and meteorological information of multiple locations, and combines it with the state space model, so as to dynamically predict the probability of geological disasters (such as landslides, debris flows, etc.) occurring in different regions, and provide early warnings as early as possible before the disasters occur. Figure 1 The flowchart of the warning method in this embodiment is shown. It can be seen from the figure that this embodiment includes the following steps:

[0048] Step 1: In the target area, that is, the area where geological disasters are prone to occur frequently, it is divided into multiple monitoring nodes, and rainfall detection stations are arranged in each monitoring node to monitor the precipitation of each monitoring point in real time, and the data is uploaded to the data transmission module through the wireless communication module. The layout of the rainfall detection stations should cover the high-risk areas where disasters may occur, and the terrain, climate, and historical disaster data should be considered to optimize the layout.

[0049] In addition, appropriate soil moisture sensors should be selected according to the soil types of different monitoring nodes, and soil moisture sensors should be arranged in the soil at different depths to collect soil moisture data at different depths. For example, in this embodiment, soil moisture sensors can be arranged at 1 cm, 15 cm, 30 cm, 45 cm, and 60 cm on the soil surface respectively. Through these soil moisture sensors arranged at different depths, the soil moisture data at different depths is correlated with the data collected by the rainfall detection stations and uploaded uniformly to facilitate subsequent data calculation.

[0050] Considering the landslide-sensitive areas, slope sensors can also be installed in the landslide-sensitive areas to monitor the slope changes of the area in real time. Combining the slope change information with the rainfall and soil moisture data helps to accurately predict the occurrence of landslides.

[0051] Step 2: Couple the collected precipitation information with the soil moisture data in terms of time, and then send it to the cloud platform.

[0052] Specifically, sending the collected data to the cloud platform includes the following sub-steps:

[0053] First, a corresponding wireless network node is set in each rainfall detection station arranged by the monitoring nodes. Multiple wireless network nodes in the same high-incidence area of geological disasters are aggregated into a wireless network node set, and the sensor data received in this wireless network node set is integrated. Specifically, it includes: First, timestamp the collected data, check whether there are duplicates in the data according to the timestamp, and delete the duplicate sensor data, thus completing the cleaning of the data. After the cleaning is completed, compress the data, and send the compressed data to the cloud platform through the TCP / IP protocol.

[0054] Step 3: After the cloud platform receives the sensor data, first decompress the data. An integrated prediction model is built in the cloud platform. Input the collected sensor data into the integrated prediction model. The integrated prediction model consists of a rainfall-soil saturation model and a state space model. The rainfall-soil saturation model is used to describe the evolution of soil moisture in space and time, and then discretize the rainfall-soil saturation model into a state space model to perform dynamic update of soil moisture.

[0055] Specifically, the rainfall-soil saturation model describes the propagation characteristics of rainfall in time and space by using partial differential equations. Its core is to establish the influence relationship of rainfall on geological conditions such as soil moisture and slope. In order to simulate the interaction between rainfall and factors such as soil moisture and slope, the following steps can be used to construct the rainfall-soil saturation model:

[0056] 1) First, use partial differential equations to describe the influence of rainfall on soil moisture at different locations (i.e., the data monitored by sensors). To simplify the analysis, assume that the soil moisture S(x, t) depends on the spatial position x and time t, and it is affected by rainfall R(x,t), the soil moisture itself S(x, t), and other factors (such as the physical properties of the soil, slope, etc.).

[0057] Then it is necessary to determine the main factors affecting soil moisture:

[0058] Influence of rainfall: Rainfall directly increases the moisture content of the soil, resulting in an increase in soil moisture.

[0059] Diffusion of soil moisture: Soil moisture is not evenly distributed. Affected by soil permeability, soil structure, etc., it will spread in space.

[0060] External factors: such as the water permeability of the soil, slope, vegetation cover, etc. These factors will affect the moisture distribution after rainfall.

[0061] Based on these influencing factors, a partial differential equation suitable for describing the variation of soil moisture with time and space can be constructed.

[0062] 2) The change of soil moisture is affected by both external rainfall and internal water transport (diffusion) in the soil. Therefore, the diffusion equation can be used to describe the spatial propagation of soil moisture:

[0063] ,

[0064] where, is used to represent the rate of change of soil moisture S(x, t) with time, that is, the change of soil moisture at time t; D is used to represent the diffusion coefficient, that is, the speed of water diffusion in the soil, which depends on the properties of the soil (such as permeability, etc.); is the Laplace operator, representing the second-order derivative of soil moisture in space, describing the diffusion behavior of moisture in space. If the soil moisture is higher at a certain point and lower around it, water will diffuse from the high-moisture area to the low-moisture area.

[0065] Through this diffusion equation, it can be expressed that in a small area, the change in the amount of water is equal to the water flow rate entering the area minus the water flow rate leaving the area. For soil, the process of water flow is mainly affected by the water permeability of the soil, so we can approximately describe it with the diffusion equation.

[0066] 3) However, in actual situations, soil moisture is not only the result of spatial diffusion but also directly affected by external factors (such as rainfall). The rainfall R(x,t) will increase the soil moisture in a certain area, and in this embodiment, we represent the influence of rainfall through a source term f(R(x,t), S(x,t)). Adding this source term to the diffusion equation gives:

[0067] ,

[0068] where, f(R(x,t), S(x,t)) represents the influence of rainfall R(x,t) on soil moisture S(x,t). It should be noted that the specific form of this function is usually finally determined based on actual sensor data and can be modeled according to the relationship between rainfall and soil moisture. Generally speaking, the greater the rainfall, the higher the soil moisture, and this influence has a non-linear characteristic. To make this function more accurate, the initial moisture of the soil and the infiltration characteristics of the soil can also be considered.

[0069] 4) In this embodiment, it is considered that the change of soil humidity is also affected by factors such as terrain slope and soil type. The slope data θ(x,t) collected by the slope sensor represents the angle of the ground inclination. In places with a larger slope, water flow is more likely to concentrate, which may trigger disasters such as landslides or debris flows.

[0070] The influence of the slope can be reflected by modifying f(R(x,t), S(x,t)). For example, in areas with a larger slope, rainfall may cause faster water flow and stronger soil humidity changes. Therefore, the slope data θ(x,t) can be used as a regulating factor.

[0071] Therefore, the final partial differential equation model obtained is:

[0072] 。

[0073] In this model, the role of the sensor data is as follows:

[0074] The data R(x,t) collected by the rainfall detection station provides real-time rainfall data, which is one of the external driving factors in the model.

[0075] The soil humidity sensor data S(x,t), which is used to monitor the soil humidity at different locations and depths, helps the model update the soil humidity state.

[0076] The slope sensor data θ(x,t), which is used to provide terrain slope information and helps adjust the intensity of the impact of rainfall on soil humidity.

[0077] By coupling these data and combining them, a soil humidity change model based on partial differential equations is obtained. This model describes how rainfall affects the evolution of soil humidity through spatial diffusion and precipitation input, and considers the influence of factors such as slope on water propagation. This model can be used to predict whether the soil humidity in a certain area reaches the conditions to trigger geological disasters (such as landslides, debris flows, etc.) under specific rainfall and terrain conditions.

[0078] In the above content, the sensor data describes the change process of soil humidity S(x,t) in space and time through partial differential equations, helping us understand and predict the evolution of soil humidity under different conditions. In this embodiment, after obtaining the rainfall-soil saturation model based on partial differential equations, a state space model is constructed to describe the influence of rainfall on the state evolution of the dynamic system of geological disasters. The core idea of the state space model is to describe the state of the system as a vector in a high-dimensional space and describe the change of the system state over time through a state transition equation.

[0079] Specifically, in this embodiment, the state space model is constructed through the following steps:

[0080] In the state space model, state variables are used to describe the dynamic behavior of the entire system, and the state transition equation is used to describe the change of the state. In this embodiment, the state space model is jointly composed of a state equation and an observation equation.

[0081] Among them, the state equation is used to describe the evolution of the system state, and this equation is shown as follows:

[0082] ,

[0083] x(t) represents the state of the system, u(t) represents the input (external drive), w(t) is the noise or system uncertainty, and A and B are the parameters of the system.

[0084] The observation equation is used to describe how to obtain the observation data from the state of the system, and this equation is shown as follows:

[0085] y(t)=Cx(t)+v(t)

[0086] y(t) is the observation data, C is the mapping matrix, and v(t) is the observation noise.

[0087] In this embodiment, the state x(t) in the state space model represents variables such as soil moisture or disaster risk, the input u(t) is the rainfall, and the observation data y(t) is the real-time data obtained through sensors, such as the measurement results of soil moisture sensors and rainfall stations.

[0088] It should be noted that the rainfall - soil saturation model based on partial differential equations and the state space model are two different modeling frameworks for solving different types of problems. The rainfall - soil saturation model based on partial differential equations describes the evolution of a continuous system, especially the spatial and temporal evolution of the system. It can well capture the propagation and diffusion of factors such as soil moisture and rainfall in space. The state space model, on the other hand, focuses more on the description of discrete - time systems, emphasizing how to perform reasoning and prediction through the state of the system and the observation data, and is especially suitable for time - series analysis and dynamic prediction.

[0089] In the prediction of geological disaster risks provided in this embodiment, by taking the partial differential equation model as a component of the state space model, it helps to describe the state transition and spatial evolution of the system. Specifically, the partial differential equation is used to provide the dynamics of the state transition in the state space model, while the state space model provides a framework for how to estimate and update these states based on the observation data for the final prediction of geological disaster risks.

[0090] Specifically, combining the two different models, namely the rainfall - soil saturation model of partial differential equations and the state - space model, includes the following steps:

[0091] First, the rainfall - soil saturation model describes the evolution of soil moisture S(x,t) in space x and time t. Regarding this model as the state equation of a continuous system means that the change in soil moisture is a continuous process, affected by factors such as spatial diffusion, rainfall, and terrain.

[0092] To incorporate this equation into the state - space model, it is first necessary to discretize this equation to meet the requirements of discrete time steps. For example, discretize time t into discrete time points t k (t k = kΔt, where Δt is the time step), and discretize space x into grid points x i (for example, grid division based on sensor positions). The discrete - form state equation can be obtained:

[0093] , at this time, the soil moisture S(x i , t k ) becomes the state variable in the state - space model. The state transition here is described by the discretization of the partial differential equation, indicating how the soil moisture transfers from time t k to t k+1 .

[0094] In the state - space model, the observation equation describes how to obtain the sensor measurement results from the state of soil moisture. And multiple sensor measurement data are related to the state variable (i.e., soil moisture). For example, the measurement value of a soil moisture sensor can be directly associated with the state variable S(x i , t k ), and the reading of a rainfall sensor affects the rainfall R(x i , t k ), thus affecting the dynamics of the system. So the observation equation can be written as:

[0095] , where y(t k ) is the sensor observation data (such as the reading of a soil moisture sensor), C is the mapping matrix (usually a parameter related to sensor positions and measurement accuracy, such as the mapping matrix from the sensor to soil moisture), and v(t k ) is the observation noise.

[0096] By combining the rainfall-soil saturation model and the state space model, the spatial propagation and temporal evolution of soil moisture, i.e., the dynamic behavior of the system, are described by the rainfall-soil saturation model. The state space model mainly estimates and predicts the system state through observation data and adjusts the model parameters according to real-time data, so as to provide more accurate disaster risk prediction.

[0097] The combination of these two models enables the consideration of the following aspects simultaneously when conducting disaster prediction: the diffusion process in space, i.e., how soil moisture diffuses between different locations; the changes over time, i.e., how soil moisture is affected by external factors such as rainfall over time; real-time observation data, i.e., using the data obtained by sensors to dynamically adjust the model and provide accurate soil moisture estimation, so as to comprehensively evaluate the disaster risk.

[0098] Step 4: The geological condition database of each geological disaster-prone and high-incidence area is included in the cloud platform. Based on the dynamic update of the soil moisture calculated in Step 3, and in combination with the data in the geological condition database, it is judged whether the predetermined risk threshold is reached to evaluate the geological disaster risk.

[0099] Specifically, judging whether the disaster trigger threshold is reached includes:

[0100] 1) Soil saturation threshold judgment: Set the threshold of soil moisture according to the specific situation of the monitoring location determined by the data in the geological condition database. When this threshold is exceeded, a warning can be issued (for example, when the soil moisture exceeds a certain critical value, landslides or floods may be triggered).

[0101] 2) High-risk area identification: Construct the spatial distribution of soil moisture and geological information such as slope through sensor data, and identify high-risk areas in combination with the geological data in the geological condition database. For example, areas with a larger slope are more likely to have landslides after rainfall.

[0102] 3) Geological disaster prediction and warning trigger: According to the identified high-risk areas, when it is detected that the soil moisture in the high-risk areas exceeds the predetermined threshold, a warning is triggered. Different levels of warnings (such as green, yellow, and red warnings) can also be set to notify local residents or relevant departments according to the level of risk.

[0103] The specific implementation manners described above further elaborate on the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above is only the specific implementation manners of the present invention and is not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A multi-point remote synchronous geological disaster early warning method based on rainfall observation, characterized in that: The early warning method comprises: S100, dividing the target area into multiple monitoring nodes, and arranging a rainfall detection station at each monitoring node, and arranging soil moisture sensors in soil layers of different depths at each monitoring node, and installing slope sensors in landslide-sensitive areas to monitor slope changes in the area in real time; S200, associating soil moisture data at different depths, rainfall detection station data, and data collected by the slope sensor, and uploading them uniformly, by temporally coupling the collected precipitation information, soil moisture data, and slope sensor data, and then sending the temporally coupled data to the cloud platform; S300, after receiving the sensor data on the cloud platform, the data is first decompressed, and then the collected sensor data is input into the comprehensive prediction model to dynamically update the soil moisture; S400, the cloud platform contains a geological condition database of various areas prone to geological disasters, and according to the dynamic update of the soil moisture calculated in step S300, based on the change of soil moisture, combined with the data in the geological condition database, it is determined whether a predetermined risk threshold is reached, and the geological disaster risk is evaluated; The integrated prediction model consists of a rainfall-soil saturation model and a state-space model. The rainfall-soil saturation model is used to describe the evolution of soil moisture in space and time, and then the rainfall-soil saturation model is discretized into the state-space model. The rainfall-soil saturation model is constructed based on partial differential equations and includes the following steps: S311. Based on partial differential equations, the effect of rainfall on soil moisture at different locations is described. The change of soil moisture is affected by both external rainfall and internal moisture transmission in the soil. The diffusion equation is used to describe the spatial propagation of soil moisture: ， in, It is used to represent the rate of change of soil moisture S(x, t) over time; D is used to represent the diffusion coefficient; is the Laplace operator, which represents the second-order derivative of soil moisture in space and describes the diffusion behavior of moisture in space; S312. At the same time, since soil moisture is also directly affected by external factors, the source term f(R(x, t), S(x, t)) is introduced to represent the influence of rainfall. The source term is added to the diffusion equation to obtain: ， Where, f(R(x,t), S(x,t)) represents the effect of rainfall R(x,t) on soil moisture S(x,t); S313, the slope data θ(x, t) collected by the slope sensor is used to represent the angle of ground inclination. In places with larger slopes, water flow is more likely to concentrate, thereby causing geological disasters. The influence of slope is reflected by modifying f(R(x,t), S(x,t)), using the slope data θ(x,t) as an adjustment factor. Finally, the rainfall-soil saturation model based on partial differential equation is obtained: 。 2. The multi-point remote synchronous geological disaster early warning method based on rainfall observation according to claim 1 is characterized in that: The soil moisture sensors are arranged at 1 cm, 15 cm, 30 cm, 45 cm and 60 cm below the soil surface, respectively.

3. The multi-point remote synchronous geological disaster early warning method based on rainfall observation according to claim 1 is characterized in that: The sending of the collected data to the cloud platform includes: S210, a corresponding wireless network node is provided in each rainfall detection station, multiple wireless network nodes in the target area are aggregated into a wireless network node set, and the received sensor data is integrated in the wireless network node set; S220, adding a timestamp to the collected data, checking whether there is duplication in the data according to the timestamp, deleting the duplicate sensor data, and completing data cleaning; S230, after cleaning is completed, the data is compressed and sent to the cloud platform via TCP / IP protocol.

4. The multi-point remote synchronous geological disaster early warning method based on rainfall observation according to claim 1 is characterized in that: The state space model is composed of state equations and observation equations. The state space model uses state variables to describe the dynamic behavior of the entire system and describes the change of state through state transition equations, where: The state equation is used to describe the evolution of the system state, which is shown as follows: ， Where x(t) represents the state of the system, u(t) represents the input, w(t) is the noise or system uncertainty, and A and B are the parameters of the system; The observation equation is used to describe the acquisition of observation data from the state of the system, which is shown as follows: y(t)=Cx(t)+v(t), y(t) is the observation data, C is the mapping matrix, and v(t) is the observation noise.

5. The multi-point remote synchronous geological disaster early warning method based on rainfall observation according to claim 1 is characterized in that: The step of discretizing the rainfall-soil saturation model into a state space model comprises the following steps: S321, discretize the rainfall-soil saturation model by discretizing time t into discrete time points t k , t k = kΔt, where Δt is the time step, discretizing the space x into grid points x i The state equation is obtained in discrete form: , Among them, soil moisture S(x i , t k ) becomes the state variable in the state space model, The discretization of the partial differential equation is used to describe the state transition, which represents how the soil moisture changes from time t k Transfer to k+1 ; S322, the observation equation describes the sensor measurement results obtained from the state of soil moisture. Multiple sensor measurement data are related to the state variables, so the observation equation can be written as: ， Among them, y(t k ) is the sensor observation data, C is the mapping matrix, v(t k ) is the observation noise.

6. The multi-point remote synchronous geological disaster early warning method based on rainfall observation according to claim 1 is characterized in that: The determination of whether the disaster triggering threshold is reached includes: S410, soil saturation threshold determination: determining the soil moisture threshold of the monitoring site according to the data in the geological condition database; S420, high-risk area identification: constructing the spatial distribution of soil moisture and geological information such as slope through sensor data, and identifying high-risk areas in combination with geological data in the geological condition database; S430, geological disaster prediction and warning triggering: based on the identified high-risk areas, when it is detected that the soil moisture in the high-risk areas exceeds a predetermined threshold, a warning is triggered.

7. A multi-point remote synchronous geological disaster early warning system based on rainfall observation, characterized in that: The early warning system includes: Data collection module, data transmission module and cloud platform module, among which, The data acquisition module is configured to divide the target area into multiple monitoring nodes, and arrange a rainfall detection station at each monitoring node, and arrange soil moisture sensors in soil layers of different depths at each monitoring node, and install slope sensors in landslide-sensitive areas to monitor the slope changes of the area in real time; The data transmission module is connected to the data acquisition module and is configured to associate soil moisture data at different depths, rainfall detection station data, and data collected by the slope sensor, and upload them uniformly, by temporally coupling the collected precipitation information, soil moisture data, and slope sensor data, and then sending the temporally coupled data to the cloud platform; The cloud platform module is connected to the data sending module and is configured as follows: Comprehensive prediction model submodule and early warning prompt submodule, The comprehensive prediction model submodule is configured to receive the sensor data collected by the data acquisition module and dynamically update the soil moisture through the comprehensive prediction model; The early warning submodule is connected to the comprehensive prediction model submodule and is configured to dynamically update the soil moisture obtained by the comprehensive prediction model submodule, determine whether a predetermined risk threshold is reached based on the change of soil moisture and the data in the geological condition database, and assess the geological disaster risk and issue an early warning; The comprehensive prediction model submodule consists of a rainfall-soil saturation model and a state-space model. The rainfall-soil saturation model is used to describe the evolution of soil moisture in space and time, and then the rainfall-soil saturation model is discretized into the state-space model. The rainfall-soil saturation model is constructed based on partial differential equations and includes the following steps: S311. Based on partial differential equations, the effect of rainfall on soil moisture at different locations is described. The change of soil moisture is affected by both external rainfall and internal moisture transmission in the soil. The diffusion equation is used to describe the spatial propagation of soil moisture: ， in, It is used to represent the rate of change of soil moisture S(x, t) over time; D is used to represent the diffusion coefficient; is the Laplace operator, which represents the second-order derivative of soil moisture in space and describes the diffusion behavior of moisture in space; S312. At the same time, since soil moisture is also directly affected by external factors, the source term f(R(x, t), S(x, t)) is introduced to represent the influence of rainfall. The source term is added to the diffusion equation to obtain: ， Where, f(R(x,t), S(x,t)) represents the effect of rainfall R(x,t) on soil moisture S(x,t); S313, the slope data θ(x, t) collected by the slope sensor is used to represent the angle of ground inclination. In places with larger slopes, water flow is more likely to concentrate, thereby causing geological disasters. The influence of slope is reflected by modifying f(R(x,t), S(x,t)), using the slope data θ(x,t) as an adjustment factor. Finally, the rainfall-soil saturation model based on partial differential equation is obtained: 。

Images