Comprehensive disaster management method and system and computer program

Abstract

The invention discloses a disaster comprehensive management method and system and a computer program. The method comprises the following steps: acquiring a historical natural information data set; constructing a training decision tree to identify disaster types, and obtaining a disaster type identification model; constructing and training a full-connection neural network to predict a disaster assessment score, and obtaining a disaster assessment score prediction model; natural information data are monitored and acquired in real time, standardization processing and feature extraction are performed, analysis is performed through the trained disaster type identification model and the disaster assessment score prediction model, and disaster types and disaster assessment scores are acquired; acquiring a disaster severity level; and processing by adopting a preset strategy. According to the technical scheme provided by the invention, a plurality of functions are integrated together, so that the management efficiency and the response speed are improved; real-time data acquisition and processing are realized, disaster risks can be found in time, and corresponding early warning and countermeasures can be taken; the artificial intelligence algorithm is used for analyzing disaster data, decision making is assisted, and the accuracy and efficiency of dealing with disasters are improved.

Description

Technical Field

[0001] The present invention relates to the field of disaster management, and particularly to a comprehensive disaster management method, system, and computer program. Background Art

[0002] In the current field of disaster management, the application of technologies presents a diverse and fragmented situation. Although various methods and tools play an indispensable role in each link of disaster management, there is often a lack of integration and real-time nature among these methods and tools, resulting in low disaster management efficiency and even possible missed best rescue opportunities.

[0003] Traditional disaster management methods mostly rely on various independent software and hardware tools. These tools may perform well in specific aspects, but lack unified management to integrate various resources and cannot achieve real-time sharing and communication of information. This scattered state leads to difficulties in data exchange, making it difficult for data between different systems to be compatible and interoperable, and unable to utilize relevant data as much as possible to timely detect disaster risks and take corresponding early warning and response measures.

[0004] In addition, traditional disaster management methods rely on manual identification of disaster categories and severity levels before corresponding measures can be taken, with low efficiency and prone to delays. Summary of the Invention

[0005] In view of the above deficiencies in the current technology, the present invention provides a comprehensive disaster management method. By integrating a variety of cutting-edge technologies, including Internet of Things sensors, cloud computing, artificial intelligence, and big data analysis, it can achieve real-time monitoring and accurate identification of various disasters.

[0006] To achieve the above object, the embodiments of the present invention adopt the following technical solutions:

[0007] A comprehensive disaster management method, comprising:

[0008] Obtain a historical natural information data set;

[0009] Construct a decision tree to identify disaster types, train the decision tree based on the historical natural information data set, and obtain a disaster type identification model;

[0010] Construct a fully connected neural network to predict disaster assessment scores, train the fully connected neural network based on the historical natural information data set, and obtain a disaster assessment score prediction model;

[0011] Use Internet of Things sensors to real-time monitor and obtain natural information data;

[0012] Standardize and extract features from natural information data, and analyze through a trained disaster type recognition model and a disaster assessment score prediction model to obtain the disaster type and the disaster assessment score;

[0013] Obtain the disaster severity level according to the disaster type and the disaster assessment score;

[0014] Take preset strategies for handling according to the disaster type and the disaster severity level.

[0015] According to one aspect of the present invention, the natural information includes one or more of river information, ocean information, meteorological information, geological information, and forest information.

[0016] According to one aspect of the present invention, the standardizing and feature extracting of the natural information data includes:

[0017] Clean the natural information data to remove noise and outliers therein;

[0018] Perform standardization processing on the cleaned data to convert data from different sources and types into a unified format;

[0019] Extract key features from the standardized data.

[0020] According to one aspect of the present invention, the extracting of key features from the standardized data includes:

[0021] Extract basic features from the natural information data;

[0022] Based on the basic features, construct derivative features according to the business logic and historical experience of disaster comprehensive management.

[0023] According to one aspect of the present invention, the derivative features include: time window features, ratio features, and aggregation features.

[0024] According to one aspect of the present invention, the taking of preset strategies for handling according to the disaster category information includes:

[0025] Evaluate the disaster severity level to determine whether it reaches the disaster warning level;

[0026] Match the corresponding emergency plan according to the disaster type and the disaster severity level;

[0027] Alarm for the disaster;

[0028] Notify the relevant departments and management personnel of the disaster type and the disaster severity level.

[0029] According to one aspect of the present invention, the disaster comprehensive management method further includes:

[0030] Periodically optimize the training of the disaster type recognition model and the disaster assessment score prediction model by using the latest natural information, actual disaster type, and disaster severity level datasets.

[0031] According to one aspect of the present invention, the comprehensive disaster management method further includes:

[0032] Display natural information, disaster types, disaster assessment scores, disaster severity levels, and emergency response plans to users by using visualization technology.

[0033] A comprehensive disaster management system includes:

[0034] A sample module for obtaining a historical natural information dataset;

[0035] A disaster type recognition model construction module for constructing a decision tree to recognize disaster types, training the decision tree based on the historical natural information dataset, and obtaining a disaster type recognition model;

[0036] A disaster assessment score prediction model construction module for constructing a fully connected neural network to predict disaster assessment scores, training the fully connected neural network based on the historical natural information dataset, and obtaining a disaster assessment score prediction model;

[0037] A monitoring module for using Internet of Things sensors to monitor and obtain natural information data in real time;

[0038] A disaster detection module for performing standardized processing and feature extraction on natural information data, and analyzing it through the trained disaster type recognition model and disaster assessment score prediction model to obtain disaster types and disaster assessment scores;

[0039] A judgment module for obtaining the disaster severity level according to the disaster type and the disaster assessment score;

[0040] An emergency response module for taking preset strategies for disposal according to the disaster category information.

[0041] A computer program, when executed, implements the steps of the comprehensive disaster management method as described above.

[0042] Advantages of the implementation of the present invention:

[0043] The present invention provides a new type of comprehensive disaster management method with high integration and real-time performance, having the following advantages:

[0044] Comprehensiveness: Compared with traditional decentralized methods or tools, this method integrates multiple functions, improving management efficiency and response speed.

[0045] Real-time performance: Through cloud computing and Internet of Things technologies, real-time data collection and processing are achieved, enabling timely detection of disaster risks and corresponding early warnings and response measures.

[0046] Intelligence: Utilize artificial intelligence and machine learning algorithms to analyze historical disaster data, assist in optimizing emergency plans and making decisions, and improve the accuracy and efficiency of disaster response.

[0047] The comprehensive disaster management provided by the present invention also has the following advantages:

[0048] Scalability: Adopt a microservices architecture design, which can flexibly add or remove functional modules according to requirements, facilitating system expansion and upgrade.

[0049] Unified standard: Provide unified user authentication and data exchange standards, facilitating integration between different systems and information sharing, reducing information silos and data redundancy. Brief Description of the Drawings

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required to be used in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0051] Figure 1 It is a flowchart of a comprehensive disaster management method according to Embodiment 1 of the present invention;

[0052] Figure 2 It is a flowchart of a comprehensive disaster management method according to Embodiment 2 of the present invention. Detailed Embodiments

[0053] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0054] Embodiment 1

[0055] As Figure 1 shown, a comprehensive disaster management method includes:

[0056] S1: Obtain a historical natural information dataset.

[0057] The natural information includes one or more of river information, ocean information, meteorological information, geological information, and forest information.

[0058] S2: Construct a decision tree to identify disaster types, train the decision tree based on historical natural information datasets, and obtain a disaster type identification model.

[0059] Specifically, when training the decision tree, by recursively splitting the data set, the model gradually constructs decision paths for different disaster categories.

[0060] Search for the best decision tree parameter combination, use K-fold cross validation, evaluate model performance, avoid overfitting, find the best parameter combination, and improve model performance.

[0061] In practical applications, decision tree models are used for training in AI-assisted disaster identification. The decision tree model is trained using historical data sets of known disaster events. By recursively splitting the data set, the model gradually constructs decision paths for different disaster types. During the training process, the model automatically selects the optimal split point to maximize the purity of each node, thereby improving the accuracy of the model.

[0062] When new data is input into the model, the decision tree model will quickly identify the possible disaster type based on the pre-built decision path. For example, when the monitored data characteristics match the path of a specific disaster, the model will identify the disaster, enabling a rapid response to potential disaster threats.

[0063] S3: Construct a fully connected neural network to predict disaster assessment scores, train a fully connected neural network based on historical natural information datasets, and obtain a disaster assessment score prediction model.

[0064] Based on the historical natural information data of disasters, including the data monitored by sensors and the actual disaster level at that time, the disaster level is converted into a score, a sample data set is constructed, and the model is trained using the sample data set. The distribution range of disaster assessment scores can be set to 0-100 points.

[0065] In order to improve the accuracy and generalization ability of the model, the cross-validation method is used to optimize the fully connected neural network during model training. By dividing the sample data set into multiple subsets, taking turns as validation sets and the rest as training sets, the performance of the model on different data distributions can be evaluated, and the network structure and hyperparameters can be adjusted accordingly. This process not only reduces the risk of overfitting, but also makes the final disaster assessment score prediction model more robust and reliable.

[0066] During the model training phase, the Early Stopping method is introduced as a regularization technique. When the loss function on the validation set no longer decreases significantly, the training is terminated in a timely manner to prevent the model from overfitting on the training set. Meanwhile, a learning rate decay strategy is used to dynamically adjust the learning rate. A relatively large learning rate is adopted initially for rapid convergence, and then the learning rate is gradually reduced to finely tune the model weights, ensuring that the model can converge near the global optimal solution.

[0067] S4: Use Internet of Things sensors to monitor and obtain natural information data in real time.

[0068] In practical applications, various types of Internet of Things sensors can be deployed in disaster-prone areas, and the data collected can be transmitted to the data processing platform in the cloud in real time through a wireless network. In this way, the real-time nature and accuracy of the data can be ensured, providing a solid foundation for subsequent analysis and processing.

[0069] In practical applications, a water level sensor can be used to monitor tides and potential floods, a meteorological sensor can be used to monitor wind speed, wind direction and rainfall, a temperature sensor and a smoke sensor can be used to monitor forest fire sources, and a seismograph can be used to monitor seismic waves, etc. Various sensors can monitor and collect corresponding natural information.

[0070] S5: Standardize and extract features from the natural information data, and analyze it through the trained disaster type recognition model and disaster assessment score prediction model to obtain the disaster type and disaster assessment score.

[0071] In practical applications, cloud computing technology can be used in the cloud to efficiently process and store a large amount of sensor data.

[0072] Among them, the standardizing and extracting features from the natural information data includes:

[0073] Clean the natural information data to remove noise and outliers;

[0074] Standardize the cleaned data to convert data from different sources and types into a unified format;

[0075] Extract key features from the standardized data.

[0076] Among them, the extracting key features from the standardized data includes:

[0077] Extract basic features from the natural information data;

[0078] Based on the business logic and historical experience of disaster comprehensive management, construct derivative features based on the basic features.

[0079] Specifically, the basic features include: time (hours, day of the week), location (area code), water level height, wind speed, rainfall, temperature, event type, numerical level, etc.

[0080] The derived features include: time window features, such as rainfall and water level changes in the past 1 hour / 3 hours / 12 hours; ratio features, such as rainfall / water level height, abnormal event incidence rate, etc.; aggregation features, such as average temperature over a period of time, total number of events in a specific area, total number of accidents, etc.

[0081] S6: Obtain the disaster severity level based on the disaster type and disaster assessment score.

[0082] In practical applications, the higher the dispersion of the disaster assessment, the higher the disaster severity level, and the more urgent it is to take corresponding measures. For example, if the disaster assessment score ranges from 0 to 100, the disaster severity level is divided into five levels in total. Below 60 points is level one, 60 - 70 points is level two, 70 - 80 is level three, 80 - 90 is level four, and 90 - 100 points is level five.

[0083] S7: Take preset strategies for disposal according to the disaster type and disaster severity level.

[0084] In practical applications, corresponding emergency plans and warning measures can be taken according to the disaster type and disaster severity level.

[0085] Step S7 includes:

[0086] S71: Evaluate the disaster severity level to see if it reaches the disaster warning level.

[0087] In practical applications, corresponding warning levels can be set in advance according to the disaster type and disaster severity level. Once a disaster is identified and data analysis shows that the disaster risk reaches the preset warning level, the system will automatically trigger the warning mechanism.

[0088] For example, when the disaster severity level is level one, it is considered a low - risk level and no warning is required; when the disaster severity level ≥ level two, it is considered that there is a disaster and warning is needed; the higher the disaster severity level, the more serious the disaster situation, and the more urgent it is for all departments to take corresponding measures to mitigate the disaster.

[0089] S72: Match the corresponding emergency plan according to the disaster type and disaster severity level.

[0090] In practical applications, corresponding emergency plans need to be set for different disaster types and disaster severity levels, and corresponding measures should be taken in a timely manner.

[0091] S73: Issue a warning about the disaster.

[0092] In practical applications, a duty officer can be set up to be responsible for checking relevant information. When a disaster is identified, an alarm is immediately issued to remind the duty officer so that timely responses can be made.

[0093] S74: Notify relevant departments and management personnel of the disaster type and the severity level of the disaster.

[0094] In practical applications, the early warning information can be sent to relevant departments and management personnel through multiple channels (such as text messages, APP notifications, broadcasts, etc.) to ensure that corresponding emergency measures can be taken in a timely manner. In this way, not only can disasters be monitored in real time, but also effective preventive and response measures can be taken before the disasters occur, minimizing the losses caused by the disasters to the greatest extent.

[0095] In practical applications, this method is data-driven, unifying the data format to facilitate data exchange, data integration, and information sharing. For example, the data format of disaster information is as follows:

[0096]

[0097]

[0098] This method constructs a complete closed-loop of monitoring, alarming, and information synchronization, driving personnel to execute tasks through data and policies, thus significantly saving communication and management costs. For example: Deploy a variety of Internet of Things sensors in the port area, including water level sensors to monitor tides and potential floods, and meteorological sensors to monitor wind speed, wind direction, rainfall, etc. Collect data from these sensors in real time and transmit it to the central processor of the port through a wireless network. The collected data is matched with a specific port area according to the location field and combined with the occurrence_time field to ensure the timeliness and relevance of the data. When the monitored data (such as water level or wind speed) exceeds the preset severity threshold, the risk is automatically evaluated and an early warning is triggered.

[0099] The beneficial effects of this embodiment are as follows: This method integrates multiple functions, improving the management efficiency and response speed compared with traditional decentralized methods or tools; through cloud computing and Internet of Things technologies, real-time data collection and processing are achieved, enabling timely discovery of disaster risks and corresponding early warning and response measures; using artificial intelligence and machine learning algorithms to analyze historical disaster data to assist in optimizing emergency plans and making decisions, improving the accuracy and efficiency of disaster response.

[0100] Embodiment 2

[0101] Such as Figure 2 As shown, a comprehensive disaster management method includes:

[0102] S1: Obtain a historical natural information data set.

[0103] The natural information includes one or more of river information, ocean information, meteorological information, geological information, and forest information.

[0104] S2: Construct a decision tree to identify disaster types, train the decision tree based on historical natural information datasets, and obtain a disaster type identification model.

[0105] Specifically, when training the decision tree, by recursively splitting the data set, the model gradually constructs decision paths for different disaster categories.

[0106] Search for the best decision tree parameter combination, use K-fold cross validation, evaluate model performance, avoid overfitting, find the best parameter combination, and improve model performance.

[0107] In practical applications, decision tree models are used for training in AI-assisted disaster identification. The decision tree model is trained using historical data sets of known disaster events. By recursively splitting the data set, the model gradually constructs decision paths for different disaster types. During the training process, the model automatically selects the optimal split point to maximize the purity of each node, thereby improving the accuracy of the model.

[0108] When new data is input into the model, the decision tree model will quickly identify the possible disaster type based on the pre-built decision path. For example, when the monitored data characteristics match the path of a specific disaster, the model will identify the disaster, enabling a rapid response to potential disaster threats.

[0109] S3: Construct a fully connected neural network to predict disaster assessment scores, train a fully connected neural network based on historical natural information datasets, and obtain a disaster assessment score prediction model.

[0110] Based on the historical natural information data of disasters, including the data monitored by sensors and the actual disaster level at that time, the disaster level is converted into a score, a sample data set is constructed, and the model is trained using the sample data set. The distribution range of disaster assessment scores can be set to 0-100 points.

[0111] In order to improve the accuracy and generalization ability of the model, the cross-validation method is used to optimize the fully connected neural network during model training. By dividing the sample data set into multiple subsets, taking turns as validation sets and the rest as training sets, the performance of the model on different data distributions can be evaluated, and the network structure and hyperparameters can be adjusted accordingly. This process not only reduces the risk of overfitting, but also makes the final disaster assessment score prediction model more robust and reliable.

[0112] During the model training phase, the Early Stopping method is introduced as a regularization technique. When the loss function on the validation set no longer decreases significantly, the training is terminated in a timely manner to avoid overfitting of the model on the training set. Meanwhile, a learning rate decay strategy is used to dynamically adjust the learning rate. A larger learning rate is adopted initially for rapid convergence, and then the learning rate is gradually decreased to finely tune the model weights, ensuring that the model can converge near the global optimal solution.

[0113] S4: Use Internet of Things sensors to monitor and obtain natural information data in real time.

[0114] In practical applications, various types of Internet of Things sensors can be deployed in disaster-prone areas, and the data collected can be transmitted to the data processing platform in the cloud in real time through a wireless network. In this way, the real-time nature and accuracy of the data can be ensured, providing a solid foundation for subsequent analysis and processing.

[0115] In practical applications, a water level sensor can be used to monitor tides and potential floods, a meteorological sensor can be used to monitor wind speed, wind direction, and rainfall, a temperature sensor and a smoke sensor can be used to monitor forest fire sources, and a seismograph can be used to monitor seismic waves, etc. Various sensors can monitor and collect corresponding natural information.

[0116] S5: Standardize and extract features from the natural information data, and analyze it through the trained disaster type recognition model and disaster assessment score prediction model to obtain the disaster type and disaster assessment score.

[0117] In practical applications, in the cloud, cloud computing technology can be used to efficiently process and store a large amount of sensor data.

[0118] Among them, the standardization processing and feature extraction of the natural information data include:

[0119] Clean the natural information data to remove noise and outliers;

[0120] Perform standardization processing on the cleaned data to convert data from different sources and types into a unified format;

[0121] Extract key features from the standardized data.

[0122] Among them, the extraction of key features from the standardized data includes:

[0123] Extract basic features from the natural information data;

[0124] Based on the business logic and historical experience of disaster comprehensive management, construct derivative features based on the basic features.

[0125] Specifically, the basic features include: time (hours, day of the week), location (area code), water level height, wind speed, rainfall, temperature, event type, numerical level, etc.

[0126] The derived features include: time window features, such as rainfall, water level changes in the past 1 hour / 3 hours / 12 hours, etc.; ratio features, such as rainfall / water level height, abnormal event incidence rate, etc.; aggregation features, such as average temperature in the past period of time, total number of events in a specific area, total number of accidents, etc.

[0127] S6: Obtain the disaster severity level according to the disaster type and disaster assessment score.

[0128] In practical applications, the higher the dispersion of the disaster assessment, the higher the disaster severity level, and the more urgent it is to take corresponding measures in a timely manner. For example, the disaster assessment score ranges from 0 to 100 points, and the disaster severity level is divided into five levels in total. Below 60 points is level one, 60 - 70 points is level two, 70 - 80 is level three, 80 - 90 is level four, and 90 - 100 points is level five.

[0129] S7: Take preset strategies for disposal according to the disaster type and disaster severity level.

[0130] In practical applications, corresponding emergency plans and warning measures can be taken according to the disaster type and disaster severity level.

[0131] Step S7 includes:

[0132] S71: Evaluate the disaster severity level to see if it reaches the disaster warning level.

[0133] In practical applications, corresponding warning levels can be set in advance according to the disaster type and disaster severity level. Once a disaster is identified and data analysis shows that the disaster risk reaches the preset warning level, the system will automatically trigger the warning mechanism.

[0134] For example, when the disaster severity level is level one, it is considered a low - risk level and no warning is required; when the disaster severity level ≥ level two, it is considered that there is a disaster and warning is needed; the higher the disaster severity level, the more serious the disaster situation, and the more urgent it is for each department to take corresponding measures to mitigate the disaster situation.

[0135] S72: Match the corresponding emergency plan according to the disaster type and disaster severity level.

[0136] In practical applications, corresponding emergency plans need to be set for different disaster types and disaster severity levels, and corresponding measures should be taken in a timely manner.

[0137] S73: Alarm for the disaster.

[0138] In practical applications, a duty officer can be set up to be responsible for checking relevant information. When a disaster is identified, an alarm is immediately issued to remind the duty officer so that timely responses can be made.

[0139] S74: Notify the relevant departments and management personnel of the disaster type and the severity level of the disaster.

[0140] In practical applications, the early warning information can be sent to the relevant departments and management personnel through various channels (such as text messages, APP notifications, broadcasts, etc.) to ensure that corresponding emergency measures can be taken in a timely manner. In this way, not only can disasters be monitored in real time, but also effective preventive and response measures can be taken before the disasters occur, minimizing the losses caused by the disasters to the greatest extent.

[0141] S8: Regularly train and optimize the disaster type recognition model and the disaster assessment score prediction model using the latest natural information, actual disaster types, and disaster severity level datasets.

[0142] To ensure the effectiveness of the disaster type recognition model and the disaster assessment score prediction model in a new environment, the models are regularly retrained using the latest disaster data to optimize their performance. In addition, according to the recognition results in practical applications, the splitting criteria of the decision tree can be adjusted to further improve the recognition accuracy of the models.

[0143] Preferably, the method further includes:

[0144] S9: Display the natural information, disaster type, disaster assessment score, disaster severity level, and emergency plan to the user using visualization technology.

[0145] In practical applications, through visualization technology, the duty officer or other relevant personnel can check the relevant information at any time and timely understand the relevant situation of disaster monitoring.

[0146] In practical applications, this method is data-driven, unifying the data format to facilitate data exchange, data integration, and information sharing. For example, the data format of disaster information is as follows:

[0147]

[0148]

[0149] This method constructs a complete closed-loop for monitoring, alarming, and information synchronization. By driving personnel to execute tasks through data and policies, it significantly saves communication and management costs. For example, a variety of Internet of Things sensors are deployed in the port area, including water level sensors to monitor tides and potential floods, and meteorological sensors to monitor wind speed, wind direction, rainfall, etc. Data is collected from these sensors in real-time and transmitted to the central processor of the port via a wireless network. The collected data is matched with specific port areas according to the location field and combined with the occurrence_time field to ensure the timeliness and relevance of the data. When the monitored data (such as water level or wind speed) exceeds the preset severity threshold, the risk is automatically evaluated and a warning is triggered.

[0150] The beneficial effect of this embodiment is that this method can also regularly train and optimize the disaster recognition model using the latest natural information and disaster data sets, adjust the splitting criteria of the decision tree according to the recognition results in actual applications, further improve the recognition accuracy of the model, and improve the performance of the model; using visualization technology, relevant personnel can view the relevant situation at any time.

[0151] Embodiment III

[0152] A comprehensive disaster management system includes:

[0153] A sample module for obtaining a historical natural information data set;

[0154] A disaster type recognition model construction module for constructing a decision tree to recognize the disaster type, training the decision tree based on the historical natural information data set, and obtaining a disaster type recognition model;

[0155] A disaster assessment score prediction model construction module for constructing a fully connected neural network to predict the disaster assessment score, training the fully connected neural network based on the historical natural information data set, and obtaining a disaster assessment score prediction model;

[0156] A monitoring module for using Internet of Things sensors to monitor and obtain natural information data in real-time;

[0157] A disaster detection module for standardizing and extracting features from the natural information data, and analyzing through the trained disaster type recognition model and disaster assessment score prediction model to obtain the disaster type and disaster assessment score;

[0158] A judgment module for obtaining the disaster severity level according to the disaster type and disaster assessment score;

[0159] An emergency response module for taking preset strategies for disposal according to the disaster category information.

[0160] Preferably, the disaster recognition module includes: a data processing module and a model analysis module.

[0161] A data processing module for processing and storing the natural information data obtained by the monitoring module.

[0162] Preferably, the emergency response module includes: a disaster assessment module, an emergency plan matching module, an alarm module, and an information notification module.

[0163] The disaster assessment module is used to assess the severity level of the disaster and determine whether it reaches the disaster warning level;

[0164] The emergency plan matching module is used to match the corresponding emergency plan according to the disaster type and the severity level of the disaster;

[0165] The alarm module is used to alarm the disaster;

[0166] The information notification module is used to notify the disaster type and the severity level of the disaster to the relevant departments and management personnel.

[0167] Preferably, the disaster comprehensive management system further includes:

[0168] A model optimization module for regularly training and optimizing the disaster type recognition model and the disaster assessment score prediction model by using the latest natural information, actual disaster types, and disaster severity level data sets.

[0169] Preferably, the disaster comprehensive management system further includes:

[0170] A visualization module for displaying the natural information, disaster types, disaster assessment scores, disaster severity levels, and emergency plans to users by using visualization technology.

[0171] This system provides a unified user authentication and data exchange standard, facilitating the integration and information sharing between different systems, and reducing information islands and data redundancy.

[0172] This system is data-driven, including the unification of data formats and the architecture design of data-driven strategies, ensuring the real-time and accurate transmission of information.

[0173] This system adopts a front-end and back-end separation architecture design to improve the scalability and flexibility of the system. Through cutting-edge technologies such as cloud computing, the Internet of Things, and artificial intelligence, real-time data processing and intelligent warning functions are achieved.

[0174] This system adopts a microservices architecture design, which can flexibly add or remove functional modules according to requirements, facilitating the expansion and upgrade of the system.

[0175] Example 4

[0176] A computer program, when executed, implements the steps of the comprehensive disaster management method as described in Embodiment 1 or Embodiment 2.

[0177] Embodiment 5

[0178] A readable storage medium stores a computer program as described in Embodiment 4. When the computer program is executed, it implements the steps of the comprehensive disaster management method as described in Embodiment 1 or Embodiment 2.

[0179] As described above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A comprehensive disaster management method, characterized in that: include: Access historical natural information datasets; Construct a decision tree to identify disaster types, train the decision tree based on historical natural information data sets, and obtain a disaster type identification model; Construct a fully connected neural network to predict disaster assessment scores. Train the fully connected neural network based on historical natural information datasets to obtain a disaster assessment score prediction model. Use IoT sensors to monitor and obtain natural information data in real time; Standardize and extract features of natural information data, analyze them through trained disaster type recognition models and disaster assessment score prediction models, and obtain disaster types and disaster assessment scores; Obtain disaster severity levels based on disaster type and disaster assessment score; According to the type of disaster and the severity of the disaster, the preset strategies are adopted to deal with it.

2. The comprehensive disaster management method according to claim 1, characterized in that: The natural information includes one or more of river information, ocean information, meteorological information, geological information, and forest information.

3. The comprehensive disaster management method according to claim 1, characterized in that: The standardization processing and feature extraction of natural information data includes: Clean the natural information data to remove noise and outliers; Standardize the cleaned data and convert data from different sources and types into a unified format; Extract key features from the standardized data.

4. The comprehensive disaster management method according to claim 3, characterized in that: The key features extracted from the standardized data include: Extract basic features from natural information data; According to the business logic and historical experience of comprehensive disaster management, derived features are constructed based on basic features.

5. The comprehensive disaster management method according to claim 4, characterized in that: The derived features include: time window features, ratio features, and aggregation features.

6. The comprehensive disaster management method according to claim 1, characterized in that: The above-mentioned adopting preset strategies for handling according to the disaster category information includes: Assess the severity of the disaster to see if it has reached the disaster warning level; Match the corresponding emergency plan according to the disaster type and disaster severity level; Provide warnings of disasters; Inform relevant departments and managers of the type and severity of disasters.

7. The comprehensive disaster management method according to claim 1, characterized in that: The integrated disaster management approach also includes: The latest natural information, actual disaster types and disaster severity level data sets are regularly used to train and optimize disaster type identification models and disaster assessment score prediction models.

8. The comprehensive disaster management method according to claim 1, characterized in that: The integrated disaster management approach also includes: Natural information, disaster types, disaster assessment scores, disaster severity levels, and emergency plans are displayed to users using visualization technology.

9. A comprehensive disaster management system, characterized in that: include: Sample module, used to obtain historical natural information datasets; The disaster type identification model building module is used to build a decision tree to identify disaster types, train the decision tree based on the historical natural information data set, and obtain a disaster type identification model; A disaster assessment score prediction model building module is used to build a fully connected neural network to predict disaster assessment scores. The fully connected neural network is trained based on a historical natural information data set to obtain a disaster assessment score prediction model. A monitoring module, which is used to monitor and obtain natural information data in real time using IoT sensors; The disaster detection module is used to perform standardization processing and feature extraction on natural information data, and obtain disaster types and disaster assessment scores through analysis using the trained disaster type recognition model and disaster assessment score prediction model; A judgment module is used to obtain the severity level of disasters based on the disaster type and disaster assessment score; The emergency response module is used to adopt preset strategies to deal with disasters based on disaster category information.

10. A computer program, characterized in that When the computer program is executed, the steps of the comprehensive disaster management method according to any one of claims 1 to 8 are implemented.

Images