Geological disaster early warning method, device, storage medium and electronic equipment during typhoon

Abstract

In the geological disaster early warning method, device, storage medium and electronic device provided by the present application during typhoon, multiple geological disaster points are selected in the target early warning area, and according to multiple preset ratios, the first non-geological disaster point of each geological disaster point in each preset ratio is selected nearby; according to each geological disaster point and each first non-geological disaster point, the optimal ratio and target prone area are determined; for each geological disaster point, the second non-geological disaster point is selected nearby in the target prone area according to the optimal ratio; the geological disaster influencing factor, each geological disaster point and each second non-geological disaster point are input into the geological disaster prediction model, and the geological disaster prediction results output by the geological disaster prediction model and the high-risk rainfall area during the typhoon in the target early warning area are used to generate geological disaster early warning information in the target early warning area during the typhoon. In this way, the accuracy of geological disaster risk prediction under typhoon weather can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of disaster warning and forecasting, and in particular to a method, device, storage medium and electronic equipment for warning geological disasters during typhoons. Background Art

[0002] In the current context of frequent extreme weather, heavy rain during typhoons triggers geological disasters, which has become a major challenge for urban management. In coastal cities, the combination of high urbanization and frequent extreme weather events has increased the risk of geological disasters.

[0003] At present, geological disaster warnings during typhoons mainly rely on typhoon paths and rainfall forecasts, as well as data from geological disaster monitoring stations. However, existing geological disaster warnings mostly focus on landslide events, and give less consideration to comprehensive geological disaster warnings such as collapse, landslides, and ground collapse during typhoons. They also fail to fully consider the vulnerability of urban areas, and the method of selecting non-geological disaster samples is also insufficient, resulting in relatively insufficient predictions of geological disaster risks during typhoon weather. Summary of the invention

[0004] The purpose of this application is to solve at least one of the above-mentioned technical defects, especially the relatively insufficient technical defect of the prior art in predicting the risk of geological disasters under typhoon weather.

[0005] In a first aspect, the present application provides a method for early warning of geological disasters during typhoons, the method comprising:

[0006] Select multiple geological disaster points in the target early warning area, and select the first non-geological disaster point corresponding to each geological disaster point in each preset ratio in the target early warning area according to multiple preset ratios, wherein at least two geological disaster points correspond to different geological disaster types;

[0007] According to the predetermined geological disaster impact factors, each geological disaster point and each first non-geological disaster point corresponding to each preset ratio, the optimal ratio is determined in each preset ratio through a preset geological disaster prediction model, the geological disaster impact factors include the impact factors of human factors, and the impact factors of human factors are used for the human risk constraints of the geological disaster prediction model;

[0008] According to each geological hazard point and each first non-geological hazard point corresponding to the optimal ratio, a target prone area in the target warning area is determined, and in the target prone area, a second non-geological hazard point corresponding to each geological hazard point is selected nearby according to the optimal ratio, and the target prone area includes an extremely low prone area and a low prone area;

[0009] The geological disaster influencing factors, various geological disaster points and various second non-geological disaster points are input into the geological disaster prediction model, and the geological disaster prediction results output by the geological disaster prediction model and the high-risk rainfall areas during typhoons in the target warning area are used to generate geological disaster warning information for the target warning area during typhoons.

[0010] In one embodiment, the process of determining the geological disaster impact factor includes:

[0011] Determine multiple geological disaster influencing factors, and determine multiple influencing factors corresponding to each geological disaster influencing factor, where the geological disaster influencing factors include human factors;

[0012] According to the multiple influencing factors corresponding to each geological hazard influencing factor, the correlation between every two geological hazard influencing factors is calculated;

[0013] According to the correlation between every two geological disaster influencing factors, the geological disaster influencing factors with high correlation are eliminated, and the influencing factors corresponding to the geological disaster influencing factors that are not eliminated and the influencing factors corresponding to human factors are used as geological disaster influencing factors.

[0014] In one embodiment, according to the predetermined geological disaster impact factor, each geological disaster point and each first non-geological disaster point corresponding to each preset ratio, the step of determining the optimal ratio among each preset ratio by using a preset geological disaster prediction model includes:

[0015] For each preset ratio, each geological hazard point and each first non-geological hazard point corresponding to the preset ratio are divided into a training set and a test set, and a geological hazard prediction model is trained with the training set and the geological hazard points, and the model performance corresponding to the trained geological hazard model is determined using the test set;

[0016] The model performances corresponding to the preset ratios are compared, and the preset ratio with the best model performance is taken as the optimal ratio.

[0017] In one embodiment, the step of determining the target prone area in the target warning area according to each geological disaster point and each first non-geological disaster point corresponding to the optimal ratio includes:

[0018] Based on the geological hazard points and the first non-geological hazard points corresponding to the optimal ratio, the target warning area is divided into extremely low susceptibility areas, low susceptibility areas, medium susceptibility areas, high susceptibility areas and extremely high susceptibility areas using the natural break method, and the extremely low susceptibility areas and low susceptibility areas are taken as target susceptibility areas.

[0019] In one embodiment, the geological disaster prediction model is an XGBoost model, and the geological disaster prediction model is constructed based on multiple decision trees. The formula corresponding to the geological disaster prediction model is as follows:

[0020]

[0021] Among them, Y represents the final prediction result, X represents the input independent variable vector, that is, the geological disaster impact factor, ⊙ k represents the random vector of parameters of the k-th tree, h refers to the prediction result of the k-th tree, and N represents the total number of decision trees.

[0022] In one embodiment, the process of determining a high-risk rainfall area includes:

[0023] Divide the target warning area into multiple grids and obtain the historical rainfall value of each grid;

[0024] For each grid, if the rainfall in the grid meets the preset conditions, the grid is regarded as a high-risk rainfall grid. The preset conditions include that the rainfall on the day exceeds the first rainfall amount,

[0025] Or, the accumulated rainfall over three days exceeds the first rainfall amount, and the accumulated rainfall over seven days exceeds the second rainfall amount,

[0026] Or, the rainfall on the day does not exceed the first rainfall amount, and the cumulative rainfall over three days is between the third rainfall amount and the first rainfall amount, and the cumulative rainfall over seven days does not exceed the second rainfall amount,

[0027] Or, the rainfall on the day does not exceed the first rainfall, and the accumulated three-day rainfall is between the fourth rainfall and the third rainfall, and the accumulated seven-day rainfall is between the fifth rainfall and the second rainfall, wherein the first rainfall, the second rainfall, the third rainfall, the fourth rainfall and the fifth rainfall are different proportions of the historical rainfall values ​​corresponding to the grid, and the first rainfall is greater than the second rainfall, greater than the fifth rainfall, greater than the third rainfall, and greater than the fourth rainfall;

[0028] Each high-risk rainfall grid is regarded as a high-risk rainfall area.

[0029] In one embodiment, the geological disaster prediction result output by the geological disaster prediction model is a geological disaster susceptibility map of the target warning area, and the steps of generating geological disaster warning information of the target warning area during the typhoon period using the geological disaster prediction result output by the geological disaster prediction model and the high-risk rainfall area during the typhoon period in the target warning area include:

[0030] According to the high-risk rainfall areas, the geological disaster susceptibility map is marked to obtain geological disaster warning information in the target warning area.

[0031] In a second aspect, the present application provides a geological disaster early warning device during a typhoon, the device comprising:

[0032] A geological disaster point selection module is used to select multiple geological disaster points in the target early warning area, and select the first non-geological disaster point corresponding to each geological disaster point in each preset ratio in the target early warning area according to multiple preset ratios, wherein at least two geological disaster points correspond to different geological disaster types;

[0033] An optimal ratio determination module is used to determine the optimal ratio in each preset ratio according to the predetermined geological disaster impact factor, each geological disaster point and each first non-geological disaster point corresponding to each preset ratio, through a preset geological disaster prediction model, wherein the geological disaster impact factor includes an impact factor of human factors, and the impact factor of human factors is used for the human risk constraint of the geological disaster prediction model;

[0034] A target prone area determination module is used to determine the target prone area in the target warning area according to each geological hazard point and each first non-geological hazard point corresponding to the optimal ratio, and in the target prone area, select the second non-geological hazard point corresponding to each geological hazard point nearby according to the optimal ratio, and the target prone area includes an extremely low prone area and a low prone area;

[0035] The geological disaster warning information generation module is used to input geological disaster influencing factors, various geological disaster points and various second non-geological disaster points into the geological disaster prediction model, and use the geological disaster prediction results output by the geological disaster prediction model and the high-risk rainfall areas during the typhoon in the target warning area to generate geological disaster warning information in the target warning area during the typhoon.

[0036] In a third aspect, the present application provides a storage medium storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the geological disaster warning method during typhoons as described in any of the above-mentioned embodiments.

[0037] In a fourth aspect, the present application provides an electronic device, comprising: one or more processors, and a memory;

[0038] The memory stores computer-readable instructions, and when the computer-readable instructions are executed by one or more processors, the steps of the geological disaster early warning method during typhoons as described in any of the above embodiments are performed.

[0039] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0040] In the method, device, storage medium and electronic device for early warning of geological disasters during typhoons provided by the present application, multiple geological disaster points are selected in the target early warning area, and in the target early warning area, the first non-geological disaster point corresponding to each geological disaster point in each preset ratio is selected nearby according to multiple preset ratios, wherein at least two geological disaster points correspond to different types of geological disasters, so as to fully consider the risks of different types of geological disasters and make the geological disaster early warning in the target early warning area more comprehensive and integrated; according to the predetermined geological disaster influencing factors, the geological disaster points and the first non-geological disaster points corresponding to each preset ratio, the optimal ratio is determined in each preset ratio through a preset geological disaster prediction model, the geological disaster influencing factors include the influencing factors of human factors, and the influencing factors of human factors are used for the artificial risk constraints of the geological disaster prediction model, so that more attention can be paid to the vulnerability of human activity areas; according to the geological disaster points and the first non-geological disaster points corresponding to the optimal ratio, the optimal ratio is determined in the target early warning area. The target prone area is selected, and the second non-geological disaster point corresponding to each geological disaster point is selected nearby according to the optimal ratio. The target prone area includes extremely low prone area and low prone area. The non-geological disaster point selected for the first time is screened for each ratio, and the optimal ratio is selected again. The non-geological disaster point is selected in the extremely low prone area and the low prone area. The area range of the selected non-disaster point can be narrowed, thereby eliminating the interference caused by random selection in the whole area, making the selected non-geological disaster point more accurate, more representative and reference, and improving the accuracy and pertinence of the early warning information for non-geological disaster areas; the geological disaster influencing factors, each geological disaster point and each second non-geological disaster point are input into the geological disaster prediction model, and the geological disaster prediction results output by the geological disaster prediction model and the high-risk rainfall area during the typhoon in the target warning area are used to generate the geological disaster warning information of the target warning area during the typhoon. By considering the rainfall during the typhoon, the final result can focus more on the impact of the typhoon on geological disasters. In this way, the accuracy of the geological disaster risk prediction under typhoon weather can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0042] Figure 1 A schematic diagram of a process for early warning of geological disasters during typhoons provided in an embodiment of the present application;

[0043] Figure 2An example diagram of a method for early warning of geological disasters during typhoons provided in an embodiment of the present application;

[0044] Figure 3 A schematic diagram of the structure of a geological disaster warning device during a typhoon provided in an embodiment of the present application;

[0045] Figure 4 A schematic diagram of the internal structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0047] Among natural disasters, some disasters may be attributed to the interaction between different hazards, making them multi-hazard events. The multiple encounters of disaster events not only increase the possibility of secondary disasters but also increase the risk in typical areas. The classification of multi-hazards depends on the interaction between its components. Specifically, multi-hazards can be divided into three categories: concurrent hazards, two or more hazard events occur simultaneously in time and / or space; sequential hazards, when the same hazard recurs within a short period of time, eventually leading to mutual dependence; cascading hazards, in which primary hazards trigger secondary or tertiary hazards or increase the possibility of other hazards, also known as disaster chains. In this application, we will focus on concurrent cascading geological hazards induced by heavy rains during typhoons.

[0048] When evaluating and managing disasters, the vulnerability of the disaster-bearing body should also be taken into account. Urban areas with high road density and areas with high population density are more susceptible to natural disasters than areas with less human activities. Especially during typhoon days, the subsequent heavy rains and geological disasters such as ground collapse and landslides and collapses often meet, which can easily cause greater losses. Extreme weather events such as typhoons and heavy rainfall are frequent in the tropical coastal areas of East Asia, which are prone to geological disasters such as landslides and collapses. In addition, due to the urbanization process, some cities in the region have suffered from severe vegetation stripping, resulting in loose soil and development of soft soil, which is more prone to soil instability and geological disasters, and is expected to cause significant economic losses and casualties. With global warming and the increase in the frequency of extreme rainfall events, coupled with continued urbanization, the demand for disaster prevention and mitigation measures in the region is expected to increase in the future. Therefore, assessing this cascading geological disaster risk warning during typhoons is crucial to effectively prevent and mitigate urban disaster losses.

[0049] Based on this, the present application provides a method for early warning of geological disasters during typhoons. The following embodiments are described by taking the method applied to electronic devices as an example. It can be understood that the electronic devices can be various devices with data processing functions, including but not limited to a single server, a server cluster, a personal laptop computer, a desktop computer, etc. The method includes:

[0050] S101: Select multiple geological disaster points in the target warning area, and select the first non-geological disaster point corresponding to each geological disaster point in each preset ratio in the target warning area according to multiple preset ratios, wherein at least two geological disaster points correspond to different geological disaster types.

[0051] Among them, the target warning area refers to a specific area where geological disaster monitoring and early warning are required. A geological disaster point refers to a specific location in the target warning area where geological disasters are known or may occur. The preset ratio refers to the ratio between geological disaster points and non-geological disaster points, which can be 1:1, 1:2, 1:3, 1:4 and 1:5. The first non-geological disaster point refers to a specific location in the target warning area where a geological disaster does not occur. Geological disaster types refer to different types of geological disasters, such as landslides, collapses, mudslides, ground collapses, etc. Nearby selection refers to searching within a certain radius near the geological disaster point and randomly sampling locations where geological disasters have not occurred. Since the geological landforms and rainfall conditions corresponding to locations farther away are usually different, nearby selection can ensure that the samples are more referenceable.

[0052] In this step, first, multiple geological disaster points can be selected in the target early warning area. The geological disaster points can be known geological disaster occurrence points or locations with potential geological disaster risks. When selecting geological disaster points, the geological disaster type is first identified to ensure that the geological disaster types corresponding to at least two geological disaster points are different, so that the selected geological disaster points can cover multiple types of geological disasters. After that, multiple selection ratios of non-geological disaster points corresponding to each geological disaster point are formulated. For example, it can be stipulated that a corresponding number of non-geological disaster points are selected around each geological disaster point. Then, in the target early warning area, the first non-geological disaster point corresponding to each geological disaster point can be selected nearby according to the preset ratio and the location of each geological disaster point based on the spatial analysis and calculation of geographic information system data. For example, for geological disaster point A, it can select the first non-geological disaster point a and the first non-geological disaster point b nearby in the preset ratio of 1:2, and the first non-geological disaster point c, the first non-geological disaster point d and the first non-geological disaster point e nearby in the preset ratio of 1:3.

[0053] Furthermore, when selecting the geological disaster point and the first non-geological disaster point, a sampling strategy of point sampling can be adopted, that is, the target warning area is divided into multiple grids, each grid represents a point, and the size of each grid can be selected according to the actual situation, for example, the spatial resolution can be 30 meters. The geological disaster point and the first non-geological disaster point are selected on each grid.

[0054] It can be understood that by setting multiple preset ratios, different situations and scenarios can be covered, thereby ensuring the comprehensiveness and accuracy of geological hazard risk assessment, and the deviation or error caused by the selection of a single ratio can be reduced, thereby improving the accuracy and robustness of data sampling.

[0055] S102: According to the predetermined geological disaster influencing factors, each geological disaster point and each first non-geological disaster point corresponding to each preset ratio, the optimal ratio is determined in each preset ratio through a preset geological disaster prediction model, the geological disaster influencing factors include the influencing factors of human factors, and the influencing factors of human factors are used for the human risk constraints of the geological disaster prediction model.

[0056] Among them, geological disaster influencing factors refer to factors corresponding to various factors that may cause geological disasters. Factors that cause geological disasters include but are not limited to topography, geological structure, soil type, vegetation coverage, rainfall, human activities, etc. Geological disaster prediction model refers to a model used to predict the potential occurrence and degree of danger of geological disasters. Human factors refer to the impact of human activities on the geological environment, including development and construction, mining activities, soil and water conservation measures, etc.

[0057] In this step, based on the geological disaster influencing factors, the geological disaster points and non-geological disaster points corresponding to each preset ratio are used to determine the performance of each preset ratio by using a geological disaster prediction model, so as to select the optimal ratio. Specifically, for each preset ratio, the geological disaster points and non-geological disaster points corresponding to the preset ratio can be predicted by the geological disaster prediction model in turn, and the optimal ratio can be determined in each preset ratio according to the geological disaster prediction results. Among them, the optimal ratio represents that the non-geological disaster points selected by the ratio are the most representative. Since geological disaster points can usually be accurately obtained based on historical disaster events, non-geological disaster points are usually randomly selected in the entire domain, with certain uncertainty and inaccuracy. By setting multiple ratios, not only can more non-geological disaster points be selected, but also a more accurate ratio can be determined based on this as the optimal ratio, so as to make further accurate selection of non-geological disaster points.

[0058] It is understandable that when using the geological disaster prediction model, the geological disaster impact factor needs to be input into the model, and the geological disaster impact factor needs to include the impact factor of human factors, which is used for the human risk constraint of the geological disaster prediction model, so as to take human factors as the human impact constraint of the geological disaster prediction model to ensure that the final prediction result of the model can take into account the vulnerability of human areas. Among them, the impact factors of human factors include but are not limited to the annual impervious location area, population density, distance from the road and building density.

[0059] S103: Determine the target prone area in the target warning area according to each geological hazard point and each first non-geological hazard point corresponding to the optimal ratio, and in the target prone area, select the second non-geological hazard point corresponding to each geological hazard point nearby according to the optimal ratio. The target prone area includes an extremely low prone area and a low prone area.

[0060] Among them, the target prone area refers to the area with certain potential geological disaster risk but relatively low risk determined in the geological disaster risk assessment. The extremely low prone area refers to the area classified as the lowest risk in the target warning area. The low prone area refers to the area with low potential geological disaster risk in the target warning area. The second non-geological disaster point refers to the location within the target prone area where no geological disaster occurs.

[0061] In this step, the spatial distribution of the geological hazard points and the first non-geological hazard points corresponding to the optimal ratio, the types of geological hazards and the frequency of geological hazard occurrence can be analyzed first to preliminarily determine the potential prone areas, and then the geological hazard risk level of the target warning area can be comprehensively evaluated through the topography, climate conditions, land use, vegetation coverage, population density and historical disaster records of each geological hazard point and each first non-geological hazard point. According to the risk assessment results, the target warning area is divided into different levels of prone areas, including extremely low prone areas, low prone areas and high prone areas, and extremely low prone areas and low prone areas are used as target prone areas.

[0062] In the target prone area, based on the spatial analysis and calculation of the geographic information system data, the second non-geological disaster point corresponding to each geological disaster point can be selected according to the optimal proportion and the location of each geological disaster point, and the selected second non-geological disaster point is different from the first non-geological disaster point.

[0063] It is understandable that although there is a higher probability of geological disasters in high-prone areas, low-prone areas are not completely immune to geological disasters. Sometimes geological disasters may also occur in low-prone areas. Although the probability is low, it may cause serious losses to the local area. Therefore, expanding the warning range to low-prone areas and extremely low-prone areas can more comprehensively protect the safety of people's lives and property. In addition, sometimes geological disasters may occur in areas that were originally not prone to occur for some reason, such as continuous heavy rainfall leading to increased risk of geological disasters. In this case, timely issuing warnings for low-prone areas and extremely low-prone areas can allow local areas to take measures in advance to reduce losses. Based on this, when issuing geological disaster warnings, not only high-prone areas but also low-prone areas should be considered to ensure the comprehensiveness of geological disaster warnings. Therefore, extremely low-prone areas and low-prone areas can be used as target prone areas, and further research can be conducted on extremely low-prone areas and low-prone areas to ensure that the geological disaster risks in the target warning areas can be fully assessed.

[0064] The first non-geological disaster point is selected from the entire target warning area, which can preliminarily represent the non-geological disaster points in the target warning area. The second non-geological disaster point is further selected from the target prone area, that is, a local selection in the target warning area. Since the target prone area is an extremely low prone area and a low prone area, the second non-geological disaster point selected from it is more representative and can further supplement the selected first non-geological disaster point, making the sample of geological disaster warning more comprehensive and accurate.

[0065] S104: Input the geological disaster influencing factors, each geological disaster point and each second non-geological disaster point into the geological disaster prediction model, and use the geological disaster prediction results output by the geological disaster prediction model and the high-risk rainfall areas during the typhoon in the target warning area to generate geological disaster warning information for the target warning area during the typhoon.

[0066] Among them, high-risk rainfall areas refer to areas where typhoon rainfall that may cause geological disasters is relatively concentrated and of high intensity. Geological disaster warning information refers to warning information generated based on geological disaster prediction models and the situation in high-risk rainfall areas, usually including warning levels, warning area scope, expected occurrence time, possible impact range, etc.

[0067] In this step, each geological disaster point and each second non-geological disaster point are divided into a training set and a test set, and the geological disaster prediction model is trained with the training set, and the trained geological disaster prediction model is tested with the test set until the trained geological disaster prediction model meets the training end condition. At the same time, the influencing factors of human factors still need to be considered during the training process. Afterwards, the geological disaster warning information of the target warning area during the typhoon can be generated based on the trained geological disaster prediction model and the high-risk rainfall area.

[0068] Furthermore, since geological disasters during typhoons are often affected by rainfall, typhoon rainfall information can be taken into account when training the geological disaster prediction model to improve the accuracy of the model in predicting geological disasters during typhoons. Specifically, when dividing the training set and the test set, a validation set can be divided. After the geological disaster test model is trained, the trained geological disaster prediction model and the high-risk rainfall area are coupled, and the model is verified using the validation set. Among them, the method of dividing each sample set into a training set, a test set, and a validation set can be divided in proportion, for example, each sample set is divided according to a ratio of 6:2:2, 60% of the sample set is a training set, 20% of the sample set is a test set, and 20% of the sample set is a validation set.

[0069] Furthermore, the trained geological disaster prediction model can be used to predict the geological disaster sensitivity level of the target warning area, and a comprehensive disaster warning classification table can be generated based on the geological disaster sensitivity level and high-risk rainfall areas. For example, extremely high geological disaster sensitivity levels and extremely high rainfall threshold probability levels can be designated as the highest warning level, and no warning will be issued for extremely low geological disaster sensitivity levels and lower rainfall threshold probabilities.

[0070] It can be understood that each geological disaster point can be paired with its corresponding first non-geological disaster point and second non-geological disaster point according to various preset ratios to form disaster-non-disaster sample sets with multiple ratios. For example, for geological disaster point A, it corresponds to the first non-geological disaster point a and the second non-geological disaster point b in a 1:1 ratio, corresponds to the first non-geological disaster point c and the first non-geological disaster point d in a 1:2 ratio, and corresponds to the second non-geological disaster point e and the second non-geological disaster point f. Then, the geological disaster point A and the first non-geological disaster point a and the second non-geological disaster point b can be combined into a disaster-non-disaster sample with a 1:1 ratio, and the geological disaster point A and the first non-geological disaster point c and the first non-geological disaster point d, as well as the second non-geological disaster point e and the second non-geological disaster point f can be combined into a disaster-non-disaster sample with a 1:2 ratio, and so on to form a disaster-non-disaster sample set with multiple ratios.

[0071] In the above embodiment, a plurality of geological hazard points are selected in the target early warning area, and in the target early warning area, the first non-geological hazard point corresponding to each geological hazard point in each preset ratio is selected nearby according to a plurality of preset ratios, wherein at least two geological hazard points correspond to different geological hazard types, so as to fully consider the risks of different types of geological hazard, and make the geological hazard early warning in the target early warning area more comprehensive and integrated; according to the predetermined geological hazard influencing factors, the geological hazard points and the first non-geological hazard points corresponding to each preset ratio, the optimal ratio is determined in each preset ratio through a preset geological hazard prediction model, the geological hazard influencing factors include the influencing factors of human factors, and the influencing factors of human factors are used for the artificial risk constraints of the geological hazard prediction model, so that more attention can be paid to the vulnerability of human activity areas; according to the geological hazard points and the first non-geological hazard points corresponding to the optimal ratio, the target prone area in the target early warning area is determined, and according to the optimal ratio, the target prone area is determined. The second non-geological disaster point corresponding to each geological disaster point is selected in proportion. The target prone area includes extremely low prone area and low prone area. Each proportion is screened by the non-geological disaster point selected for the first time, and the optimal proportion is selected again. The non-geological disaster point is selected in the extremely low prone area and the low prone area. The area range of the selected non-disaster point can be narrowed, thereby eliminating the interference caused by random selection in the whole area, making the selected non-geological disaster point more accurate, more representative and reference, and improving the accuracy and pertinence of the early warning information for the non-geological disaster area; the geological disaster influencing factors, each geological disaster point and each second non-geological disaster point are input into the geological disaster prediction model, and the geological disaster prediction results output by the geological disaster prediction model and the high-risk rainfall area during the typhoon in the target warning area are used to generate the geological disaster warning information of the target warning area during the typhoon. By considering the rainfall during the typhoon, the final result can focus more on the impact of the typhoon on geological disasters. In this way, the accuracy of the geological disaster risk prediction under typhoon weather can be improved.

[0072] In one embodiment, the process of determining the geological disaster impact factor includes:

[0073] Determine multiple geological disaster influencing factors, and determine multiple influencing factors corresponding to each geological disaster influencing factor, where the geological disaster influencing factors include human factors;

[0074] According to the multiple influencing factors corresponding to each geological hazard influencing factor, the correlation between every two geological hazard influencing factors is calculated;

[0075] According to the correlation between every two geological disaster influencing factors, the geological disaster influencing factors with high correlation are eliminated, and the influencing factors corresponding to the geological disaster influencing factors that are not eliminated and the influencing factors corresponding to human factors are used as geological disaster influencing factors.

[0076] Specifically, the main influencing factors of each geological disaster are first determined, such as geomorphic features, terrain relief and elevation, etc. in geomorphology, soil type, distance to water and land cover type in geography, rainfall, surface temperature and vegetation index in environment, ground sand humidity index, sediment transport factor and water flow power index in hydrology, population density, building density and distance to roads in human activities. After that, multiple influencing factors corresponding to each geological disaster influencing factor are obtained, and then statistical methods are used to calculate the correlation between each two geological disaster influencing factors to understand the degree of association between the various factors. Usually, factors with high correlation may have redundant information, and removing them can simplify the model and reduce multicollinearity. Therefore, according to the results of correlation analysis, the geological disaster influencing factors with high correlation are removed. In order to ensure that human factors can be fully considered in the subsequent model training, when screening various influencing factors, there may be a situation where human factors are completely removed. Therefore, in order to ensure that the influencing factors used in model training include the influencing factors of human factors, the geological disaster influencing factors that have not been removed and their corresponding influencing factors, as well as the influencing factors corresponding to human factors, are required as the final geological disaster influencing factors.

[0077] In this embodiment, factors with high correlation usually lead to multicollinearity problems, that is, there is a high correlation between different independent variables, which may interfere with the stability and explanatory power of the model. Therefore, eliminating factors with high correlation can reduce the impact of multicollinearity on the model. The geological disaster influencing factors and human factors influencing factors after elimination are more representative and independent, which helps to more clearly understand the relationship between each influencing factor and its actual contribution to geological disasters, and further improve the prediction accuracy and reliability of the model.

[0078] In one example, the correlation coefficient and variance inflation factor can be used to calculate the correlation between every two geological disaster influencing factors, where the expression of the correlation coefficient is as follows:

[0079]

[0080] Among them, CC represents the correlation coefficient, F i represents the i-th influencing factor of geological disaster influencing factor F, O i represents the i-th influencing factor of geological disaster influencing factor O, It represents the average value of each influencing factor corresponding to the geological disaster influencing factor F, It represents the average value of each influencing factor corresponding to the geological hazard influencing factor O, and n represents the total number of influencing factors corresponding to the geological hazard influencing factor F and the total number of influencing factors corresponding to the geological hazard influencing factor O.

[0081] The expression of variance inflation factor is as follows:

[0082]

[0083] Among them, VIF stands for variance inflation factor.

[0084] Specifically, when the correlation coefficient value is greater than or equal to 0.5 and the variance inflation factor value is greater than or equal to 4, it means that the correlation between the disaster influencing factors of the two addresses is high.

[0085] It can be understood that the correlation coefficient can measure the degree of linear correlation between two variables, while the variance inflation factor is mainly used to detect multicollinearity. By using the correlation coefficient and the variance inflation factor at the same time, the correlation between two factors can be more comprehensively evaluated, including linear correlation and collinearity. By comprehensively considering linear correlation and multicollinearity, the degree of influence between influencing factors can be better understood, thereby making more accurate predictions and analyses.

[0086] In one embodiment, according to the predetermined geological disaster impact factors, the geological disaster points and the first non-geological disaster points corresponding to each preset ratio, the step of determining the optimal ratio among the preset ratios by using a preset geological disaster prediction model includes:

[0087] For each preset ratio, each geological hazard point and each first non-geological hazard point corresponding to the preset ratio are divided into a training set and a test set, and a geological hazard prediction model is trained with the training set and the geological hazard points, and the model performance corresponding to the trained geological hazard model is determined using the test set;

[0088] The model performances corresponding to the preset ratios are compared, and the preset ratio with the best model performance is taken as the optimal ratio.

[0089] Specifically, for each preset ratio, each geological hazard point and each first non-geological hazard point corresponding to the ratio are divided into a training set and a test set. Ensure that the data distribution of the training set and the test set is uniform and representative. Use the training set data to train the geological hazard prediction model, and you can choose a suitable machine learning algorithm or deep learning model. During the training process, you can perform parameter tuning and other operations as needed. Use the test set data to evaluate the performance of the trained geological hazard prediction model. Common evaluation indicators include accuracy, recall rate, F1 value, etc., to help evaluate the prediction effect and accuracy of the model. For each preset ratio, evaluate it according to the model performance indicators on the test set. Compare the model performance corresponding to each ratio, and select the best performing preset ratio as the optimal ratio based on the model performance comparison results. This ratio can make the geological hazard prediction model reach the best state in terms of prediction accuracy and effect.

[0090] In this embodiment, by dividing the data set into a training set and a test set, and evaluating the model performance on the test set, the generalization ability of the model can be better evaluated to avoid overfitting or underfitting problems. By comparing the model performance corresponding to each preset ratio, the preset ratio that performs best under the current data set can be found, thereby improving the accuracy and effect of the geological disaster prediction model.

[0091] In one embodiment, the step of determining the target prone area in the target early warning area according to each geological disaster point and each first non-geological disaster point corresponding to the optimal ratio includes:

[0092] Based on the geological hazard points and the first non-geological hazard points corresponding to the optimal ratio, the target warning area is divided into extremely low susceptibility areas, low susceptibility areas, medium susceptibility areas, high susceptibility areas and extremely high susceptibility areas using the natural break method, and the extremely low susceptibility areas and low susceptibility areas are taken as target susceptibility areas.

[0093] Specifically, first, the geological hazard risk value of each geological hazard point and each first non-geological hazard point can be calculated based on the optimal ratio corresponding to each geological hazard point and each first non-geological hazard point to determine the potential risk level of each geological hazard point and each first non-geological hazard point. Then, statistical analysis is performed on each geological hazard risk value to determine the distribution characteristics of each geological hazard risk value in the target warning area, and based on the distribution characteristics, natural discontinuities are found in the distribution of each geological hazard risk value, that is, the location where the data distribution changes significantly. According to the natural discontinuities, the target warning area of ​​the distribution of geological hazard risk values ​​is divided into different levels, such as extremely low susceptibility areas, low susceptibility areas, medium susceptibility areas, high susceptibility areas and extremely high susceptibility areas, and extremely low susceptibility areas and low susceptibility areas are used as target susceptibility areas.

[0094] In this embodiment, the natural break method does not rely on subjectively set thresholds or standards, but determines the division points through the inherent distribution characteristics of the data, which is more objective and scientific, and can better adapt to the changes and laws of different data. Therefore, using the natural break method to divide each prone area can avoid human interference and subjective errors, ensuring that the division results are more accurate and reliable.

[0095] In one embodiment, the geological disaster prediction model is an XGBoost model, and the geological disaster prediction model is constructed based on multiple decision trees. The formula corresponding to the geological disaster prediction model is as follows:

[0096]

[0097] Among them, Y represents the final prediction result, X represents the input independent variable vector, that is, the geological disaster impact factor, ⊙ krepresents the random vector of parameters of the k-th tree, h refers to the prediction result of the k-th tree, and N represents the total number of decision trees.

[0098] Specifically, the XGBoost model is used for predictive calculations. The model mode is to make decisions with multiple decision trees, and the result of each tree is the difference between the target value and the prediction results of all previous trees. All the results are accumulated to get the final result, so as to improve the effect of the entire model. The number of decision trees required in the model can be selected according to the actual situation, for example, it can be selected in the range of 2000 to 3000. A decision tree is a tree structure used for analysis and prediction, especially in classification and regression problems. It consists of a series of decision nodes, each node represents a test of an attribute or a group of attributes, and each branch represents a different result of a test output. Each leaf node of the decision tree represents a category or prediction result.

[0099] After obtaining the required number of decision trees, the outputs of each decision tree are voted, and the class with the most votes is used as the final decision. For the prediction problem, which is essentially a regression problem, the relationship between the relevant influencing factors and geological hazards can be established. In the formula of this embodiment, in order to determine the prediction result of the input variable X corresponding to the k-th tree, all decision trees vote for the final prediction result.

[0100] In the present embodiment, XGBoost is an integrated learning model based on a decision tree. By optimizing the objective function and adopting the gradient boosting algorithm, it is possible to achieve very good performance in the prediction task, so the complex relationship between the factors affecting geological hazards can be effectively captured, and the prediction accuracy can be improved. And XGBoost can handle high dimensions, sparse features and nonlinear relationships, and is suitable for data sets containing a variety of geological hazard influencing factors and human factors. Thus, the interaction between each influencing factor can be adaptively learned to better fit the data. In addition, XGBoost uses methods such as regularization technology, pruning and leaf node splitting to effectively avoid overfitting problems and improve the generalization ability of the model, which is particularly important for practical application scenarios such as geological hazard prediction. At the same time, XGBoost performs well in processing large-scale data, can efficiently process in parallel during training, and accelerates the training process of the model, which is very beneficial for the geological hazard prediction system that needs to be frequently updated.

[0101] In one embodiment, the process of determining a high-risk rainfall area includes:

[0102] Divide the target warning area into multiple grids and obtain the historical rainfall value of each grid;

[0103] For each grid, if the rainfall in the grid meets the preset conditions, the grid is regarded as a high-risk rainfall grid. The preset conditions include that the rainfall on the day exceeds the first rainfall amount,

[0104] Or, the accumulated rainfall over three days exceeds the first rainfall amount, and the accumulated rainfall over seven days exceeds the second rainfall amount,

[0105] Or, the rainfall on the day does not exceed the first rainfall amount, and the cumulative rainfall over three days is between the third rainfall amount and the first rainfall amount, and the cumulative rainfall over seven days does not exceed the second rainfall amount,

[0106] Or, the rainfall on the day does not exceed the first rainfall, and the accumulated three-day rainfall is between the fourth rainfall and the third rainfall, and the accumulated seven-day rainfall is between the fifth rainfall and the second rainfall, wherein the first rainfall, the second rainfall, the third rainfall, the fourth rainfall and the fifth rainfall are different proportions of the historical rainfall values ​​corresponding to the grid, and the first rainfall is greater than the second rainfall, greater than the fifth rainfall, greater than the third rainfall, and greater than the fourth rainfall;

[0107] Each high-risk rainfall grid is regarded as a high-risk rainfall area.

[0108] Specifically, the target warning area is divided into multiple grids, and the historical rainfall values ​​of each grid are obtained using remote sensing data, meteorological station data and other sources. Then, preset conditions are defined, including the proportional relationship of the first rainfall, the second rainfall, the third rainfall, the fourth rainfall and the fifth rainfall, as well as the conditional combination of high-risk rainfall. The historical rainfall values ​​of each grid are traversed, and it is determined whether the conditions for high-risk rainfall are met according to the preset conditions. For each grid, the daily rainfall, the accumulated three-day rainfall and the accumulated seven-day rainfall are compared with the preset conditions to determine whether it is a high-risk rainfall grid. The areas identified as high-risk rainfall grids are grouped together to form a high-risk rainfall area.

[0109] In an example, if the proportion corresponding to the first rainfall is 95%, the proportion corresponding to the second rainfall is 90%, the proportion corresponding to the third rainfall is 75%, the proportion corresponding to the fourth rainfall is 40%, and the proportion corresponding to the fifth rainfall is 80%, then the condition for high-risk rainfall is whether the rainfall on the day exceeds 95% of the historical rainfall value of the grid, or the cumulative three-day rainfall and the cumulative seven-day rainfall exceed 95% and 90% of the historical rainfall values ​​of the grid respectively, or the rainfall on the day is less than 95% of the historical rainfall value of the grid, the cumulative three-day rainfall is between 75-95% of the historical rainfall value of the grid and the cumulative seven-day rainfall does not exceed 90%, or the daily rainfall is less than 95% of the historical rainfall value of the grid, the cumulative three-day rainfall is between 40-75% of the historical rainfall value of the grid and the cumulative seven-day rainfall is between 80%-90%.

[0110] In this embodiment, by subdividing the area into multiple grids, the rainfall conditions in each small area can be observed and monitored more finely, and through the refined division, the rainfall conditions in each small area can be more accurately grasped, thereby improving the accuracy of early warning of high-risk rainfall events.

[0111] In one embodiment, the geological disaster prediction result output by the geological disaster prediction model is a geological disaster susceptibility map of the target warning area, and the steps of generating geological disaster warning information of the target warning area during the typhoon period by using the geological disaster prediction result output by the geological disaster prediction model and the high-risk rainfall area during the typhoon period in the target warning area include:

[0112] According to the high-risk rainfall areas, the geological disaster susceptibility map is marked to obtain geological disaster warning information in the target warning area.

[0113] Specifically, the geological disaster prediction model is used to predict the susceptibility of geological disasters in the target warning area and generate a geological disaster susceptibility map. The geological disaster susceptibility map can identify the potential risk level of geological disasters in different areas and provide a basis for subsequent warnings. High-risk rainfall areas are marked on the geological disaster susceptibility map. Combined with the geological disaster susceptibility map and the high-risk rainfall area marks, geological disaster warning information for the target warning area is generated. The warning information may include high-risk geological disaster points, possible disaster types, recommended preventive measures, etc.

[0114] In this embodiment, the geological disaster susceptibility map realizes the possibility of disasters under the influence of environmental factors, that is, possible inducing factors, which facilitates preventive intervention in extremely high and high-prone areas. High-risk rainfall areas reflect the different probability levels of historical disasters triggered by cumulative rainfall. Relying on the geological disaster susceptibility map and high-risk rainfall areas, it is possible to forecast the time of disaster risks in high-prone areas under the condition of mastering meteorological conditions, thereby realizing disaster time and space warning.

[0115] To facilitate understanding of the solution of the present application, a specific example is provided below for illustration.

[0116] like Figure 2As shown in the figure, first, determine the factors affecting geological disasters, including but not limited to geomorphic features, terrain relief, elevation, slope, aspect, terrain position index, terrain ruggedness index, slope length and slope curvature, geomorphic soil type, lithology, distance to water, distance to fault and land cover type, hydrological terrain moisture index, sediment transport factor and water flow power index, environmental rainfall, surface temperature, vegetation index and soil moisture, human activities impervious surface, population density, building density and distance to road. Then, eliminate multiple linear relationships of each influencing factor to complete factor selection, and then select samples by self-supervision using the optimal ratio through the number of historical geological disaster events and the number of historical non-geological disaster events to construct a typhoon and rainstorm geological disaster database. 60% of the typhoon and rainstorm geological disaster database is used as a training set, 20% as a model training test set, and 20% as a model verification test set. The influencing factors of human activities are used as the vulnerability constraints of the model to train XGBoots. The trained XGBoots are used to determine the geological disaster susceptibility mapping, and based on the heavy rain threshold during the typhoon, the model is verified by means of Accuracy and AUC to obtain the geological disaster warning map during the typhoon.

[0117] The following describes the geological disaster warning device during typhoons provided by the embodiment of the present application. The geological disaster warning device during typhoons described below and the geological disaster warning method during typhoons described above can be referred to each other. Figure 3 As shown, the present application provides a geological disaster early warning device during a typhoon, the device comprising:

[0118] The geological disaster point selection module 201 is used to select multiple geological disaster points in the target early warning area, and select the first non-geological disaster point corresponding to each geological disaster point in each preset ratio in the target early warning area according to multiple preset ratios, wherein at least two geological disaster points correspond to different geological disaster types;

[0119] The optimal ratio determination module 202 is used to determine the optimal ratio in each preset ratio according to the predetermined geological disaster impact factor, each geological disaster point and each first non-geological disaster point corresponding to each preset ratio, through a preset geological disaster prediction model, wherein the geological disaster impact factor includes an impact factor of human factors, and the impact factor of human factors is used for the human risk constraint of the geological disaster prediction model;

[0120] The target prone area determination module 203 is used to determine the target prone area in the target warning area according to each geological hazard point and each first non-geological hazard point corresponding to the optimal ratio, and select the second non-geological hazard point corresponding to each geological hazard point in the target prone area according to the optimal ratio. The target prone area includes an extremely low prone area and a low prone area.

[0121] The geological disaster warning information generation module 204 is used to input the geological disaster influencing factors, each geological disaster point and each second non-geological disaster point into the geological disaster prediction model, and use the geological disaster prediction results output by the geological disaster prediction model and the high-risk rainfall areas during the typhoon in the target warning area to generate geological disaster warning information for the target warning area during the typhoon.

[0122] In one embodiment, the optimal ratio determination module 202 includes:

[0123] An influencing factor determination unit is used to determine a plurality of geological disaster influencing factors and a plurality of influencing factors corresponding to each geological disaster influencing factor, wherein the geological disaster influencing factors include human factors;

[0124] A correlation calculation unit, used for calculating the correlation between every two geological disaster influencing factors according to a plurality of influencing factors corresponding to each geological disaster influencing factor;

[0125] The geological disaster impact factor determination unit is used to eliminate the geological disaster impact factors with high correlation according to the correlation between every two geological disaster impact factors, and use the impact factors corresponding to the geological disaster impact factors that have not been eliminated and the impact factors corresponding to the human factors as the geological disaster impact factors.

[0126] In one embodiment, the optimal ratio determination module 202 includes:

[0127] A model performance determination unit is used to divide each geological hazard point and each first non-geological hazard point corresponding to each preset ratio into a training set and a test set, and train a geological hazard prediction model with the training set and the geological hazard points, and use the test set to determine the model performance corresponding to the trained geological hazard model;

[0128] The optimal ratio determination unit is used to compare the model performances corresponding to the preset ratios and take the preset ratio with the best model performance as the optimal ratio.

[0129] In one embodiment, the target prone area determination module 203 includes:

[0130] The target prone area determination unit is used to divide the target warning area into extremely low prone area, low prone area, medium prone area, high prone area and extremely high prone area based on each geological hazard point and each first non-geological hazard point corresponding to the optimal ratio, using the natural break method, and use the extremely low prone area and low prone area as the target prone area.

[0131] In one embodiment, the geological disaster prediction model is an XGBoost model, and the geological disaster prediction model is constructed based on multiple decision trees. The formula corresponding to the geological disaster prediction model is as follows:

[0132]

[0133] Among them, Y represents the final prediction result, X represents the input dependent variable vector, that is, the geological disaster impact factor, ⊙ k represents the random vector of parameters of the k-th tree, h refers to the prediction result of the k-th tree, and N represents the total number of decision trees.

[0134] In one embodiment, the geological disaster warning information generation module 204 includes:

[0135] A historical rainfall value acquisition unit is used to divide the target warning area into multiple grids and obtain the historical rainfall value of each grid;

[0136] The high-risk rainfall grid determination unit is used to, for each grid, if the rainfall of the grid meets a preset condition, determine the grid as a high-risk rainfall grid, the preset condition including that the rainfall on the day exceeds a first rainfall amount,

[0137] Or, the accumulated rainfall over three days exceeds the first rainfall amount, and the accumulated rainfall over seven days exceeds the second rainfall amount,

[0138] Or, the rainfall on the day does not exceed the first rainfall amount, and the cumulative rainfall over three days is between the third rainfall amount and the first rainfall amount, and the cumulative rainfall over seven days does not exceed the second rainfall amount,

[0139] Or, the rainfall on the day does not exceed the first rainfall, and the accumulated three-day rainfall is between the fourth rainfall and the third rainfall, and the accumulated seven-day rainfall is between the fifth rainfall and the second rainfall, wherein the first rainfall, the second rainfall, the third rainfall, the fourth rainfall and the fifth rainfall are different proportions of the historical rainfall values ​​corresponding to the grid, and the first rainfall is greater than the second rainfall, greater than the fifth rainfall, greater than the third rainfall, and greater than the fourth rainfall;

[0140] The high-risk rainfall area determination unit is used to regard each high-risk rainfall grid as a high-risk rainfall area.

[0141] In one embodiment, the geological disaster prediction result output by the geological disaster prediction model is a geological disaster susceptibility map of the target warning area, and the geological disaster warning information generation module 204 includes:

[0142] The geological disaster warning information generation unit is used to mark the geological disaster susceptibility map according to the high-risk rainfall area and obtain the geological disaster warning information in the target warning area.

[0143] In one embodiment, the present application also provides a storage medium, which stores computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the method for early warning of geological disasters during typhoons as described in any of the above embodiments.

[0144] In one embodiment, the present application also provides an electronic device having computer-readable instructions stored therein. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the method for early warning of geological disasters during typhoons as described in any one of the above embodiments.

[0145] Indicatively, if Figure 4 As shown, Figure 4 This is a schematic diagram of the internal structure of an electronic device provided in an embodiment of the present application. The electronic device 300 can be provided as a server. Figure 4 The electronic device 300 includes a processing component 302, which further includes one or more processors, and a memory resource represented by a memory 301, for storing instructions executable by the processing component 302, such as an application. The application stored in the memory 301 may include one or more modules each corresponding to a set of instructions. In addition, the processing component 302 is configured to execute instructions to execute the geological disaster early warning method during typhoons of any of the above embodiments.

[0146] The electronic device 300 may further include a power supply component 303 configured to perform power management of the electronic device 300, a wired or wireless network interface 304 configured to connect the electronic device 300 to a network, and an input / output (I / O) interface 305. The electronic device 300 may operate based on an operating system stored in the memory 301, such as Windows Server™, Mac OS X™, Unix™, Linux™, Free BSD™, or the like.

[0147] Those skilled in the art will understand that Figure 4 The structure shown in the figure is merely a block diagram of a partial structure related to the scheme of the present application, and does not constitute a limitation on the electronic device to which the scheme of the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0148] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the term "include", "comprise" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not clearly listed, or also includes elements inherent to such process, method, article or equipment. In the absence of more restrictions, the elements defined by the sentence "include one..." do not exclude the existence of other identical elements in the process, method, article or equipment including the elements. Herein, "one", "one", "said", "the" and "it" may also include plural forms, unless the context clearly indicates another way. A plurality refers to at least two cases, such as 2, 3, 5 or 8, etc. "And / or" includes any and all combinations of the relevant listed items.

[0149] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can refer to each other.

[0150] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for early warning of geological disasters during typhoons, characterized in that: The method comprises: Selecting a plurality of geological disaster points in the target early warning area, and selecting a first non-geological disaster point corresponding to each of the geological disaster points in each of the preset proportions in the target early warning area according to a plurality of preset proportions, wherein at least two of the geological disaster points correspond to different geological disaster types; According to the predetermined geological disaster impact factor, each geological disaster point and each first non-geological disaster point corresponding to each preset ratio, an optimal ratio is determined in each preset ratio through a preset geological disaster prediction model, wherein the geological disaster impact factor includes an impact factor of human factors, and the impact factor of human factors is used for the human risk constraint of the geological disaster prediction model; According to each geological hazard point and each first non-geological hazard point corresponding to the optimal ratio, a target prone area in the target warning area is determined, and in the target prone area, a second non-geological hazard point corresponding to each geological hazard point is selected nearby according to the optimal ratio, and the target prone area includes an extremely low prone area and a low prone area; The geological disaster influencing factors, each of the geological disaster points and each of the second non-geological disaster points are input into the geological disaster prediction model, and the geological disaster prediction results output by the geological disaster prediction model and the high-risk rainfall areas during the typhoon in the target warning area are used to generate geological disaster warning information for the target warning area during the typhoon.

2. The method for early warning of geological disasters during typhoons according to claim 1, characterized in that: The process of determining the geological disaster impact factor includes: Determine a plurality of geological disaster influencing factors, and determine a plurality of influencing factors corresponding to each of the geological disaster influencing factors, wherein the geological disaster influencing factors include human factors; Calculate the correlation between each two of the geological disaster influencing factors according to the multiple influencing factors corresponding to each of the geological disaster influencing factors; According to the correlation between every two of the geological disaster influencing factors, the geological disaster influencing factors with high correlation are eliminated, and the influencing factors corresponding to the geological disaster influencing factors that are not eliminated and the influencing factors corresponding to the human factors are used as the geological disaster influencing factors.

3. The method for early warning of geological disasters during typhoons according to claim 1, characterized in that: The step of determining the optimal ratio among the preset ratios by using a preset geological disaster prediction model according to the predetermined geological disaster impact factors, the geological disaster points and the first non-geological disaster points corresponding to each preset ratio, comprises: For each of the preset ratios, each geological hazard point and each first non-geological hazard point corresponding to the preset ratio is divided into a training set and a test set, and the training set and the geological hazard points are used to train the geological hazard prediction model, and the test set is used to determine the model performance corresponding to the trained geological hazard model; The model performances corresponding to the preset ratios are compared, and the preset ratio with the best model performance is taken as the optimal ratio.

4. The method for early warning of geological disasters during typhoons according to claim 1, characterized in that: The step of determining the target prone area in the target early warning area according to each geological disaster point and each first non-geological disaster point corresponding to the optimal ratio comprises: Based on the geological disaster points and the first non-geological disaster points corresponding to the optimal ratio, the target warning area is divided into extremely low-prone areas, low-prone areas, medium-prone areas, high-prone areas and extremely high-prone areas using the natural break method, and the extremely low-prone areas and low-prone areas are used as the target prone areas.

5. The method for early warning of geological disasters during typhoons according to claim 1, characterized in that: The geological disaster prediction model is an XGBoost model, which is constructed based on multiple decision trees. The formula corresponding to the geological disaster prediction model is as follows: Among them, Y represents the final prediction result, X represents the input independent variable vector, that is, the geological disaster impact factor, ⊙ k represents the random vector of parameters of the k-th tree, h refers to the prediction result of the k-th tree, and N represents the total number of decision trees.

6. The method for early warning of geological disasters during typhoons according to claim 1, characterized in that: The process of determining the high-risk rainfall area includes: Divide the target warning area into a plurality of grids, and obtain the historical rainfall value of each grid; For each grid, if the rainfall in the grid meets a preset condition, the grid is regarded as a high-risk rainfall grid, wherein the preset condition includes that the rainfall on the day exceeds a first rainfall amount, Or, the accumulated rainfall over three days exceeds the first rainfall amount, and the accumulated rainfall over seven days exceeds the second rainfall amount, Or, the rainfall on the day does not exceed the first rainfall amount, and the cumulative rainfall over three days is between the third rainfall amount and the first rainfall amount, and the cumulative rainfall over seven days does not exceed the second rainfall amount, Or, the rainfall on the day does not exceed the first rainfall, and the accumulated rainfall for three days is between the fourth rainfall and the third rainfall, and the accumulated rainfall for seven days is between the fifth rainfall and the second rainfall, wherein the first rainfall, the second rainfall, the third rainfall, the fourth rainfall and the fifth rainfall are different proportions of the historical rainfall values ​​corresponding to the grid, and the first rainfall is greater than the second rainfall, greater than the fifth rainfall, greater than the third rainfall and greater than the fourth rainfall; Each of the high-risk rainfall grids is used as the high-risk rainfall area.

7. The method for early warning of geological disasters during typhoons according to claim 1, characterized in that: The geological disaster prediction result output by the geological disaster prediction model is a geological disaster susceptibility map of the target warning area. The step of using the geological disaster prediction result output by the geological disaster prediction model and the high-risk rainfall area during the typhoon in the target warning area to generate geological disaster warning information of the target warning area during the typhoon period includes: According to the high-risk rainfall areas, the geological disaster susceptibility map is marked to obtain geological disaster warning information in the target warning area.

8. A geological disaster early warning device during typhoon, characterized in that: The device comprises: A geological disaster point selection module is used to select multiple geological disaster points in the target early warning area, and select the first non-geological disaster point corresponding to each of the geological disaster points in each of the preset proportions in the target early warning area according to multiple preset proportions, wherein at least two of the geological disaster points correspond to different geological disaster types; An optimal ratio determination module is used to determine the optimal ratio among the preset ratios according to the predetermined geological disaster impact factors, the geological disaster points and the first non-geological disaster points corresponding to each preset ratio, through a preset geological disaster prediction model, wherein the geological disaster impact factors include the impact factors of human factors, and the impact factors of human factors are used for the human risk constraints of the geological disaster prediction model; A target prone area determination module is used to determine the target prone area in the target warning area according to each geological hazard point and each first non-geological hazard point corresponding to the optimal ratio, and in the target prone area, select the second non-geological hazard point corresponding to each geological hazard point nearby according to the optimal ratio, wherein the target prone area includes an extremely low prone area and a low prone area; The geological disaster warning information generation module is used to input the geological disaster influencing factors, each of the geological disaster points and each of the second non-geological disaster points into the geological disaster prediction model, and use the geological disaster prediction results output by the geological disaster prediction model and the high-risk rainfall areas during the typhoon in the target warning area to generate geological disaster warning information for the target warning area during the typhoon.

9. A storage medium, characterized in that: The storage medium stores computer-readable instructions, which, when executed by one or more processors, enable the one or more processors to execute the steps of the method for early warning of geological disasters during typhoons as described in any one of claims 1 to 7.

10. An electronic device, characterized in that: include: one or more processors, and memory; The memory stores computer-readable instructions, and when the computer-readable instructions are executed by the one or more processors, the steps of the method for early warning of geological disasters during typhoons as described in any one of claims 1 to 7 are performed.

Images