A method and system for early warning of critical landslides caused by rainfall

Abstract

This invention relates to the technical field of geological disaster early warning, and in particular to a method and system for critical early warning of rainfall-induced landslides. The method includes acquiring multi-source geographic data of the target area and generating gridded data through spatial interpolation; extracting topographic parameters and topographic humidity index based on a DEM model, and determining landslide susceptibility evaluation factors by combining stratigraphic lithology, soil type, and human activity data; allocating weights using the analytic hierarchy process (AHP) to divide landslide-prone areas and locate prediction regions; dynamically fitting landslide critical curves for different stratigraphic lithologies based on historical disaster data and soil permeability coefficients, and outputting the probability of landslide occurrence using a random forest algorithm; multiplying the susceptibility by the probability results to generate gridded hazard values ​​and classifying early warning levels. This application improves the accuracy and timeliness of landslide early warning analysis.

Description

Technical Field

[0001] This invention relates to the technical field of geological disaster early warning, and in particular to a critical early warning method and system for rainfall-induced landslides. Background Technology

[0002] With the impact of climate change and human activities, natural disasters are becoming more frequent. Among them, landslides have become an indispensable type of natural disaster. Rainfall is one of the main factors that induce landslides. Rainfall-induced landslides are currently a key focus and hot topic in landslide disaster research both domestically and internationally. According to statistics, more than 80% of geological disasters in China are directly or indirectly affected by rainfall. Therefore, studying the impact mechanism and early warning of regional rainfall-induced landslides is of great significance for regional disaster prevention and mitigation.

[0003] Currently, early warning of rainfall-induced landslide disasters mainly relies on traditional monitoring methods, such as manual patrols and remote sensing monitoring. However, these methods have certain limitations. First, traditional manual patrols are inefficient and difficult to achieve full coverage. Second, while remote sensing monitoring has the advantage of wide coverage, its accuracy in identifying landslides in complex terrain or vegetated areas is significantly insufficient due to limitations in sensor precision. In addition, some automated early warning systems exist, but they generally suffer from weak multi-source data integration capabilities, insufficient scientific rigor in threshold models, and poor timeliness and adaptability of early warnings. In summary, existing technologies are insufficient to meet the needs of accurate early warning of landslide disasters in complex environments, and therefore, there is room for improvement. Summary of the Invention

[0004] To improve the accuracy and timeliness of early warning analysis for rainfall-induced landslides, this application provides a critical early warning method and system for rainfall-induced landslides.

[0005] Firstly, the above-mentioned inventive objective of this application is achieved through the following technical solution:

[0006] A method for early warning of critical landslides caused by rainfall, the method comprising the following steps:

[0007] Acquire multi-source geographic data for the target study area, including rainfall data, topographic data, stratigraphic lithology data, soil type data, vegetation cover data, and human engineering activity data; perform gridding processing on the multi-source geographic data to generate gridded data.

[0008] Based on the gridded data, landslide susceptibility evaluation factors are determined. Based on the landslide susceptibility evaluation factors and their weights, landslide susceptibility results for the target study area are generated. Based on the landslide susceptibility results, landslide prediction areas are determined.

[0009] Historical landslide disaster data and real-time rainfall data of the landslide prediction area are obtained, effective rainfall is calculated based on the permeability coefficient of the soil and rock mass, and landslide critical curves under different geological lithologies are dynamically fitted according to the effective rainfall.

[0010] Based on the landslide critical curve, critical rainfall parameters are obtained, and the probability of rainfall-induced landslides in the landslide prediction area is determined according to the critical rainfall parameters.

[0011] Based on the landslide susceptibility results and the probability of rainfall-induced landslides, a gridded hazard value is generated, and a landslide early warning signal is generated based on the gridded hazard value.

[0012] By adopting the above technical solution, in the process of landslide early warning analysis, multi-source geographic data of the target study area is collected. This multi-source geographic data includes rainfall data, topographic data, stratigraphic lithology data, soil type data, land cover data, and human engineering activity data. Rainfall data can be obtained in real time through meteorological stations, while the other data can be obtained in real time through remote sensing monitoring. The collected multi-source geographic data is then gridded to generate gridded data. Landslide susceptibility evaluation factors are generated using this gridded data to facilitate the evaluation and analysis of landslide susceptibility in the target study area. The weight coefficients of these landslide susceptibility factors are determined using the analytic hierarchy process (AHP). Based on these factors and their weight coefficients, landslide susceptibility is evaluated, resulting in landslide susceptibility assessments. This allows for regional classification of landslide susceptibility in the target study area, identifying areas prone to landslide disasters and thus determining landslide prediction areas. Historical landslide disaster data is collected for these identified landslide prediction areas. Based on this historical data and real-time rainfall data, combined with the regional... Effective rainfall is calculated using the permeability coefficient of soil and rock. This effective rainfall is then used to dynamically fit landslide critical curves under different geological lithologies, adapting to different strata and improving the accuracy and reliability of rainfall-induced landslide early warning. Analysis of the landslide critical curves yields critical rainfall parameters, which are the minimum combined threshold of rainfall intensity and duration required to trigger a landslide disaster under different geological lithologies. Based on these critical rainfall parameters, the probability of rainfall-induced landslides in the region is determined, enabling dynamic matching of landslide occurrence probabilities and improving the response speed of rainfall-induced landslide early warnings. Based on landslide susceptibility and rainfall-induced landslide occurrence probabilities, a gridded hazard value for the target study area is generated. This gridded hazard value represents the landslide risk value corresponding to each grid cell after the target study area is gridded. Corresponding landslide early warning signals are generated based on these gridded hazard values, thus realizing the rainfall-induced landslide early warning analysis process for the target study area. The entire early warning analysis process integrates multi-source geographic data and performs dynamic threshold matching, improving the accuracy and timeliness of rainfall-induced landslide early warning analysis.

[0013] In a preferred embodiment, this application can be further configured such that: the landslide susceptibility assessment factors include rainfall factors, topographic factors, hydrological environmental factors, geological factors, land cover factors, soil factors, and human activity factors; and the determination of landslide susceptibility assessment factors based on the gridded data specifically includes:

[0014] Based on the rainfall data, a continuous gridded rainfall layer is converted using spatial interpolation to form rainfall factors;

[0015] The topographic data is input into a preset DEM model to extract topographic parameters and topographic humidity index of the target study area. Topographic factors are generated based on the topographic parameters, and hydrological environmental factors are determined based on the topographic humidity index.

[0016] By integrating stratigraphic lithology data and soil type data, stratigraphic lithology classification, soil texture and distance from fault zones are analyzed to form geological factors and soil factors;

[0017] Based on vegetation cover data and human engineering activity data, road, building and vegetation distribution maps are analyzed. The land cover factor is determined according to the vegetation distribution map. The distance of each grid cell in the target study area to the nearest road is calculated based on the road and building distribution map to form the human activity factor.

[0018] By adopting the above technical solution, real-time rainfall data, including real-time rainfall, historical rainfall, and rainfall forecast data, is acquired and converted into a gridded rainfall layer through Kriging spatial interpolation to form rainfall factors. This solves the local error problem caused by data discretization and provides high-resolution and continuous rainfall data for landslide susceptibility analysis. The DEM model extracts topographic parameters (such as elevation, slope, curvature, etc.) and topographic humidity index of the target study area from the topographic data, accurately quantifying surface runoff and seepage characteristics. For example, areas with a slope greater than 25° combined with a topographic humidity index greater than 8.5 can identify high-risk slopes with strong seepage, providing key hydrological and topographic factors for landslide susceptibility analysis. Furthermore, stratigraphic lithology data and soil... Type data is used to analyze stratigraphic lithology and soil type data to identify stratigraphic lithology classification, soil texture, and distance from fault zones, constructing geological and soil factors. For example, in shale areas, due to low shear strength and a distance of less than 500m from fault zones, the probability of landslide susceptibility is significantly increased, providing geological and soil factors for landslide susceptibility analysis. By analyzing vegetation cover and human engineering activity data, distribution maps of roads, buildings, and vegetation within the target study area are obtained. Based on the vegetation distribution map, the surface cover factor within the area is determined. Using the road and building distribution maps, the distance of each grid cell within the target study area to the nearest road is calculated as a human activity factor, providing multi-source evaluation factors for landslide susceptibility analysis, thereby improving the accuracy of landslide susceptibility analysis.

[0019] In a preferred embodiment, this application can be further configured as follows: generating landslide susceptibility results for the target study area based on the landslide susceptibility evaluation factors and their weights, and determining the landslide prediction area based on the landslide susceptibility results, specifically includes:

[0020] The weight coefficients of each landslide susceptibility evaluation factor were assigned using the analytic hierarchy process (AHP). The landslide susceptibility evaluation factors were then normalized and weighted and superimposed to calculate the landslide susceptibility results for each grid unit in the target study area.

[0021] Based on the landslide susceptibility results of each grid cell, different levels of landslide-prone areas are divided to determine the landslide prediction area.

[0022] By adopting the above technical solution, the weight coefficients of landslide susceptibility evaluation factors are allocated using the analytic hierarchy process (AHP), avoiding the subjective bias of traditional methods. This makes the weight allocation of landslide susceptibility evaluation factors more closely aligned with actual geological conditions. The various landslide susceptibility evaluation factors are normalized and weighted, integrating multi-source factors into a single landslide susceptibility index, thus solving the problem of inconsistent dimensions among multiple factors. Based on the normalized and weighted landslide susceptibility factors, the landslide susceptibility results for each grid unit in the target study area are calculated. According to the landslide susceptibility results of each grid unit, the target study area is divided into different levels of landslide-prone zones. Using these different levels of landslide-prone zones, landslide prediction areas are determined within the target study area. These landslide prediction areas refer to the areas where rainfall-induced landslides are further analyzed for early warning, achieving spatialized positioning of landslide prediction areas. This helps to focus monitoring resources on high-risk areas and improve the efficiency of landslide early warning.

[0023] In a preferred embodiment, this application can be further configured as follows: acquiring historical landslide disaster data and real-time rainfall data of the landslide prediction area, calculating effective rainfall based on the permeability coefficient of the soil and rock mass, and dynamically fitting landslide critical curves under different geological lithologies based on the effective rainfall, specifically including:

[0024] An initial critical curve is fitted based on the historical landslide disaster data, and the initial critical rainfall parameters are obtained based on the initial critical curve.

[0025] Based on the lithological classification of the strata within the landslide prediction area, the permeability coefficient of the soil and rock mass is determined, and the saturation state of the soil and rock mass is determined based on the permeability coefficient of the soil and rock mass and real-time rainfall data.

[0026] The initial critical rainfall parameters are adjusted in real time according to the saturation state of the soil and rock mass. Based on the adjusted critical rainfall parameters, landslide critical rainfall thresholds under different geological lithologies are generated. Landslide critical curves are generated based on the landslide critical rainfall thresholds.

[0027] By adopting the above technical solution, an initial critical curve is fitted using historical landslide disaster data to establish a benchmark relationship between rainfall intensity and duration. The initial critical rainfall parameter is determined using the initial critical curve. The permeability coefficient of the soil and rock mass within the landslide prediction area is determined through lithological analysis. Differentiated permeability parameters of soil and rock mass are determined by classifying the lithology of the soil and rock mass within the landslide prediction area, such as a permeability parameter of 0.8 for clay and 0.9 for sand. The saturation state of the soil and rock mass is determined based on the permeability coefficient and real-time rainfall data in the area, quantifying the cumulative permeability effect of previous rainfall in the area, thus improving the accuracy of rainfall-induced landslide prediction and analysis. The initial critical rainfall parameter is dynamically adjusted using the soil and rock mass saturation state, and a landslide critical rainfall threshold is generated. A corresponding landslide critical curve is generated based on the landslide critical rainfall threshold to adapt to different lithological formations, thereby improving the accuracy and reliability of rainfall-induced landslide early warning.

[0028] In a preferred embodiment, this application can be further configured as follows: obtaining critical rainfall parameters based on the landslide critical curve, and determining the probability of rainfall-induced landslides in the landslide prediction area based on the critical rainfall parameters, specifically includes:

[0029] Based on the landslide critical curve, critical rainfall parameters are statistically analyzed to construct a rainfall-type landslide dataset.

[0030] Based on the rainfall-induced landslide dataset, the random forest algorithm is used to combine the topographic parameters and topographic humidity index of the landslide prediction area to output the probability of rainfall-induced landslides occurring in the landslide prediction area.

[0031] By adopting the above technical solution, all critical rainfall parameters are statistically analyzed based on the landslide critical curve and integrated to form a rainfall-type landslide dataset. Using the random forest algorithm, combined with terrain parameters and terrain humidity index, a probabilistic landslide early warning model is constructed, outputting the probability results of rainfall-type landslide occurrence, realizing the function of landslide early warning prediction and analysis. Through dynamic adaptation of geological parameters and multi-factor probabilistic modeling, the accuracy and reliability of rainfall-type landslide early warning are significantly improved.

[0032] In a preferred embodiment, this application can be further configured as follows: generating a gridded hazard value based on the landslide susceptibility results and the probability results of rainfall-induced landslides, and generating a landslide early warning signal based on the gridded hazard value, specifically including:

[0033] The landslide susceptibility result of each grid cell in the landslide prediction area is multiplied by the probability result of rainfall-induced landslide occurrence to obtain the gridded hazard value;

[0034] Landslide warning signals are divided into different levels based on gridded hazard values. Based on landslide warning signals of the same level, adjacent grid units are merged to form a landslide warning area.

[0035] By adopting the above technical solution, the landslide susceptibility result of each grid unit is multiplied by the probability of rainfall-induced landslides to obtain a gridded hazard value. This achieves two-way coupling of static geological conditions and dynamic rainfall processes, accurately quantifies comprehensive risks, avoids misjudgments caused by single indicators, reduces the false alarm rate of landslide warnings, divides landslide warning signals into different levels based on gridded hazard values, and merges adjacent grid units with the same level of warning to form a continuous landslide warning area. Through hazard fusion and spatialized warning, the accuracy of landslide warnings and the efficiency of emergency response are significantly improved.

[0036] In a preferred embodiment, this application can be further configured as follows: after generating a gridded hazard value based on the landslide susceptibility result and the probability result of rainfall-induced landslide occurrence, and generating a landslide early warning signal based on the gridded hazard value, the rainfall-induced landslide critical early warning method further includes:

[0037] Real-time data collection of landslide signs reported by the public; data preprocessing of the landslide signs data to obtain valid data;

[0038] Based on the Bayesian network model, the effective data and gridded hazard values ​​are fused to obtain a corrected landslide early warning signal. An emergency response signal is triggered based on the corrected landslide early warning signal, and a targeted early warning notification is sent to the target area.

[0039] By adopting the above technical solution, a mobile application is used to receive landslide warning data reported by the public. The landslide warning data includes descriptions of landslide precursor phenomena, geographical locations, timestamps, and multimedia evidence. The landslide warning data is preprocessed to obtain valid data, including verifying whether the reported location is within the landslide prediction area through GPS positioning. Image recognition algorithms are used to analyze the uploaded multimedia evidence to identify ground cracks, vegetation tilting, or surface displacement characteristics. Real-time sensor data is compared, and if the soil moisture or displacement monitoring values ​​in the area are consistent with the reported phenomena, they are marked as valid data. The valid data and gridded hazard values ​​are fused using a Bayesian network model to obtain a corrected landslide warning signal. An emergency response signal is triggered based on the corrected landslide warning signal, and a targeted warning notification is sent to the target area. Combining public data to fill monitoring blind spots significantly improves the accuracy and real-time performance of landslide warnings in complex terrain.

[0040] Secondly, the above-mentioned inventive objective of this application is achieved through the following technical solutions:

[0041] A rainfall-based landslide critical early warning system, the rainfall-based landslide critical early warning system comprising:

[0042] The multi-source data processing module is used to acquire multi-source geographic data of the target study area. The multi-source geographic data includes rainfall data, topographic data, stratigraphic lithology data, soil type data, vegetation cover data, and human engineering activity data. The multi-source geographic data is then processed into grids to generate gridded data.

[0043] The landslide susceptibility analysis module is used to determine landslide susceptibility evaluation factors based on the gridded data, generate landslide susceptibility results for the target study area based on the landslide susceptibility evaluation factors and their weights, and determine the landslide prediction area based on the landslide susceptibility results.

[0044] The critical curve generation module is used to acquire historical landslide disaster data and real-time rainfall data of the landslide prediction area, calculate the effective rainfall based on the permeability coefficient of the soil and rock mass, and dynamically fit the landslide critical curve under different geological lithologies based on the effective rainfall.

[0045] The landslide probability prediction module is used to obtain critical rainfall parameters based on the landslide critical curve, and to determine the probability of rainfall-induced landslides in the landslide prediction area based on the critical rainfall parameters.

[0046] The landslide early warning signal generation module is used to generate a gridded hazard value based on the landslide susceptibility results and the probability results of rainfall-induced landslides, and to generate a landslide early warning signal based on the gridded hazard value.

[0047] By adopting the above technical solution, in the process of landslide early warning analysis, multi-source geographic data of the target study area is collected. This multi-source geographic data includes rainfall data, topographic data, stratigraphic lithology data, soil type data, land cover data, and human engineering activity data. Rainfall data can be obtained in real time through meteorological stations, while the other data can be obtained in real time through remote sensing monitoring. The collected multi-source geographic data is then gridded to generate gridded data. Landslide susceptibility evaluation factors are generated using this gridded data to facilitate the evaluation and analysis of landslide susceptibility in the target study area. The weight coefficients of these landslide susceptibility factors are determined using the analytic hierarchy process (AHP). Based on these factors and their weight coefficients, landslide susceptibility is evaluated, resulting in landslide susceptibility assessments. This allows for regional classification of landslide susceptibility in the target study area, identifying areas prone to landslide disasters and thus determining landslide prediction areas. Historical landslide disaster data is collected for these identified landslide prediction areas. Based on this historical data and real-time rainfall data, combined with the regional... Effective rainfall is calculated using the permeability coefficient of soil and rock. This effective rainfall is then used to dynamically fit landslide critical curves under different geological lithologies, adapting to different strata and improving the accuracy and reliability of rainfall-induced landslide early warning. Analysis of the landslide critical curves yields critical rainfall parameters, which are the minimum combined threshold of rainfall intensity and duration required to trigger a landslide disaster under different geological lithologies. Based on these critical rainfall parameters, the probability of rainfall-induced landslides in the region is determined, enabling dynamic matching of landslide occurrence probabilities and improving the response speed of rainfall-induced landslide early warnings. Based on landslide susceptibility and rainfall-induced landslide occurrence probabilities, a gridded hazard value for the target study area is generated. This gridded hazard value represents the landslide risk value corresponding to each grid cell after the target study area is gridded. Corresponding landslide early warning signals are generated based on these gridded hazard values, thus realizing the rainfall-induced landslide early warning analysis process for the target study area. The entire early warning analysis process integrates multi-source geographic data and performs dynamic threshold matching, improving the accuracy and timeliness of rainfall-induced landslide early warning analysis.

[0048] Thirdly, the above-mentioned objectives of this application are achieved through the following technical solutions:

[0049] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described rainfall-induced landslide critical early warning method.

[0050] Fourthly, the above-mentioned objectives of this application are achieved through the following technical solutions:

[0051] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described rainfall-induced landslide critical early warning method.

[0052] In summary, this application includes at least one of the following beneficial technical effects:

[0053] 1. In the process of landslide early warning analysis, multi-source geographic data of the target study area is collected. This multi-source geographic data includes rainfall data, topographic data, stratigraphic lithology data, soil type data, land cover data, and human engineering activity data. The collected multi-source geographic data is processed into gridded data. Landslide susceptibility evaluation factors are generated using the resulting gridded data to facilitate landslide susceptibility assessment of the target study area. The weight coefficients of these landslide susceptibility factors are determined using the analytic hierarchy process (AHP). Based on these landslide susceptibility factors and their weight coefficients, a landslide susceptibility assessment is conducted to obtain the landslide susceptibility results, enabling the assessment of landslide risks in the target study area. Landslide susceptibility is classified into regional levels to identify areas in the target study area prone to landslide disasters, thus obtaining landslide prediction areas. Historical landslide disaster data is collected for the identified landslide prediction areas. Based on the historical landslide disaster data and the rainfall data of the area, critical rainfall parameters under different geological lithologies are calculated. The probability of rainfall-induced landslides in the area is determined based on the critical rainfall parameters, which can dynamically match the probability of landslides in the area and improve the speed of early warning response to rainfall-induced landslides. Based on the landslide susceptibility results and the probability of rainfall-induced landslides, a gridded hazard value is generated for the target study area, and a corresponding landslide early warning signal is generated based on the gridded hazard value.

[0054] 2. By acquiring real-time rainfall data, including real-time rainfall, historical rainfall, and rainfall forecast data, and generating a gridded rainfall layer through Kriging space interpolation to form rainfall factors, the problem of local errors caused by data discretization is solved. This provides high-resolution and continuous rainfall data for landslide susceptibility analysis. Using a DEM model, topographic parameters (such as elevation, slope, curvature, etc.) and topographic humidity index of the target study area are extracted from the topographic data to accurately quantify surface runoff and seepage characteristics. For example, areas with a slope greater than 25° combined with a topographic humidity index greater than 8.5 can identify high-risk slopes with strong seepage, providing key hydrological and topographic factors for landslide susceptibility analysis. Furthermore, stratigraphic lithology data and soil type data are integrated. By analyzing stratigraphic lithology data and soil type data, stratigraphic lithology classification, soil texture, and distance from fault zones are obtained, and geological and soil factors are constructed. For example, in shale areas, due to low shear strength and a distance of less than 500m from fault zones, the probability of landslide susceptibility is greatly increased, providing geological and soil factors for landslide susceptibility analysis. By analyzing vegetation cover data and human engineering activity data, road, building, and vegetation distribution maps are obtained within the target study area. Based on the vegetation distribution map, the surface cover factor within the area is determined. Using the road and building distribution maps, the distance of each grid cell within the target study area to the nearest road is calculated as a human activity factor, providing multi-source evaluation factors for landslide susceptibility analysis, thereby improving the accuracy of landslide susceptibility analysis.

[0055] 3. By fitting an initial critical curve to historical landslide disaster data, a baseline relationship between rainfall intensity and duration is established. The initial critical rainfall parameter is determined using the initial critical curve. Through lithological analysis of the strata within the landslide prediction area, the permeability coefficient of the soil and rock mass is determined. Differentiated permeability parameters of soil and rock mass are determined by classifying the lithology of the strata within the landslide prediction area, such as 0.8 for clay and 0.9 for sand. Based on the permeability coefficient of the soil and rock mass and real-time rainfall data in the area, the saturation state of the soil and rock mass is determined, quantifying the cumulative permeability effect of previous rainfall in the area, thus improving the accuracy of rainfall-induced landslide prediction and analysis. The initial critical rainfall parameter is dynamically adjusted using the soil and rock mass saturation state, and a landslide critical rainfall threshold is generated. Based on the landslide critical rainfall threshold, a corresponding landslide critical curve is generated to adapt to different lithologies, thereby improving the accuracy and reliability of rainfall-induced landslide early warning.

[0056] 4. Utilize a mobile application to receive landslide warning data reported by the public. This data includes descriptions of precursory landslide phenomena, geographical locations, timestamps, and multimedia evidence. Preprocessing of the landslide warning data yields valid data, including verifying via GPS location whether the reported location is within the landslide prediction area. Image recognition algorithms are used to analyze the uploaded multimedia evidence, identifying ground cracks, vegetation tilting, or surface displacement characteristics. Real-time sensor data is compared; if soil moisture or displacement monitoring values ​​within the area match the reported phenomena, the data is marked as valid. The valid data and gridded hazard values ​​are fused using a Bayesian network model to obtain a corrected landslide warning signal. Based on the corrected landslide warning signal, an emergency response signal is triggered, and targeted warning notifications are sent to the target area. Combining public data to fill monitoring blind spots significantly improves the accuracy and real-time performance of landslide warnings in complex terrain. Attached Figure Description

[0057] Figure 1 This is a flowchart of a critical early warning method for rainfall-induced landslides according to one embodiment of this application;

[0058] Figure 2 This is a flowchart illustrating the implementation of step S20 in a rainfall-induced landslide critical early warning method according to an embodiment of this application.

[0059] Figure 3 This is another implementation flowchart of step S20 in a critical early warning method for rainfall-induced landslides according to an embodiment of this application;

[0060] Figure 4 This is a flowchart illustrating the implementation of step S30 in a rainfall-induced landslide critical early warning method according to an embodiment of this application.

[0061] Figure 5 This is a flowchart illustrating the implementation of step S40 in a rainfall-induced landslide critical early warning method according to an embodiment of this application.

[0062] Figure 6 This is a flowchart illustrating the implementation of step S50 in a rainfall-induced landslide critical early warning method according to an embodiment of this application.

[0063] Figure 7 This is another implementation flowchart of a critical early warning method for rainfall-induced landslides in one embodiment of this application;

[0064] Figure 8 This is a principle block diagram of a rainfall-induced landslide critical early warning system according to one embodiment of this application;

[0065] Figure 9 This is a schematic diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0066] The present application will be further described in detail below with reference to the accompanying drawings.

[0067] In one embodiment, such as Figure 1 As shown, this application discloses a critical early warning method for rainfall-induced landslides, which specifically includes the following steps:

[0068] S10: Obtain multi-source geographic data of the target study area, including rainfall data, topographic data, stratigraphic lithology data, soil type data, vegetation cover data, and human engineering activity data; perform gridding processing on the multi-source geographic data to generate gridded data.

[0069] Specifically, in the process of landslide early warning analysis, rainfall data is acquired in real time through meteorological stations, and topographic data, stratigraphic lithology data, soil type data, land cover data, and human engineering activity data are acquired in real time through remote sensing monitoring. The collected multi-source geographic data is then processed into gridded data to generate gridded data. The resulting gridded data is used to generate landslide susceptibility evaluation factors to facilitate the evaluation and analysis of landslide susceptibility in the target study area.

[0070] Furthermore, by integrating multi-source geographic data such as rainfall, topography, geology, soil, vegetation, and human activities (e.g., topographic humidity index, stratigraphic lithology, distance from fault zones, etc.), a multi-factor evaluation system is constructed, which solves the problem of single data in traditional early warning models and improves the accuracy of landslide early warning.

[0071] S20: Determine landslide susceptibility evaluation factors based on the gridded data, generate landslide susceptibility results for the target study area based on the landslide susceptibility evaluation factors and their weights, and determine the landslide prediction area based on the landslide susceptibility results.

[0072] Specifically, the weight coefficients of the obtained landslide susceptibility factors are determined by the analytic hierarchy process (AHP), such as slope weight accounting for 25%, stratigraphic lithology weight accounting for 20%, and rainfall weight accounting for 18%. Based on the landslide susceptibility factors and their weight coefficients, landslide susceptibility is evaluated, and landslide susceptibility results are obtained. This allows for the classification of landslide susceptibility in the target study area, thereby identifying areas in the target study area prone to landslide disasters and obtaining landslide prediction areas.

[0073] S30: Obtain historical landslide disaster data and real-time rainfall data for the landslide prediction area, calculate the effective rainfall based on the permeability coefficient of the soil and rock, and dynamically fit the critical curve of landslide under different geological lithologies based on the effective rainfall.

[0074] Specifically, based on historical landslide disaster data and real-time rainfall data in the region, and combined with the permeability coefficient of the soil and rock mass in the region, the effective rainfall is calculated. The obtained effective rainfall is then used to dynamically fit the critical curve of landslides under different geological lithologies, so as to adapt to different strata lithologies and thus improve the accuracy and reliability of rainfall-induced landslide early warning.

[0075] S40: Obtain critical rainfall parameters based on the landslide critical curve, and determine the probability of rainfall-induced landslides in the landslide prediction area based on the critical rainfall parameters.

[0076] Specifically, the critical rainfall parameter refers to the minimum combined threshold of rainfall intensity and duration required to trigger a landslide disaster under different geological lithologies. By determining the probability of rainfall-induced landslides in a region based on the critical rainfall parameter, the probability of landslides in that region can be dynamically matched, thereby improving the speed of early warning and response to rainfall-induced landslides.

[0077] S50: Based on the landslide susceptibility results and the probability results of rainfall-induced landslides, generate gridded hazard values, and generate landslide early warning signals based on the gridded hazard values.

[0078] Specifically, based on the landslide susceptibility results and the probability of rainfall-induced landslides, a gridded hazard value is generated for the target study area. The gridded hazard value refers to the landslide hazard value corresponding to each grid after the target study area is gridded. A corresponding landslide early warning signal is generated based on the gridded hazard value.

[0079] In this embodiment, during the landslide early warning analysis, multi-source geographic data of the target study area is collected. This multi-source geographic data includes rainfall data, topographic data, stratigraphic lithology data, soil type data, land cover data, and human engineering activity data. Rainfall data can be acquired in real-time through weather stations, while the other data can be acquired in real-time through remote sensing monitoring. The collected multi-source geographic data is then gridded to generate gridded data. Landslide susceptibility evaluation factors are generated using this gridded data to facilitate the landslide susceptibility evaluation and analysis of the target study area. The weight coefficients of these landslide susceptibility factors are determined using the analytic hierarchy process (AHP). Based on these landslide susceptibility factors and their weight coefficients, a landslide susceptibility evaluation is conducted, resulting in landslide susceptibility assessments. This allows for regional classification of landslide susceptibility in the target study area, identifying areas prone to landslide disasters and thus determining landslide prediction areas. Historical landslide disaster data is collected for the identified landslide prediction areas. Based on this historical landslide disaster data and real-time rainfall data of the area, combined with the local soil and rock conditions, the analysis is conducted. Effective rainfall is calculated using volume permeability coefficients. This effective rainfall data is then dynamically fitted to landslide critical curves under different geological lithologies to adapt to varying strata, thereby improving the accuracy and reliability of rainfall-induced landslide early warnings. Analysis of the landslide critical curves yields critical rainfall parameters, which are the minimum combined threshold of rainfall intensity and duration required to trigger a landslide disaster under different geological lithologies. Based on these critical rainfall parameters, the probability of rainfall-induced landslides in the region is determined, enabling dynamic matching of landslide occurrence probabilities and improving the response speed of rainfall-induced landslide early warnings. Based on landslide susceptibility and rainfall-induced landslide occurrence probabilities, a gridded hazard value for the target study area is generated. This gridded hazard value represents the landslide risk value corresponding to each grid cell after the target study area is gridded. Corresponding landslide early warning signals are generated based on these gridded hazard values, thus realizing the rainfall-induced landslide early warning analysis process for the target study area. The entire early warning analysis process integrates multi-source geographic data and performs dynamic threshold matching to improve the accuracy and timeliness of rainfall-induced landslide early warning analysis.

[0080] In one embodiment, such as Figure 2 As shown, in step S20, the landslide susceptibility assessment factors include rainfall factors, topographic factors, hydrological environmental factors, geological factors, land cover factors, soil factors, and human activity factors. The landslide susceptibility assessment factors are determined based on gridded data, specifically including:

[0081] S11: Based on rainfall data, a continuous gridded rainfall layer is converted using spatial interpolation to form rainfall factors.

[0082] Specifically, rainfall data is acquired in real time through meteorological stations, including real-time rainfall, historical rainfall, and rainfall forecast data. The discrete station data is converted into a spatially continuous 1km×1km gridded rainfall layer using the Kriging spatial interpolation method to form a rainfall factor. This solves the problem of local errors caused by data discretization and provides high-resolution and continuous rainfall data for landslide susceptibility analysis.

[0083] S12: Input the topographic data into the preset DEM model, extract the topographic parameters and topographic humidity index of the target study area, generate topographic factors based on the topographic parameters, and determine the hydrological environmental factors based on the topographic humidity index.

[0084] Specifically, by using the DEM model to extract topographic parameters (such as elevation, slope, curvature, etc.) and topographic humidity index of the target study area from topographic data, the surface runoff and seepage characteristics can be accurately quantified. For example, when the slope is greater than 25° and the topographic humidity index is greater than 8.5, high-risk slopes with strong seepage can be identified, providing key hydrological and topographic factors for landslide susceptibility analysis.

[0085] S13: Integrate stratigraphic lithology data and soil type data to analyze stratigraphic lithology classification, soil texture and distance from fault zones, and form geological factors and soil factors.

[0086] Specifically, by integrating stratigraphic lithology data and soil type data, stratigraphic lithology classification, soil texture, and distance from fault zones are analyzed from the stratigraphic lithology data and soil type data to construct geological factors and soil factors. For example, in shale areas, due to low shear strength and a distance of less than 500m from the fault zone, the probability of landslide susceptibility is greatly increased, providing geological factors and soil factors for landslide susceptibility analysis.

[0087] S14: Based on vegetation cover data and human engineering activity data, analyze the road, building and vegetation distribution maps, determine the land cover factor according to the vegetation distribution map, and calculate the distance of each grid cell in the target study area from the nearest road based on the road and building distribution maps to form the human activity factor.

[0088] Specifically, vegetation cover data and human engineering activity data are analyzed to obtain road, building and vegetation distribution maps within the target study area. The land cover factor within the area is determined based on the vegetation distribution map. The distance from each grid cell to the nearest road within the target study area is calculated using the road distribution map and building distribution map, which serves as the human activity factor.

[0089] In one embodiment, such as Figure 3As shown, in step S20, the landslide susceptibility results for the target study area are generated based on the landslide susceptibility evaluation factors and their weights, and the landslide prediction area is determined based on the landslide susceptibility results. Specifically, this includes:

[0090] S21: The weight coefficients of each landslide susceptibility evaluation factor are allocated using the analytic hierarchy process (AHP). The landslide susceptibility evaluation factors are then normalized and weighted and superimposed to calculate the landslide susceptibility results for each grid unit in the target study area.

[0091] Specifically, the weight coefficients of landslide susceptibility evaluation factors are allocated using the analytic hierarchy process (AHP) to avoid the subjective bias of traditional methods, making the weight allocation of landslide susceptibility evaluation factors more closely reflect actual geological conditions. The various landslide susceptibility evaluation factors are then normalized and weighted, integrating multi-source factors into a single landslide susceptibility index. Based on the normalized and weighted landslide susceptibility factors, the landslide susceptibility results for each grid unit in the target study area are calculated.

[0092]

[0093] Where a represents the weighting coefficient, f represents the evaluation factor for landslide hazard susceptibility, and n represents the number of evaluation factors.

[0094] S22: Based on the landslide susceptibility results of each grid unit, landslide-prone areas of different levels are divided to determine the landslide prediction area.

[0095] Specifically, based on the landslide susceptibility results of each grid unit, the target study area is divided into different levels of landslide-prone areas, such as high-susceptibility areas (F>0.7), medium-susceptibility areas (F=0.4-0.7), and low-susceptibility areas (F<0.4). By using the different levels of landslide-prone areas, landslide prediction areas are determined within the target study area. The landslide prediction area refers to the area where rainfall-induced landslides are further analyzed for early warning, thus realizing the spatial positioning of landslide prediction areas. This helps to focus monitoring resources on high-risk areas and improve the efficiency of landslide early warning.

[0096] In one embodiment, such as Figure 4 As shown, in step S30, historical landslide disaster data and real-time rainfall data of the landslide prediction area are obtained, and the effective rainfall is calculated by combining the permeability coefficient of the soil and rock mass. Based on the effective rainfall, the critical curve of landslide under different geological lithologies is dynamically fitted, specifically including:

[0097] S31: Fit an initial critical curve based on the historical landslide disaster data, and obtain the initial critical rainfall parameters based on the initial critical curve.

[0098] Specifically, using historical landslide disaster data, an ID threshold model is employed for critical threshold analysis. Taking the day of the landslide disaster as the starting point, the duration of continuous rainfall preceding the landslide is used as the sustained rainfall time. Combined with the effective rainfall intensity at the time of the landslide disaster, an initial landslide critical curve is fitted to establish a baseline relationship between rainfall intensity and rainfall duration.

[0099] I = c + α·D β ;

[0100] Where I is the effective rainfall intensity, D is the rainfall duration, α and β are statistical parameters, and c≥0 are parameters obtained from historical landslide disaster data.

[0101] S32: Based on the lithological classification of the strata within the landslide prediction area, determine the permeability coefficient of the soil and rock mass, and determine the saturation state of the soil and rock mass based on the permeability coefficient of the soil and rock mass and real-time rainfall data.

[0102] Specifically, by classifying the lithology of the strata within the landslide prediction area, differentiated soil and rock permeability parameters are determined. For example, the permeability parameter for clay is 0.8, and for sand it is 0.9. These permeability parameters, along with real-time rainfall data for the area, are used to quantify the saturation state of the soil and rock mass and the cumulative infiltration effect of previous rainfall, thus improving the accuracy of rainfall-induced landslide prediction and analysis.

[0103]

[0104] Among them, S t Let R be the saturation degree of the soil and rock mass at time t, k be the permeability coefficient, and R be the saturation degree of the soil and rock mass at time t. i Let λ be the rainfall on day i, and λ be the attenuation coefficient.

[0105] S33: Adjust the initial critical rainfall parameters in real time according to the saturation state of the soil and rock mass, generate landslide critical rainfall thresholds under different geological lithologies based on the adjusted critical rainfall parameters, and generate landslide critical curves based on the landslide critical rainfall thresholds.

[0106] Specifically, the initial critical rainfall parameters are dynamically adjusted based on the saturation state of the soil and rock mass, and a critical rainfall threshold for landslides is generated. Based on the critical rainfall threshold, a corresponding critical curve for landslides is generated to adapt to different lithologies, thereby improving the accuracy and reliability of rainfall-induced landslide early warning.

[0107] In one embodiment, such as Figure 5 As shown, in step S40, the critical rainfall parameters are obtained based on the landslide critical curve, and the probability of rainfall-induced landslides in the landslide prediction area is determined based on the critical rainfall parameters. Specifically, this includes:

[0108] S41: Based on the landslide critical curve, statistically analyze the critical rainfall parameters and construct a rainfall-type landslide dataset.

[0109] S42: Based on the rainfall-induced landslide dataset, the random forest algorithm is used to combine the topographic parameters and topographic humidity index of the landslide prediction area to output the probability of rainfall-induced landslides occurring in the landslide prediction area.

[0110] Specifically, based on the landslide critical curve, all critical rainfall parameters are statistically analyzed and integrated to form a rainfall-type landslide dataset. Using the random forest algorithm, combined with topographic parameters and topographic humidity index, a probabilistic landslide early warning model is constructed, outputting the probability of rainfall-type landslide occurrence, realizing the function of landslide early warning prediction and analysis. Through dynamic adaptation of geological parameters and multi-factor probabilistic modeling, the accuracy and reliability of rainfall-type landslide early warning are significantly improved.

[0111] In one embodiment, such as Figure 6 As shown, in step S50, based on the landslide susceptibility results and the probability results of rainfall-induced landslides, a gridded hazard value is generated, and a landslide early warning signal is generated based on the gridded hazard value, specifically including:

[0112] S51: Multiply the landslide susceptibility result of each grid cell in the landslide prediction area with the probability of occurrence of rainfall-induced landslides to obtain the gridded hazard value.

[0113] Specifically, the landslide susceptibility result of each grid cell is multiplied by the probability of rainfall-induced landslides to obtain the gridded hazard value: R = P × F, where R represents the gridded hazard value, P is the probability of rainfall-induced landslides, and F is the landslide susceptibility result. This achieves two-way coupling between static geological conditions and dynamic rainfall processes, accurately quantifies comprehensive risks, avoids misjudgments caused by single indicators, and reduces the false alarm rate of landslide warnings.

[0114] S52: Landslide warning signals are divided into different levels based on gridded hazard values. Based on landslide warning signals of the same level, adjacent grid units are merged to form a landslide warning area.

[0115] Specifically, landslide warning signals are divided into five different levels based on gridded hazard values: extremely low (0.0–0.2), low (0.2–0.4), medium (0.4–0.6), high (0.6–0.8), and extremely high (0.8–1.0). Adjacent grid units of the same warning level are merged to form continuous landslide warning areas. Through hazard fusion and spatialized warning, the accuracy of landslide warnings and the efficiency of emergency response are significantly improved.

[0116] In one embodiment, such as Figure 7 As shown, after step S50, that is, after generating gridded hazard values ​​based on landslide susceptibility results and rainfall-induced landslide occurrence probability results, and generating landslide early warning signals based on gridded hazard values, the rainfall-induced landslide critical early warning method further includes:

[0117] S60: Collect landslide sign data reported by the public in real time, and perform data preprocessing on the landslide sign data to obtain valid data.

[0118] Specifically, a mobile application is used to receive landslide warning data reported by the public. This data includes descriptions of precursory landslide phenomena, geographical locations, timestamps, and multimedia evidence. The data is preprocessed to obtain valid data, including verifying whether the reported location is within the landslide prediction area via GPS positioning. Image recognition algorithms are used to analyze the uploaded multimedia evidence to identify ground cracks, vegetation tilting, or surface displacement characteristics. Real-time sensor data is compared, and if the soil moisture or displacement monitoring values ​​in the area are consistent with the reported phenomena, the data is marked as valid. This, combined with public data, fills in monitoring blind spots and significantly improves the accuracy and real-time performance of landslide warnings in complex terrain.

[0119] S70: Based on the Bayesian network model, the effective data and gridded hazard values ​​are fused to obtain a corrected landslide early warning signal. An emergency response signal is triggered based on the corrected landslide early warning signal, and a targeted early warning notification is sent to the target area.

[0120] Specifically, effective data and gridded hazard values ​​are fused using a Bayesian network model to obtain a corrected landslide early warning signal. When the corrected warning level reaches "extremely high", an emergency response signal is triggered, and a targeted early warning notification is sent to the target area.

[0121] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0122] In one embodiment, a rainfall-based landslide critical early warning system is provided, which corresponds one-to-one with the rainfall-based landslide critical early warning method described in the above embodiments. For example... Figure 8 As shown, this rainfall-based landslide critical early warning system includes a multi-source data processing module, a landslide susceptibility analysis module, a critical curve generation module, a landslide probability prediction module, and a landslide early warning signal generation module. Detailed descriptions of each functional module are as follows:

[0123] The multi-source data processing module is used to acquire multi-source geographic data of the target study area. The multi-source geographic data includes rainfall data, topographic data, stratigraphic lithology data, soil type data, vegetation cover data, and human engineering activity data. The multi-source geographic data is then processed into grids to generate gridded data.

[0124] The landslide susceptibility analysis module is used to determine landslide susceptibility evaluation factors based on the gridded data, generate landslide susceptibility results for the target study area based on the landslide susceptibility evaluation factors and their weights, and determine the landslide prediction area based on the landslide susceptibility results.

[0125] The critical curve generation module is used to acquire historical landslide disaster data and real-time rainfall data of the landslide prediction area, calculate the effective rainfall based on the permeability coefficient of the soil and rock mass, and dynamically fit the landslide critical curve under different geological lithologies based on the effective rainfall.

[0126] The landslide probability prediction module is used to obtain critical rainfall parameters based on the landslide critical curve, and to determine the probability of rainfall-induced landslides in the landslide prediction area based on the critical rainfall parameters.

[0127] The landslide early warning signal generation module is used to generate a gridded hazard value based on the landslide susceptibility results and the probability results of rainfall-induced landslides, and to generate a landslide early warning signal based on the gridded hazard value.

[0128] Specific limitations regarding the rainfall-based landslide critical early warning system can be found in the limitations of the rainfall-based landslide critical early warning method described above, and will not be repeated here. Each module in the aforementioned rainfall-based landslide critical early warning system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in an electronic device, or stored in the memory of an electronic device in software form, so that the processor can call and execute the corresponding operations of each module.

[0129] In one embodiment, an electronic device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 9 As shown, this electronic device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores the database. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a rainfall-induced landslide critical early warning method.

[0130] In one embodiment, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps:

[0131] Acquire multi-source geographic data for the target study area, including rainfall data, topographic data, stratigraphic lithology data, soil type data, vegetation cover data, and human engineering activity data; perform gridding processing on the multi-source geographic data to generate gridded data.

[0132] Based on the gridded data, landslide susceptibility evaluation factors are determined. Based on the landslide susceptibility evaluation factors and their weights, landslide susceptibility results for the target study area are generated. Based on the landslide susceptibility results, landslide prediction areas are determined.

[0133] Historical landslide disaster data and real-time rainfall data of the landslide prediction area are obtained, effective rainfall is calculated based on the permeability coefficient of the soil and rock mass, and landslide critical curves under different geological lithologies are dynamically fitted according to the effective rainfall.

[0134] Based on the landslide critical curve, critical rainfall parameters are obtained, and the probability of rainfall-induced landslides in the landslide prediction area is determined according to the critical rainfall parameters.

[0135] Based on the landslide susceptibility results and the probability of rainfall-induced landslides, a gridded hazard value is generated, and a landslide early warning signal is generated based on the gridded hazard value.

[0136] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0137] Acquire multi-source geographic data for the target study area, including rainfall data, topographic data, stratigraphic lithology data, soil type data, vegetation cover data, and human engineering activity data; perform gridding processing on the multi-source geographic data to generate gridded data.

[0138] Based on the gridded data, landslide susceptibility evaluation factors are determined. Based on the landslide susceptibility evaluation factors and their weights, landslide susceptibility results for the target study area are generated. Based on the landslide susceptibility results, landslide prediction areas are determined.

[0139] Historical landslide disaster data and real-time rainfall data of the landslide prediction area are obtained, effective rainfall is calculated based on the permeability coefficient of the soil and rock mass, and landslide critical curves under different geological lithologies are dynamically fitted according to the effective rainfall.

[0140] Based on the landslide critical curve, critical rainfall parameters are obtained, and the probability of rainfall-induced landslides in the landslide prediction area is determined according to the critical rainfall parameters.

[0141] Based on the landslide susceptibility results and the probability of rainfall-induced landslides, a gridded hazard value is generated, and a landslide early warning signal is generated based on the gridded hazard value.

[0142] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0143] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0144] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for early warning of critical landslides caused by rainfall, characterized in that, The critical early warning method for rainfall-induced landslides includes the following steps: Acquire multi-source geographic data for the target study area, including rainfall data, topographic data, stratigraphic lithology data, soil type data, vegetation cover data, and human engineering activity data; perform gridding processing on the multi-source geographic data to generate gridded data. Based on the gridded data, landslide susceptibility evaluation factors are determined. Based on the landslide susceptibility evaluation factors and their weights, landslide susceptibility results for the target study area are generated. Based on the landslide susceptibility results, landslide prediction areas are determined. Historical landslide disaster data and real-time rainfall data of the landslide prediction area are obtained, effective rainfall is calculated based on the permeability coefficient of the soil and rock mass, and landslide critical curves under different geological lithologies are dynamically fitted according to the effective rainfall. Based on the landslide critical curve, critical rainfall parameters are obtained, and the probability of rainfall-induced landslides in the landslide prediction area is determined according to the critical rainfall parameters. Based on the landslide susceptibility results and the probability results of rainfall-induced landslides, a gridded hazard value is generated, and a landslide early warning signal is generated based on the gridded hazard value; The process of acquiring historical landslide disaster data and real-time rainfall data for the predicted landslide area, calculating effective rainfall based on the permeability coefficient of the soil and rock mass, and dynamically fitting landslide critical curves under different geological lithologies based on the effective rainfall specifically includes: An initial critical curve is fitted based on the historical landslide disaster data, and the initial critical rainfall parameters are obtained based on the initial critical curve. Based on the lithological classification of the strata within the landslide prediction area, the permeability coefficient of the soil and rock mass is determined, and the saturation state of the soil and rock mass is determined based on the permeability coefficient of the soil and rock mass and real-time rainfall data. The initial critical rainfall parameters are adjusted in real time according to the saturation state of the soil and rock mass. Based on the adjusted critical rainfall parameters, landslide critical rainfall thresholds under different geological lithologies are generated. Landslide critical curves are generated based on the landslide critical rainfall thresholds.

2. The critical early warning method for rainfall-induced landslides according to claim 1, characterized in that, The landslide susceptibility assessment factors include rainfall factors, topographic factors, hydrological environmental factors, geological factors, land cover factors, soil factors, and human activity factors. The determination of landslide susceptibility assessment factors based on the gridded data specifically includes: Based on the rainfall data, a continuous gridded rainfall layer is converted using spatial interpolation to form rainfall factors; The topographic data is input into a preset DEM model to extract topographic parameters and topographic humidity index of the target study area. Topographic factors are generated based on the topographic parameters, and hydrological environmental factors are determined based on the topographic humidity index. By integrating stratigraphic lithology data and soil type data, stratigraphic lithology classification, soil texture and distance from fault zones are analyzed to form geological factors and soil factors; Based on vegetation cover data and human engineering activity data, road, building and vegetation distribution maps are analyzed. The land cover factor is determined according to the vegetation distribution map. The distance of each grid cell in the target study area to the nearest road is calculated based on the road and building distribution map to form the human activity factor.

3. The critical early warning method for rainfall-induced landslides according to claim 1, characterized in that, The process of generating landslide susceptibility results for the target study area based on the landslide susceptibility evaluation factors and their weights, and determining the landslide prediction area based on the landslide susceptibility results, specifically includes: The weight coefficients of each landslide susceptibility evaluation factor were assigned using the analytic hierarchy process (AHP). The landslide susceptibility evaluation factors were then normalized and weighted and superimposed to calculate the landslide susceptibility results for each grid unit in the target study area. Based on the landslide susceptibility results of each grid cell, different levels of landslide-prone areas are divided to determine the landslide prediction area.

4. The critical early warning method for rainfall-induced landslides according to claim 1, characterized in that, The process of obtaining critical rainfall parameters based on the landslide critical curve and determining the probability of rainfall-induced landslides in the landslide prediction area based on the critical rainfall parameters specifically includes: Based on the landslide critical curve, critical rainfall parameters are statistically analyzed to construct a rainfall-type landslide dataset. Based on the rainfall-induced landslide dataset, the random forest algorithm is used to combine the topographic parameters and topographic humidity index of the landslide prediction area to output the probability of rainfall-induced landslides occurring in the landslide prediction area.

5. The critical early warning method for rainfall-induced landslides according to claim 1, characterized in that, Based on the landslide susceptibility results and the probability results of rainfall-induced landslides, a gridded hazard value is generated, and a landslide early warning signal is generated according to the gridded hazard value, specifically including: The landslide susceptibility result of each grid cell in the landslide prediction area is multiplied by the probability result of rainfall-induced landslide occurrence to obtain the gridded hazard value; Landslide warning signals are divided into different levels based on gridded hazard values. Based on landslide warning signals of the same level, adjacent grid units are merged to form a landslide warning area.

6. The critical early warning method for rainfall-induced landslides according to claim 1, characterized in that, After generating gridded hazard values ​​based on the landslide susceptibility results and the probability of rainfall-induced landslides, and generating landslide early warning signals based on the gridded hazard values, the rainfall-induced landslide critical early warning method further includes: Real-time data collection of landslide signs reported by the public; data preprocessing of the landslide signs data to obtain valid data; Based on the Bayesian network model, the effective data and gridded hazard values ​​are fused to dynamically correct the landslide early warning signal. The corrected landslide early warning signal triggers the emergency response signal and sends a targeted early warning notification to the target study area.

7. A rainfall-based landslide critical early warning system, characterized in that, The rainfall-induced landslide critical early warning system includes: The multi-source data processing module is used to acquire multi-source geographic data of the target study area. The multi-source geographic data includes rainfall data, topographic data, stratigraphic lithology data, soil type data, vegetation cover data, and human engineering activity data. The multi-source geographic data is then processed into grids to generate gridded data. The landslide susceptibility analysis module is used to determine landslide susceptibility evaluation factors based on the gridded data, generate landslide susceptibility results for the target study area based on the landslide susceptibility evaluation factors and their weights, and determine the landslide prediction area based on the landslide susceptibility results. The critical curve generation module is used to acquire historical landslide disaster data and real-time rainfall data of the landslide prediction area, calculate the effective rainfall based on the permeability coefficient of the soil and rock mass, and dynamically fit the landslide critical curve under different geological lithologies based on the effective rainfall. The landslide probability prediction module is used to obtain critical rainfall parameters based on the landslide critical curve, and to determine the probability of rainfall-induced landslides in the landslide prediction area based on the critical rainfall parameters. The landslide early warning signal generation module is used to generate a gridded hazard value based on the landslide susceptibility results and the probability results of rainfall-induced landslides, and to generate a landslide early warning signal based on the gridded hazard value.

8. A rainfall-based landslide critical early warning system according to claim 7, characterized in that, The critical curve generation module includes: The initial critical rainfall parameter generation module is used to fit an initial critical curve based on the historical landslide disaster data and obtain the initial critical rainfall parameter according to the initial critical curve. The soil and rock saturation state determination module is used to determine the soil and rock permeability coefficient based on the lithological classification of the strata within the landslide prediction area, and to determine the soil and rock saturation state based on the soil and rock permeability coefficient and real-time rainfall data. The landslide critical curve generation module is used to adjust the initial critical rainfall parameters in real time according to the saturation state of the soil and rock mass, generate landslide critical rainfall thresholds under different geological lithologies based on the adjusted critical rainfall parameters, and generate landslide critical curves based on the landslide critical rainfall thresholds.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the critical early warning method for rainfall-induced landslides as described in any one of claims 1 to 6.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the critical early warning method for rainfall-induced landslides as described in any one of claims 1 to 6.

Images