Method and device for determining risk level of flood disaster and electronic equipment

Abstract

The invention discloses a flood disaster risk level determination method and device and electronic equipment. The method comprises the steps that a current text data set, related to the flood disaster, of a target area, current remote sensing image data and current meteorological data of the target area are acquired, and the current text data set at least comprises public opinion information, related to the flood disaster, of the target area; based on the current text data set, an emotional tendency index is determined, and the emotional tendency index represents the influence degree of subjective emotional information on the flood disaster risk level of the target area; and based on the emotional tendency index, the current remote sensing image data and the current meteorological data, determining the risk level of the flood disaster of the target area. According to the method and the device, the technical problem of low accuracy of risk level determination results of flood disasters in related technologies is solved.

Description

Technical Field

[0001] The present application relates to the field of flood disaster risk assessment, and specifically to a method, device, and electronic equipment for determining the risk level of flood disasters. Background Art

[0002] Floods are a frequent and severe natural disaster worldwide, posing a significant threat to human life and property safety and socioeconomic development. Accurate and timely flood risk classification and early warning are crucial for reducing disaster losses.

[0003] Related technologies typically determine flood risk levels based on geographic information or meteorological data for the target area. Relying on a single type of data limits the comprehensiveness and accuracy of risk assessments. Furthermore, these technologies typically use static data from a fixed point in time to assess flood risk, failing to reflect the latest developments in the disaster, reducing the timeliness and effectiveness of early warnings. Consequently, related technologies suffer from the problem of low accuracy in determining flood risk levels.

[0004] To address the above-mentioned problems, no effective solutions have been proposed so far. Summary of the Invention

[0005] The embodiments of the present application provide a method, device, and electronic device for determining the risk level of flood disasters, so as to at least solve the technical problem of low accuracy of the results of determining the risk level of flood disasters existing in the related art.

[0006] According to one aspect of an embodiment of the present application, a method for determining the risk level of flood disasters is provided, including: obtaining a current text dataset about flood disasters in a target area, current remote sensing image data, and current meteorological data of the target area, wherein the current text dataset at least includes public opinion information about flood disasters in the target area; determining a sentiment tendency index based on the current text dataset, wherein the sentiment tendency index represents the degree of influence of subjective sentiment information on the flood disaster risk level of the target area; determining the flood disaster risk level of the target area based on the sentiment tendency index, the current remote sensing image data, and the current meteorological data.

[0007] According to another aspect of an embodiment of the present application, a device for determining the risk level of flood disasters is provided, including: a data acquisition module, used to acquire a current text data set about flood disasters in a target area, current remote sensing image data, and current meteorological data of the target area, wherein the current text data set at least includes public opinion information about flood disasters in the target area; a sentiment tendency index determination module, used to determine a sentiment tendency index based on the current text data set, wherein the sentiment tendency index represents the degree of influence of subjective sentiment information on the flood disaster risk level of the target area; a risk level determination module, used to determine the risk level of flood disasters in the target area based on the sentiment tendency index, current remote sensing image data, and current meteorological data.

[0008] According to another aspect of an embodiment of the present application, a non-volatile storage medium is provided, which stores a plurality of instructions, wherein the instructions are suitable for being loaded and executed by a processor in any one of the methods for determining the risk level of flood disasters.

[0009] According to another aspect of an embodiment of the present application, an electronic device is provided, comprising: one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors implement any one of the methods for determining the risk level of flood disasters.

[0010] In an embodiment of the present application, by obtaining a current text dataset about flood disasters in a target area, current remote sensing image data, and current meteorological data of the target area, wherein the current text dataset at least includes public opinion information about flood disasters in the target area; based on the current text dataset, determining a sentiment tendency index, wherein the sentiment tendency index represents the degree of influence of subjective sentiment information on the flood disaster risk level in the target area; and determining the flood disaster risk level in the target area based on the sentiment tendency index, current remote sensing image data, and current meteorological data. The purpose is achieved by performing sentiment analysis on the current text dataset about flood disasters in the target area to obtain a sentiment tendency index, and determining the flood disaster risk level in the target area by combining the remote sensing image data and meteorological data, thereby achieving a technical effect of improving the accuracy of the results of determining the flood disaster risk level in the target area, thereby solving the technical problem of low accuracy of the flood disaster risk level determination results existing in related technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0012] Figure 1This is a flow chart of an optional method for determining the risk level of flood disasters provided in an embodiment of the present application;

[0013] Figure 2 This is a block diagram of an optional method for determining the risk level of flood disasters provided in an embodiment of the present application;

[0014] Figure 3 This is a schematic diagram of an optional flood disaster risk level determination device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0015] In order to enable those skilled in the art to better understand the present invention, 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 part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0016] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in a sequence other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0017] According to an embodiment of the present application, a method embodiment of a method for determining the risk level of flood disasters is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0018] Figure 1 is a flow chart of an optional method for determining the risk level of flood disasters provided in an embodiment of the present application, such as Figure 1 As shown, the method includes the following steps:

[0019] Step S102: obtaining a current text dataset about flood disasters in a target area, current remote sensing image data, and current meteorological data of the target area, wherein the current text dataset at least includes public opinion information about flood disasters in the target area;

[0020] It is understandable that a current text dataset regarding flood disasters in the target area can be obtained, such as public opinion information regarding flood disasters in the target area obtained from mainstream social media. Satellite remote sensing technology can be used to collect current remote sensing image data and current meteorological data for the target area. By obtaining text data regarding flood disasters in the target area and combining it with multimodal data such as the current remote sensing image data and current meteorological data of the target area to assess the risk level of the target area, the comprehensiveness and accuracy of risk assessment results caused by a single data source can be avoided, and the accuracy of flood risk level determination results can be improved. At the same time, by real-time collection of the above multimodal data, the timeliness of risk level determination results and the ability to respond immediately to flood disasters can be enhanced.

[0021] Alternatively, the current text dataset can be obtained based on public posts, comments, private messages, and forwarded content related to flood disasters in the target area posted by users on mainstream social platforms. Keywords such as "flood," "waterlogging," "rescue," and "disaster" can be set on mainstream social platforms, and the platform's search function and data collection interface can be used to capture relevant text data. Alternatively, similar keyword searches and API (Application Programming Interface) calls can be used on social platforms to obtain text data related to flood disasters.

[0022] Step S104: determining a sentiment index based on the current text dataset, wherein the sentiment index represents the degree of influence of subjective sentiment information on the flood disaster risk level of the target area;

[0023] As can be understood, based on the current text dataset, a sentiment index is determined to describe the degree to which subjective emotional information influences the flood risk level of the target area. The inclusion of this sentiment index allows flood risk assessment to rely not only on geographic data (i.e., remote sensing imagery) and meteorological data, but also consider the public's emotional reactions and disaster perceptions, achieving a comprehensive assessment of flood risk and improving the accuracy and comprehensiveness of the assessment results.

[0024] In an optional embodiment, when there are multiple types of sentiment polarities, a sentiment tendency index is determined based on the current text data set, including: preprocessing the current text data set to obtain a standard text data set; based on the standard text data set, using a predetermined sentiment analysis model to obtain sentiment polarities corresponding to multiple text data included in the standard text data set, wherein the sentiment analysis model has pre-learned the correlation between text data related to flood disasters and sentiment levels, and the sentiment levels are used to describe the sentiment tendency of text data with respect to flood disasters; based on the sentiment polarities corresponding to multiple text data, the number of text data corresponding to the multiple sentiment polarities is determined; based on the number of text data corresponding to the multiple sentiment polarities, the weights corresponding to the multiple sentiment polarities, and the total number of text data included in the standard text data set, the sentiment tendency index is determined.

[0025] It is understood that if the sentiment polarity used to describe the sentiment orientation of text data regarding flood disasters includes multiple types, such as positive, neutral, and negative, a sentiment orientation index is determined as follows. First, the data in the current text dataset is preprocessed, such as by data cleaning, word segmentation, and removal of stop words and noise data, to obtain a standard text dataset. The standard text dataset is then input into a sentiment analysis model that has pre-learned the correlation between text data related to flood disasters and sentiment levels, to obtain the sentiment polarity corresponding to each of the multiple text data included in the standard text dataset. Based on the sentiment polarity determination results, the number of text data corresponding to each of the multiple sentiment polarities is determined, such as the number of text data with positive sentiment polarity. Based on the number of text data corresponding to each of the multiple sentiment polarities, the weights corresponding to each of the multiple sentiment polarities, and the total number of text data included in the standard text dataset, a sentiment orientation index is determined. As flood disasters develop, public sentiment may change. By continuously monitoring and updating the sentiment orientation index, the risk level of flood disasters and the content of early warning information can be dynamically adjusted to ensure the timeliness and effectiveness of early warnings, thereby facilitating the development of effective and timely flood disaster emergency management and disaster prevention and mitigation measures.

[0026] Optionally, the text data in the current text dataset can be preprocessed in the following ways: First, the text data can be cleaned by removing HTML (Hyper Text Markup Language) tags, special symbols, duplicate characters, and non-text content, etc., to ensure the unity of the text data structure. For example, regular expressions can be used to match and delete HTML tags, and programs can be written to remove special symbols and duplicate characters in the text data. Second, deep learning-based tokenization tools can be used, such as Jieba Tokenizer (mainly for Chinese text data), BERT Tokenizer (i.e., the tokenizer based on the BERT (Bidirectional Encoder Representations from Transformers) model, mainly for English text), etc., to tokenize the text data in combination with the proprietary noun library for flood disasters, and optimize the semantic parsing ability. At the same time, high-frequency meaningless words, such as "de" (的), "le" (了), "shi" (是), etc., can be removed through a predefined stop word list, and noise data, such as advertising links, emoticons, irrelevant picture descriptions, etc., can be eliminated using regular expressions and semantic analysis techniques.

[0027] Optionally, the preprocessed text data can be initially labeled by manual annotation or with the assistance of natural language processing tools, marking the core themes related to flood disasters, such as disaster situation descriptions, rescue needs, emotional expressions, etc., and then adding corresponding tags to each text data. And the preprocessed and labeled text data is stored in a distributed database (such as HBase, which is a distributed, versioned, multi-dimensional columnar storage database with characteristics such as high reliability, high performance, column-oriented, and scalable), and timestamp and geographical tags are used for index management to ensure data traceability and efficient invocation. The above-labeled text data can not only be used to construct the dataset for the sentiment analysis model, but also provide rich historical text data for subsequent flood disaster analysis.

[0028] Optionally, when constructing the above sentiment analysis model, a standard dataset labeled with sentiment polarity (such as the IMDB (Internet Movie Database) movie review dataset, the SST (Stanford Sentiment Treebank) sentiment analysis dataset, etc.) can be used for pre-training to obtain a pre-trained model, and then the model is adapted to the flood disaster scenario through transfer learning, that is, on the basis of the pre-trained model, the model is fine-tuned using text data related to flood disasters to obtain a sentiment analysis model that can better understand and process text data in the field of flood disasters.

[0029] Optionally, an adversarial training mechanism can be introduced into the sentiment analysis model to improve the model's robustness against false information and ambiguous semantics. Specifically, a GAN (Generative Adversarial Network) can be used to generate adversarial text data samples about flood disasters. This adversarial text data sample can be mixed with real text datasets collected through social platforms and then fed into the sentiment analysis model for training. This allows the sentiment analysis model to learn how to identify and resist interference from false information and ambiguous semantics.

[0030] Alternatively, a multi-task learning approach can be used to simultaneously optimize the sentiment analysis model's sentiment classification and key entity recognition capabilities. During the training process of the sentiment analysis model, in addition to sentiment classification of text data (determining the sentiment polarity of the text data), key entities in the text data (such as the names of disaster-stricken areas and rescue teams) are also identified and labeled. This allows for improved comprehensive understanding and processing of text data by the sentiment analysis model through shared model parameters and joint training.

[0031] Alternatively, a deep learning model based on the Transformer architecture can be used to perform sentiment analysis on the preprocessed text data to determine the sentiment polarity of the text data and quantify the sentiment intensity through an attention mechanism. The sentiment polarity information can include three categories: positive, neutral, and negative.

[0032] Alternatively, the sentiment weight (i.e., sentiment index) S can be determined based on the sentiment polarity of multiple text data in the current text dataset. First, different weights are assigned to the three sentiment polarities of positive, neutral, and negative. For example, the positive sentiment weight is set to 0.2, the neutral sentiment weight is set to 0.5, and the negative sentiment weight is set to 0.8. Then, the sentiment weight S is calculated based on the proportion of different sentiment polarities in the text.

[0033] Step S106: Determine the flood disaster risk level of the target area based on the sentiment index, the current remote sensing image data, and the current meteorological data.

[0034] It can be understood that the flood risk level of a target area can be determined based on sentiment indicators that reflect social perceptions and reactions to flood disasters, combined with current remote sensing imagery and meteorological data. By integrating sentiment analysis results with environmental data, a comprehensive and accurate assessment of flood risk can be achieved, enhancing the timeliness and responsiveness of early warning systems and providing strong data support and decision-making basis for emergency management.

[0035] In an optional embodiment, the risk level of flood disasters in the target area is determined based on the sentiment tendency index, current remote sensing image data, and current meteorological data, including: performing feature extraction based on the current remote sensing image data to obtain spatial features, wherein the spatial features are used to describe the geographical environment and water dynamic characteristics of the target area, and the water dynamic characteristics are used to describe the distribution of surface water bodies in the target area and their changes over time; performing feature extraction based on the current meteorological data to obtain temporal features, wherein the temporal features represent the temporal trend of the meteorological changes in the target area; based on the spatial features and the temporal features, a flood disaster risk assessment model is used to obtain a risk assessment result of flood disasters in the target area, wherein the risk assessment result includes at least a physical risk assessment value of flood disasters in the target area and a probability of occurrence of flood disasters in the target area, the flood disaster risk assessment model pre-learns the spatial features obtained based on the remote sensing image data, and the temporal features obtained based on the meteorological data and the correlation between the risk assessment results, and the physical risk assessment value is used to quantify the degree of damage to the target area caused by flood disasters; and determining the risk level based on the sentiment tendency index and the physical risk assessment value.

[0036] It can be understood that the image data processing module in the flood disaster risk assessment model, such as a graph convolutional network model, extracts features from the current remote sensing image data to obtain spatial features describing the geographic environment and water dynamics of the target area (i.e., describing the distribution of surface water bodies in the target area and their changes over time). The meteorological data processing module in the flood disaster risk assessment model, such as a temporal convolutional network model, extracts features from the current meteorological data to obtain temporal features representing the temporal trends of the target area's meteorological conditions. Based on these spatial and temporal features, a flood disaster risk assessment model that has pre-learned the correlation between spatial features derived from remote sensing image data and temporal features derived from meteorological data and risk assessment results is used to obtain a flood disaster risk assessment result for the target area. This risk assessment result includes at least a physical risk assessment value used to quantify the extent of damage caused by flood disasters in the target area, as well as the probability of flood disasters occurring in the target area. The flood disaster risk level of the target area is determined based on the sentiment index and the physical risk assessment value. A flood disaster risk level determination method based on sentiment tendency indicators, current remote sensing image data and current meteorological data, with the help of deep learning models and data fusion technology, can achieve a comprehensive, dynamic and accurate assessment of flood disaster risks, enhance the effectiveness of the early warning system, and provide strong support for flood disaster emergency management.

[0037] Optionally, since the flood disaster risk assessment model mentioned above needs to process not only meteorological data but also remote sensing image data, that is, it needs to process multimodal data, it is necessary to construct a multimodal flood disaster risk assessment model to determine the flood disaster risk assessment results. For remote sensing image data, a graph convolutional network model can be used to process it and extract the spatial features of the remote sensing image data; for meteorological data, a time series convolutional network model can be used to process it and extract the temporal features of the meteorological data.

[0038] Optionally, when processing remote sensing image data, the remote sensing images can be converted into quantitative indicators such as the rate of change of surface water levels and the speed of expansion of waterlogged areas. For example, by analyzing and comparing satellite remote sensing images at different time points, and using image processing algorithms to calculate indicators such as the rate of change of water levels and the speed of expansion of waterlogged areas, the above indicators can be used as one of the input features of the flood disaster risk assessment model to improve the prediction accuracy of the flood disaster risk assessment model. At the same time, geographical entities such as cities, rivers, and dams in the target area are used as nodes, and the spatial relationships between entities (such as adjacent relationships, inclusion relationships, etc.) are used as edges to construct a geographical entity relationship graph. Then, a graph convolutional network is used to perform a convolution operation on the relationship graph to extract spatial correlation features (i.e., spatial features).

[0039] Optionally, when processing meteorological data, the dynamic evolution trend of meteorological data can be captured through a time series convolutional network, and cross-modally aligned with spatial features. The meteorological data (such as rainfall, temperature, humidity, etc.) detected in real time by the weather station are arranged in chronological order and input into the time series convolutional network for training to learn the dynamic evolution law of meteorological data and extract time series features (i.e., time features). At the same time, the time series features extracted by the time series convolutional network and the spatial correlation features extracted by the graph convolutional network are cross-modally aligned through the feature fusion layer of the flood disaster risk assessment model to form a comprehensive multimodal feature representation.

[0040] In an optional embodiment, the risk level is determined based on the sentiment tendency index and the physical risk assessment value, including: obtaining the number of occurrences of flood disasters in the target area within a predetermined historical time period; determining the historical disaster frequency of flood disasters in the target area within the predetermined historical time period based on the number of occurrences; weighting the sentiment tendency index, the physical risk assessment value, and the historical disaster frequency to obtain a comprehensive risk assessment value; and determining the risk level based on the comprehensive risk assessment value.

[0041] As can be understood, flood disaster records for the target area within a predetermined historical time period are collected and the number of flood disasters occurring is counted. Based on the number of occurrences, the historical flood disaster frequency (i.e., the frequency of disasters occurring per unit time) for the target area within the predetermined historical time period is calculated. Weights are then determined for the sentiment index, physical risk assessment value, and historical disaster frequency. A weighted combination of these weights is then performed to obtain a comprehensive risk assessment. Based on this comprehensive risk assessment, the flood risk level for the target area is determined. By integrating sentiment indexes, physical risk assessment values, and historical disaster frequency, a more comprehensive assessment of flood risk levels is achieved. This integration of sentiment indexes, physical risk assessment values, and historical disaster frequency allows for a more comprehensive assessment of flood risk levels. This process not only considers environmental factors but also social perception and response to disasters, thereby improving the comprehensiveness and accuracy of risk level determination. Furthermore, the comprehensive analysis of historical disaster frequency, sentiment indexes, and physical risk assessment values ​​provides decision makers with multi-dimensional data support, enabling flood emergency management and resource allocation decisions to be based on richer and more comprehensive information, thereby enhancing the scientific and effective nature of these decisions.

[0042] Alternatively, the physical risk assessment value P can be determined not only based on the flood risk assessment model, but also based on remote sensing imagery and meteorological data, combined with multi-source data such as the target area's topographical data and historical flood records. GIS (Geographic Information System) technology can be used to perform spatial and overlay analysis on multi-source data to assess the physical risk of floods and obtain the physical risk assessment value P.

[0043] Alternatively, the flood disaster frequency H in each region may be calculated by analyzing flood disaster records in the past few decades (ie, a predetermined historical time period).

[0044] Optionally, a weighted calculation is performed based on the sentiment index S, the physical risk assessment value P, and the historical disaster frequency H to obtain a comprehensive risk assessment value R of flood disasters in the target area. The comprehensive risk assessment value R can be determined as follows:

[0045] R=α·S+β·P+γ·H

[0046] Among them, α, β, and γ are the sentiment tendency index, physical risk assessment value, and weight of historical disaster frequency determined by the entropy weight method, respectively.

[0047] In an optional embodiment, the risk level is determined based on the comprehensive risk assessment value, including: when the comprehensive risk assessment value is less than a preset first threshold, determining the risk level as low risk; or when the comprehensive risk assessment value is greater than or equal to the first threshold and less than a preset second threshold, determining the risk level as medium risk, wherein the first threshold is less than the second threshold; or when the comprehensive risk assessment value is greater than or equal to the second threshold, determining the risk level as high risk.

[0048] It can be understood that if the comprehensive risk assessment value is less than the pre-set first threshold, the flood risk level is assessed as low risk; if the comprehensive risk assessment value is greater than or equal to the first threshold and less than the pre-set second threshold, the flood risk level is assessed as medium risk; and if the comprehensive risk assessment value is greater than or equal to the second threshold, the flood risk level is assessed as high risk. By setting clear thresholds and determining risk levels based on comprehensive risk assessment values, it is possible to more accurately determine when, where, and what level of disaster warnings need to be issued, reducing false alarms and missed alarms and improving the effectiveness and credibility of warning information.

[0049] Optionally, the risk of flood disasters can be divided into three risk levels: low risk, medium risk, and high risk, based on a preset threshold and a comprehensive risk assessment value R. For example, if the first threshold is 0.3 and the second threshold is 0.6, then when R < 0.3, the risk is low; when 0.3 <= R < 0.6, the risk is medium; and when R >= 0.6, the risk is high.

[0050] In an optional embodiment, after determining the risk level based on the sentiment tendency index and the physical risk assessment value, the method further includes: determining the geographic coordinate boundary information of the flood disaster based on the current remote sensing image data, wherein the geographic coordinate boundary information represents the actual impact range of the flood disaster; and determining the response measures to the flood disaster based on the risk level, the probability of occurrence of the flood disaster, and the geographic coordinate boundary information.

[0051] As can be understood, current remote sensing imagery data can be used to determine geographic coordinate boundary information representing the actual impact range of a flood disaster. Response measures to flood disasters can be determined based on the flood risk level, probability of occurrence, and geographic coordinate boundary information. Determining geographic coordinate boundary information based on remote sensing imagery can improve the geographic positioning accuracy of response measures, ensuring that resources can be quickly and accurately deployed to the areas most in need, reducing the indiscriminate deployment and waste of rescue resources.

[0052] Alternatively, the latest satellite remote sensing imagery and advanced image processing techniques can be used to identify and track changes in surface water bodies. By analyzing the dynamic expansion of waterlogged areas and rising water levels, the actual impact of flooding can be determined, generating precise geographic coordinate boundary information for delineating the most severely affected and potentially affected areas. By combining satellite remote sensing imagery with geographic information system (GIS) technology, spatial analysis methods can be used to precisely delineate high-risk areas affected by flooding and generate geographic coordinate boundary information. Satellite remote sensing imagery is used to determine the inundation area, combined with terrain, water system, and other data from the GIS. Using spatial interpolation and buffer analysis, the geographic coordinate boundaries of the affected area can be determined.

[0053] Optionally, a flood disaster risk assessment model based on multi-source data fusion can be used to learn and analyze multi-source data, predict the probability of flood disasters within a specific time period in the future (such as 24 hours, 48 ​​hours, etc.), and quantify it in numerical form.

[0054] Optionally, graded emergency response recommendations (i.e., response measures to flood disasters) can be generated based on the risk level and specific characteristics of the affected area, including the geographic coordinate boundary information of the flood disaster. For low-risk areas, it is recommended to strengthen detection and early warning, and make preventive preparations; for medium-risk areas, it is recommended to organize personnel evacuation, transfer important materials, and carry out emergency rescue and other work; for high-risk areas, it is recommended to immediately activate the emergency plan and carry out comprehensive emergency rescue and disaster relief work.

[0055] Optionally, the risk level of flood disasters can be regularly reassessed based on real-time multi-source data input, and the warning information content can be dynamically updated to ensure the timeliness and accuracy of the warning information. The above warning information includes the probability of flood disasters, the location of the affected area (i.e., geographic coordinate boundary information), and emergency response recommendations. At regular intervals (such as every hour, every half day, etc.), multi-source data is re-collected and analyzed, and parameters such as sentiment tendency weights, physical risk assessment values, and historical disaster frequencies are updated, as well as the comprehensive risk index. The warning information and emergency response recommendations are adjusted according to the new risk level.

[0056] In an optional embodiment, the method further includes: performing error analysis based on the physical risk assessment value and the actual risk value of the flood disaster to obtain an assessment error of the flood disaster risk assessment model; based on the assessment error and a predetermined error threshold, correcting the model parameters of the flood disaster risk assessment model to obtain an updated flood disaster risk assessment model.

[0057] It can be understood that the actual flood risk value is determined, and an error analysis is performed based on the actual risk value and the physical risk assessment value of flood disasters predicted using the flood risk assessment model to obtain the assessment error of the flood risk assessment model. If the assessment error is less than or equal to the preset error threshold, the flood risk assessment model's prediction accuracy meets the requirements. If the assessment error is greater than the preset error threshold, the flood risk assessment model's prediction accuracy is poor. Based on the assessment error, the model parameters of the flood risk assessment model are modified to obtain an updated flood risk assessment model with satisfactory prediction accuracy. Dynamic adjustment of model parameters enables the flood risk assessment model to better adapt to environmental changes and data updates, especially when facing emerging disaster patterns or regional characteristics, so that the model's predictive ability is not affected by inherent biases.

[0058] Optionally, after each flood disaster occurs, the actual risk value of the flood disaster can be obtained by collecting and recording data such as the actual disaster impact area and the extent of the disaster. The actual risk value is compared with the physical risk assessment value predicted by the model to calculate the assessment error. The assessment error can be determined using methods such as mean square error, absolute error, and normalized root mean square error. A statistical analysis is performed on the assessment error, such as plotting an error distribution histogram and calculating the mean and standard deviation of the error. The model's prediction accuracy is then determined based on the statistical analysis results and an error threshold. If the error exceeds the error threshold, it indicates that there is a significant deviation between the model prediction and the actual situation, and parameter correction is required. Based on the results of the error analysis, the causes of the model's inaccurate predictions are identified, such as an unreasonable model structure, improper setting of certain feature weights, insufficient or unbalanced training data, etc. Model parameters are then adjusted based on these causes, including but not limited to changing the number of network layers, adjusting the learning rate, rebalancing feature weights, and increasing the amount of training data. The model is retrained using the corrected parameters and tested on a validation set to ensure that the corrected model's prediction accuracy meets the requirements.

[0059] Through the above steps S102 to S106, the purpose of obtaining a sentiment tendency index by performing sentiment analysis on the current text data set about flood disasters in the target area collected can be achieved, and the risk level of flood disasters in the target area can be determined by combining remote sensing image data and meteorological data, thereby achieving the technical effect of improving the accuracy of the results of determining the risk level of flood disasters in the target area, and thus solving the technical problem of low accuracy of the results of determining the risk level of flood disasters existing in related technologies.

[0060] Based on the above-mentioned embodiments and alternative embodiments, the present application proposes an alternative implementation method for determining the risk level of flood disasters. This implementation method integrates multi-source data such as text data on social platforms, remote sensing image data, and meteorological data, and proposes a method for social media text sentiment analysis, flood disaster risk classification, and early warning based on deep learning. This method uses deep learning technology to挖掘 the potential associations between text data and multi-source data,实现 more accurate and timely flood disaster risk classification and early warning, and provide more effective decision-making support for emergency management.

[0061] Figure 2 is a block diagram of an alternative method for determining the risk level of flood disasters provided by an embodiment of the present application. As Figure 2 shown, the steps of a method for social media text sentiment analysis, flood disaster risk classification, and early warning based on deep learning include:

[0062] Step S1, collect text data on social media platforms, including posts, comments, and private messages related to flood disasters posted by users, and preprocess the text data, including data cleaning, word segmentation, removing stop words, and noise data.

[0063] Collect text data on social media platforms. Obtain text data according to the public posts, comments, private messages, and forwarded content related to flood disasters posted by users in mainstream social media platforms, and focus on text data covering a wide range of user groups and geographical areas. Then, through keyword matching and scraping technology, based on a preset keyword library related to flood disasters such as "flood", "waterlogging", "rescue", "affected", etc., perform real-time scraping and preliminary screening on social platforms,剔除 irrelevant data, and perform format standardization processing on the scraped text data, removing HTML tags, special symbols, duplicate characters, and non-text content to ensure the unity of data structure. Further, use deep learning-based word segmentation tools such as Jieba segmentation, BERT Tokenizer, etc., combined with a proprietary noun library for flood disasters to segment the text and optimize the semantic parsing ability. At the same time, remove high-frequency meaningless words such as "de", "le", "shi", etc. through a predefined stop word list, and use regular expressions and semantic analysis techniques to剔除 noise data such as advertising links, emoticons, and irrelevant picture descriptions. Finally, perform preliminary annotation on the cleaned text data, marking the core themes related to flood disasters such as disaster situation description, rescue needs, and emotion expression, providing basic data support for subsequent sentiment analysis and risk assessment. Store the cleaned and annotated text data in a distributed database, and use time stamps and geographical tags for index management to ensure the traceability and efficient invocation of data.

[0064] In step S2, a deep learning model based on the Transformer architecture is used to perform sentiment analysis on the preprocessed text data, determining the sentiment polarity of the text data and quantifying the sentiment intensity through an attention mechanism. The sentiment polarity information includes three categories: positive, neutral, and negative.

[0065] First, a standard dataset labeled with sentiment polarity is used for pre-training to obtain a pre-trained model, and then transfer learning is used to adapt it to flood disaster scenarios. That is, based on the pre-trained model, the model is fine-tuned using text data related to flood disasters to obtain a sentiment analysis model that can better understand and process text data in the field of flood disasters; then, an adversarial training mechanism is introduced to improve the model's robustness to false information and ambiguous semantics; finally, through multi-task learning, the sentiment classification ability and key entity recognition ability of the sentiment analysis model are simultaneously optimized.

[0066] Step S3 integrates satellite remote sensing images (i.e., remote sensing image data), meteorological data, topographic data, and historical flood disaster records to construct a multimodal flood disaster risk assessment model, in which a graph neural network is used to extract spatial correlation features, and a convolutional neural network is used to extract time series features.

[0067] First, satellite remote sensing images are converted into quantitative indicators such as the rate of change of surface water levels and the expansion speed of waterlogged areas. Then, a geographic entity relationship diagram is constructed, and a graph convolutional network is used to capture the spatial dependencies among cities, rivers, and dams within the basin. Finally, a temporal convolutional network is used to capture the dynamic evolution trend of meteorological data, extract time series features, and perform cross-modal alignment with spatial features.

[0068] Step S4: Combining the sentiment polarity weights of sentiment analysis with the risk assessment results of multi-source data, a weighted fusion algorithm is used to generate a comprehensive risk index (i.e., a comprehensive risk assessment value), and the flood disaster risk is divided into three risk levels: low risk, medium risk, and high risk according to a preset threshold;

[0069] The method for determining the comprehensive risk assessment value R is the same as described above and will not be repeated here. For example, if the first threshold is 0.3 and the second threshold is 0.6, then when R < 0.3, the risk is low; when 0.3 <= R < 0.6, the risk is medium; and when R >= 0.6, the risk is high.

[0070] Step S5: Dynamically generate early warning information including the probability of flood disasters, location of disaster-affected areas, and emergency response recommendations based on the risk level.

[0071] First, based on a flood disaster risk assessment model that integrates multi-source data, the probability of flood disasters occurring within a specific time period in the future (such as 24 hours, 48 ​​hours, etc.) is calculated and quantified in numerical form. Then, by combining satellite remote sensing image data with geographic information system technology, spatial analysis methods are used to accurately delineate the scope of high-risk areas affected by flood disasters and generate geographic coordinate boundary information. At the same time, graded emergency response recommendations are generated based on the risk level and the specific characteristics of the affected area. Finally, the risk level of flood disasters is regularly reassessed based on real-time multi-source data input, and the warning information content is dynamically updated to ensure the timeliness and accuracy of the warning information.

[0072] The above optional implementation method achieves at least the following effects: by obtaining text data about flood disasters in the target area and evaluating the risk level of the target area in combination with multimodal data such as the current remote sensing image data and current meteorological data of the target area, the comprehensiveness and accuracy of the risk assessment results caused by a single data source can be avoided, and the accuracy of the flood disaster risk level determination results can be improved; by setting clear thresholds and determining the risk level based on the comprehensive risk assessment value, it can more accurately judge when and where to issue what level of disaster warning, reduce false alarms and missed alarms, and improve the effectiveness and credibility of the warning information; based on the determination of geographic coordinate boundary information of remote sensing images, the geographic positioning accuracy of response measures can be improved, ensuring that resources can be quickly and accurately deployed to the areas where they are most needed, and reducing the blind deployment and waste of rescue resources.

[0073] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0074] This embodiment also provides a device for determining the risk level of flood disasters. This device is used to implement the above-mentioned embodiments and preferred embodiments, and the details already described will not be repeated. As used below, the terms "module" and "device" can refer to a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0075] According to an embodiment of the present application, there is also provided an embodiment of a device for implementing a method for determining a risk level of a flood disaster. Figure 3 FIG. 1 is a schematic diagram of a device for determining the risk level of flood disasters according to an embodiment of the present application. Figure 3As shown, the above-mentioned flood disaster risk level determination device includes a data acquisition module 302, a sentiment index determination module 304, and a risk level determination module 306. The device is described below.

[0076] The data acquisition module 302 is used to acquire a current text dataset about flood disasters in a target area, current remote sensing image data, and current meteorological data of the target area, wherein the current text dataset at least includes public opinion information about flood disasters in the target area;

[0077] The sentiment index determination module 304 is connected to the data acquisition module 302 and is used to determine a sentiment index based on the current text data set, wherein the sentiment index represents the degree of influence of subjective sentiment information on the flood disaster risk level of the target area;

[0078] The risk level determination module 306 is connected to the sentiment index determination module 304 and is used to determine the risk level of flood disasters in the target area based on the sentiment index, current remote sensing image data, and current meteorological data.

[0079] In a device for determining the risk level of flood disasters provided in an embodiment of the present application, a data acquisition module 302 is provided for acquiring a current text dataset about flood disasters in a target area, current remote sensing image data, and current meteorological data of the target area, wherein the current text dataset includes at least public opinion information about flood disasters in the target area; an emotional tendency index determination module 304 is connected to the data acquisition module 302 and is configured to determine an emotional tendency index based on the current text dataset, wherein the emotional tendency index represents the degree of influence of subjective emotional information on the flood disaster risk level in the target area; and a risk level determination module 306 is connected to the emotional tendency index determination module 304 and is configured to determine the flood disaster risk level in the target area based on the emotional tendency index, current remote sensing image data, and current meteorological data. This achieves the purpose of performing emotional analysis on the current text dataset about flood disasters in the target area to obtain an emotional tendency index, and then determining the flood disaster risk level in the target area by combining the remote sensing image data and meteorological data, thereby achieving the technical effect of improving the accuracy of the flood disaster risk level determination result for the target area, thereby solving the technical problem of low accuracy of the flood disaster risk level determination result existing in the related art.

[0080] It should be noted that the above modules can be implemented by software or hardware. For example, for the latter, it can be implemented in the following ways: the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0081] It should be noted that the data acquisition module 302, sentiment index determination module 304, and risk level determination module 306 correspond to steps S102 to S106 in the embodiment. The examples and application scenarios implemented by these modules and corresponding steps are the same, but are not limited to the contents disclosed in the above embodiment. It should be noted that these modules, as part of the device, can be run on a computer terminal.

[0082] It should be noted that the optional or preferred implementation of this embodiment can be found in the relevant description in the embodiment, which will not be repeated here.

[0083] The above-mentioned flood disaster risk level determination device can also include a processor and a memory. The data acquisition module 302, the emotional tendency index determination module 304, the risk level determination module 306, etc. are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize the corresponding functions.

[0084] The processor includes a kernel, which retrieves the corresponding program unit from memory. There can be one or more kernels. Memory may include non-permanent memory in a computer-readable medium, random access memory (RAM), and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory includes at least one memory chip.

[0085] An embodiment of the present application provides a non-volatile storage medium having a program stored thereon, which, when executed by a processor, implements a method for determining a risk level of a flood disaster.

[0086] An embodiment of the present application provides an electronic device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, the following steps are implemented: obtaining a current text dataset about flood disasters in a target area, current remote sensing image data, and current meteorological data of the target area, wherein the current text dataset includes at least public opinion information about flood disasters in the target area; determining a sentiment tendency index based on the current text dataset, wherein the sentiment tendency index represents the degree of influence of subjective sentiment information on the flood disaster risk level in the target area; determining the flood disaster risk level in the target area based on the sentiment tendency index, current remote sensing image data, and current meteorological data. The device herein may be a server, a PC, or the like.

[0087] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing an initialized program having the following method steps: obtaining a current text data set about flood disasters in a target area, current remote sensing image data, and current meteorological data of the target area, wherein the current text data set at least includes public opinion information about flood disasters in the target area; determining a sentiment tendency index based on the current text data set, wherein the sentiment tendency index represents the degree of influence of subjective sentiment information on the flood disaster risk level of the target area; determining the flood disaster risk level of the target area based on the sentiment tendency index, the current remote sensing image data, and the current meteorological data.

[0088] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0089] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0090] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0091] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0092] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0093] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0094] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0095] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0096] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0097] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A method for determining the risk level of flood disasters, characterized in that: include: Acquire a current text dataset about flood disasters in a target area, including current remote sensing image data and current meteorological data of the target area, wherein the current text dataset at least includes public opinion information about flood disasters in the target area; Determining a sentiment index based on the current text dataset, wherein the sentiment index represents the degree of influence of subjective sentiment information on the flood disaster risk level of the target area; The risk level of flood disasters in the target area is determined based on the sentiment index, the current remote sensing image data, and the current meteorological data.

2. The method according to claim 1, characterized in that In the case where there are multiple types of sentiment polarity, determining the sentiment tendency index based on the current text dataset includes: Preprocessing the current text dataset to obtain a standard text dataset; Based on the standard text dataset, a predetermined sentiment analysis model is used to obtain sentiment polarities corresponding to a plurality of text data included in the standard text dataset, wherein the sentiment analysis model has previously learned an association between text data related to flood disasters and sentiment polarities, and the sentiment polarities are used to describe the sentiment tendency of the text data with respect to flood disasters; Determining the number of text data corresponding to the plurality of sentiment polarities based on the sentiment polarities corresponding to the plurality of text data; The sentiment tendency index is determined based on the amount of text data corresponding to the multiple sentiment polarities, the weights corresponding to the multiple sentiment polarities, and the total amount of text data included in the standard text data set.

3. The method according to any one of claims 1 to 2, characterized in that Determining the flood disaster risk level of the target area based on the sentiment index, the current remote sensing image data, and the current meteorological data includes: Performing feature extraction based on the current remote sensing image data to obtain spatial features, wherein the spatial features are used to describe the geographical environment and water dynamic characteristics of the target area, and the water dynamic characteristics are used to describe the distribution of surface water bodies in the target area and their changes over time; Extracting features based on the current meteorological data to obtain time features, wherein the time features represent a changing trend of the meteorological conditions in the target area over time; Based on the spatial characteristics and the temporal characteristics, a flood disaster risk assessment model is used to obtain a risk assessment result of flood disasters in the target area, wherein the risk assessment result includes at least a physical risk assessment value of flood disasters in the target area and a probability of occurrence of flood disasters in the target area. The flood disaster risk assessment model pre-learns the spatial characteristics obtained based on remote sensing image data and the temporal characteristics obtained based on meteorological data, and the correlation between the risk assessment results. The physical risk assessment value is used to quantify the extent of damage caused by flood disasters in the target area. The risk level is determined based on the emotional tendency index and the physical risk assessment value.

4. The method according to claim 3, characterized in that The determining of the risk level based on the emotional tendency index and the physical risk assessment value includes: Obtaining the number of flood disasters that occurred in the target area within a predetermined historical time period; Determining a historical disaster frequency of flood disasters in the target area within the predetermined historical time period based on the number of occurrences; Performing weighted processing on the sentiment index, the physical risk assessment value, and the historical disaster frequency to obtain a comprehensive risk assessment value; Based on the comprehensive risk assessment value, the risk level is determined.

5. The method according to claim 4, characterized in that Determining the risk level based on the comprehensive risk assessment value includes: When the comprehensive risk assessment value is less than a preset first threshold, determining the risk level as low risk; or If the comprehensive risk assessment value is greater than or equal to the first threshold and less than a preset second threshold, determining the risk level as medium risk, wherein the first threshold is less than the second threshold; or When the comprehensive risk assessment value is greater than or equal to the second threshold, the risk level is determined to be high risk.

6. The method according to claim 3, characterized in that After determining the risk level based on the emotional tendency index and the physical risk assessment value, the method further includes: Determining geographic coordinate boundary information of the flood disaster based on the current remote sensing image data, wherein the geographic coordinate boundary information represents the actual impact range of the flood disaster; Based on the risk level, the probability of occurrence of the flood disaster, and the geographic coordinate boundary information, response measures to the flood disaster are determined.

7. The method according to claim 3, characterized in that The method further comprises: Performing error analysis based on the physical risk assessment value and the actual risk value of flood disasters to obtain an assessment error of the flood disaster risk assessment model; Based on the assessment error and a predetermined error threshold, the model parameters of the flood disaster risk assessment model are corrected to obtain an updated flood disaster risk assessment model.

8. A device for determining the risk level of flood disasters, characterized in that: include: a data acquisition module, configured to acquire a current text dataset related to flood disasters in a target area, current remote sensing image data and current meteorological data of the target area, wherein the current text dataset at least includes public opinion information related to flood disasters in the target area; A sentiment tendency index determination module is used to determine a sentiment tendency index based on the current text data set, wherein the sentiment tendency index represents the degree of influence of subjective sentiment information on the flood disaster risk level of the target area; The risk level determination module is used to determine the risk level of flood disasters in the target area based on the sentiment tendency index, the current remote sensing image data, and the current meteorological data.

9. A non-volatile storage medium, characterized in that: The non-volatile storage medium stores a plurality of instructions, which are suitable for being loaded by a processor and executed by the method for determining the risk level of flood disasters according to any one of claims 1 to 7.

10. An electronic device, characterized in that: include: One or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method for determining the risk level of flood disasters as described in any one of claims 1 to 7.

Images