Flood disaster-oriented disaster situation analysis method and device
By constructing a semantic network of disaster themes and water accumulation points, and combining it with deep learning technology to extract data from multimedia information, the problem of inaccurate disaster assessment in existing technologies has been solved, and more efficient disaster analysis and rescue decision support have been achieved.
Patent Information
- Application Number
- CN202510869064.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-26
- Publication Date
- 2025-10-31
AI Technical Summary
In existing technologies, flood disaster analysis based on social media only uses the location of water accumulation points as an indicator of the importance of the disaster, which makes it difficult to accurately assess the actual disaster situation at the disaster site, leading to misleading rescue decisions.
We construct a disaster-themed semantic network and a waterlogging point semantic network. We extract disaster-themed data and waterlogging point location data from multimedia information using deep learning technology, and combine co-occurrence relationships to construct a hybrid semantic network for disaster analysis.
It improved the accuracy of flood disaster analysis, enhanced disaster relief efficiency, and provided more accurate disaster assessment support.
Smart Images

Figure CN120876186A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of deep learning technology, and in particular to a method and apparatus for disaster analysis of floods. Background Technology
[0002] Disasters caused by natural disasters pose a serious threat to people's lives and property. The rapid and accurate collection and correlation analysis of disaster relief information is of great significance for improving the efficiency of urban disaster relief.
[0003] Crowdsourced data, exemplified by social media, can reflect disaster information from the public's perspective. Compared to satellite remote sensing, it has higher "spatiotemporal resolution" and situational awareness, and has become an important data source for flood control and disaster relief.
[0004] In related technologies, existing flood disaster analysis technologies based on social media analyze disaster information by analyzing the location of water accumulation points at the disaster site, which facilitates the provision of rescue services. However, since water accumulation points at different locations have different disaster risks, such as different degrees of damage to infrastructure or buildings or different densities of trapped people, it is difficult to accurately assess the true disaster situation at the disaster site by using the location of water accumulation points as an indicator of the importance of the disaster. This can mislead rescue decisions and lead to problems such as untimely rescue. Summary of the Invention
[0005] This invention provides a disaster analysis method and apparatus for flood disasters, which addresses the shortcomings of existing technologies that rely solely on the location of water accumulation points as an indicator of the severity of the disaster, making it difficult to accurately assess the actual disaster situation at the disaster site and thus misleading rescue decisions. The method described in this invention improves the accuracy of flood disaster analysis and enhances disaster relief efficiency.
[0006] This invention provides a disaster analysis method for flood disasters, comprising: Multiple disaster-themed data and multiple waterlogging point location data were obtained from multimedia information corresponding to the disaster-stricken areas; A disaster theme semantic network is constructed based on the multiple disaster theme data and the co-occurrence relationships among them; a water accumulation point semantic network is constructed based on the multiple water accumulation point location data and the co-occurrence relationships among them. A hybrid semantic network is constructed based on the co-occurrence relationship between the target disaster theme data in the disaster theme semantic network and the target water accumulation point location data in the water accumulation point semantic network; wherein, the target disaster theme data belongs to the multiple disaster theme data, and the target water accumulation point location data belongs to the multiple water accumulation point location data; Based on the hybrid semantic network, the spatial correlation and disaster semantics of water accumulation points in the disaster-stricken area are analyzed according to real-time disaster data to obtain disaster analysis results.
[0007] According to a disaster analysis method for flood disasters provided by the present invention, the step of constructing a disaster theme semantic network based on the multiple disaster theme data and the co-occurrence relationships among the disaster theme data includes: The disaster theme semantic network is constructed using each disaster theme data as node vectors, the co-occurrence relationship between any two disaster theme data as edge vectors, and the co-occurrence frequency between any two disaster theme data as edge weights.
[0008] According to a disaster analysis method for flooding disasters provided by the present invention, the step of constructing a semantic network of water accumulation points based on the multiple water accumulation point location data and the co-occurrence relationships among the water accumulation point location data includes: The semantic network of water accumulation points is constructed using the location data of each water accumulation point as node vectors, the co-occurrence relationship between any two water accumulation point location data as edge vectors, and the co-occurrence frequency between any two water accumulation point location data as edge weights.
[0009] According to a disaster analysis method for flood disasters provided by the present invention, the step of constructing a hybrid semantic network based on the co-occurrence relationship between target disaster theme data in the disaster theme semantic network and target water accumulation point location data in the water accumulation point semantic network includes: The hybrid semantic network is constructed using the intersection nodes between the disaster-themed semantic network and the water accumulation point semantic network as node vectors, the co-occurrence relationship between any two intersection nodes as edge vectors, and the co-occurrence frequency between any two intersection nodes as edge weights.
[0010] According to a disaster analysis method for flood disasters provided by the present invention, the multiple disaster-related data are obtained through the following steps: The multimedia information is then subjected to word segmentation, stop word removal, and keyword extraction to obtain a disaster data set. The disaster dataset is filtered by topic using a deep learning classification model to obtain multiple disaster-themed data.
[0011] According to a disaster analysis method for floods provided by the present invention, the location data of water accumulation points are obtained through the following steps: The multimedia information is sequentially segmented, stop words are removed, and keywords are extracted to obtain the location words of the water accumulation point; The location words of the water accumulation point are processed based on the sequence prediction model to obtain the location data of the water accumulation point.
[0012] The present invention also provides a disaster analysis device for flood disasters, comprising: The data acquisition module obtains multiple disaster-themed data and multiple water accumulation point location data from the multimedia information corresponding to the disaster-stricken areas; The first network construction module is used to construct a disaster theme semantic network based on the multiple disaster theme data and the co-occurrence relationship between each disaster theme data; and to construct a water accumulation point semantic network based on the multiple water accumulation point location data and the co-occurrence relationship between each water accumulation point location data. The second network construction module is used to construct a hybrid semantic network based on the co-occurrence relationship between the target disaster theme data in the disaster theme semantic network and the target water accumulation point location data in the water accumulation point semantic network; wherein, the target disaster theme data belongs to the plurality of disaster theme data, and the target water accumulation point location data belongs to the plurality of water accumulation point location data; The disaster analysis module is used to analyze the spatial correlation and disaster semantics of water accumulation points in the disaster-stricken area based on the hybrid semantic network and real-time disaster data, and obtain disaster analysis results.
[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the disaster analysis method for flood disasters as described above.
[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the disaster analysis method for flood disasters as described above.
[0015] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the disaster analysis method for flood disasters as described above.
[0016] The present invention provides a disaster analysis method and apparatus for flood disasters. This method acquires multiple disaster theme data and multiple waterlogging point location data from multimedia information corresponding to the disaster-stricken area. It then constructs a disaster theme semantic network based on the co-occurrence relationships between the multiple disaster theme data and the data on each disaster theme; constructs a waterlogging point semantic network based on the co-occurrence relationships between the multiple waterlogging point location data and the data on each waterlogging point location; and constructs a hybrid semantic network based on the co-occurrence relationships between the target disaster theme data in the disaster theme semantic network and the target waterlogging point location data in the waterlogging point semantic network. Finally, based on the hybrid semantic network, it analyzes the spatial association and disaster semantics of waterlogging points in the disaster-stricken area using real-time disaster data to obtain disaster analysis results. By combining the spatial commonalities and semantic characteristics of different affected waterlogging points, the method improves the accuracy of flood disaster analysis and increases disaster relief efficiency. Attached Figure Description
[0017] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.
[0018] Figure 1 This is one of the flowcharts of the disaster analysis method for flood disasters provided by this invention.
[0019] Figure 2 This is a schematic diagram of the structure of the hybrid semantic network provided by the present invention.
[0020] Figure 3 This invention provides a disaster situation display diagram of different water accumulation points based on a hybrid semantic network.
[0021] Figure 4 This is the second flowchart of the disaster analysis method for flood disasters provided by this invention.
[0022] Figure 5 This is a schematic diagram of the disaster analysis device for flood disasters provided by the present invention.
[0023] Figure 6 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation
[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.
[0025] The following is combined Figures 1-5 This invention describes a disaster analysis method and apparatus for flood disasters.
[0026] Figure 1 This is one of the flowcharts illustrating the disaster analysis method for flood disasters provided by this invention, such as... Figure 1 As shown, the method includes the following steps: Step 110: Obtain multiple disaster-related data and multiple waterlogging location data from the multimedia information corresponding to the disaster-stricken areas.
[0027] In this embodiment, the disaster-stricken area is the region that has been impacted by disasters such as urban flooding, secondary waterlogging caused by geological disasters, or dam-break floods.
[0028] In this embodiment, disaster-related data includes one or more of the following themes: rainfall, building damage, people trapped (population density, health status), and road conditions (such as whether the road is clear or the depth of water accumulation).
[0029] In this embodiment, a water accumulation point refers to a low-lying area where water accumulates due to drainage system failure or topographical factors, typically manifested as road flooding or water ingress into underground facilities; the location information of the water accumulation point includes latitude and longitude coordinates.
[0030] In this embodiment, the location data of water accumulation points can be obtained through methods such as model prediction or sensor monitoring.
[0031] Step 120: Construct a disaster theme semantic network based on multiple disaster theme data and the co-occurrence relationship between each disaster theme data; construct a water accumulation point semantic network based on multiple water accumulation point location data and the co-occurrence relationship between each water accumulation point location data.
[0032] In this step, the disaster themes or water accumulation points that appear simultaneously in the current scenario have a co-occurrence relationship.
[0033] In this step, the disaster-themed semantic network can be constructed by using different types of disaster themes as network nodes and the co-occurrence relationships between disaster themes as edges.
[0034] For example, if there are 10 disaster themes at a flood disaster site, and every two disaster themes co-occur, then the semantic network for this disaster theme includes 10 nodes. 10 Edge.
[0035] In this step, the water accumulation point semantic network can be constructed by using different water accumulation point locations as network nodes and the co-occurrence relationships between water accumulation point locations as edges.
[0036] For example, if there are 5 waterlogged locations at a flood disaster site, and every 2 waterlogged locations have a co-occurrence relationship, then the semantic network of this waterlogged location includes 5 nodes. 5 Edge.
[0037] Step 130: Construct a hybrid semantic network based on the co-occurrence relationship between the target disaster theme data in the disaster theme semantic network and the target water accumulation point location data in the water accumulation point semantic network; wherein, the target disaster theme data belongs to multiple disaster theme data, and the target water accumulation point location data belongs to multiple water accumulation point location data.
[0038] In this step, each node in the hybrid semantic network represents either disaster-related data or water accumulation location data; the edge between two adjacent nodes in the hybrid semantic network indicates that both disaster and water accumulation occur simultaneously.
[0039] For example, the disaster-themed semantic network includes 10 nodes, 2 10 The semantic network of water accumulation points consists of 5 nodes and 2 edges. 5 If there are 4 overlapping or intersecting nodes between two networks, then there are 4 nodes in the hybrid semantic network. All nodes have a co-occurrence relationship, and the weight of each node is determined based on the frequency of simultaneous occurrence of disasters and water accumulation points.
[0040] Step 140: Based on the hybrid semantic network, analyze the spatial correlation and disaster semantics of water accumulation points in the disaster-stricken area according to real-time disaster data to obtain disaster analysis results.
[0041] In this step, after constructing the above-mentioned hybrid semantic network, the corresponding thematic disaster features and water accumulation point location features extracted from the collected real-time disaster data are input into the hybrid semantic network, and the disaster analysis results are output, including the spatial correlation results of water accumulation points and the disaster semantic results.
[0042] In this embodiment, the spatial association results and disaster semantic results of the aforementioned water accumulation points are visualized in the form of a network diagram, which facilitates intuitive reading of the distribution of different types of disaster themes (semantic characteristics) at each water accumulation point (spatial characteristics) in the current disaster-stricken area, and provides decision support for disaster reduction and emergency response.
[0043] The disaster analysis method for floods provided in this invention obtains multiple disaster theme data and multiple waterlogging point location data from multimedia information corresponding to the disaster-stricken area. It then constructs a disaster theme semantic network based on the co-occurrence relationships between the multiple disaster theme data and the data on each disaster theme; constructs a waterlogging point semantic network based on the co-occurrence relationships between the multiple waterlogging point location data and the data on each waterlogging point location; and constructs a hybrid semantic network based on the co-occurrence relationships between the target disaster theme data in the disaster theme semantic network and the target waterlogging point location data in the waterlogging point semantic network. Finally, based on the hybrid semantic network, it analyzes the spatial association and disaster semantics of waterlogging points in the disaster-stricken area using real-time disaster data to obtain disaster analysis results. By combining the spatial commonalities and semantic characteristics of different affected waterlogging points, the method improves the accuracy of flood disaster analysis and increases disaster relief efficiency.
[0044] In some embodiments, constructing a disaster theme semantic network based on multiple disaster theme data and the co-occurrence relationship between each disaster theme data includes: using each disaster theme data as a node vector, using the co-occurrence relationship between any two disaster theme data as an edge vector, and using the co-occurrence frequency between any two disaster theme data as an edge weight to construct the disaster theme semantic network.
[0045] In this embodiment, multiple disaster-themed data are obtained from multimedia information corresponding to the disaster-stricken areas to form a disaster-themed word set T=[ In constructing the main disaster response network, That is, the nodes of the network being constructed. .
[0046] In this embodiment, the weight of an edge in the disaster topic semantic network can be determined based on the frequency of occurrence of any two disaster topics that have a co-occurrence relationship; that is, the higher the frequency, the greater the weight of the edge.
[0047] Specifically, first calculate and The co-occurrence of nodes indicates the connections between them; the degree of a node represents the number of edges connected to that node. Therefore, for a node... Its degree It can be represented as: ; in, yes and The co-occurrence relationship is represented if and Co-occurrence, then the node and If there is an edge, then ;otherwise .
[0048] In this embodiment, the weight of the edge It can reflect the co-occurrence frequency among keywords, indicating the strength of the semantic relationship between disaster-related topics. It can be represented as: ; in, express and The number of co-occurrences in the same text; a disaster-themed semantic network is constructed based on the above node vectors, edge vectors, and edge weights.
[0049] The disaster analysis method for flood disasters provided in this embodiment of the invention constructs a disaster theme semantic network by using each disaster theme data as node vectors, the co-occurrence relationship between any two disaster theme data as edge vectors, and the co-occurrence frequency between any two disaster theme data as edge weights. This provides associated semantic characteristics for different waterlogging points, thereby improving the accuracy of disaster analysis.
[0050] In some embodiments, constructing a semantic network for water accumulation points based on multiple water accumulation point location data and the co-occurrence relationship between each water accumulation point location data includes: using each water accumulation point location data as a node vector, using the co-occurrence relationship between any two water accumulation point location data as an edge vector, and using the co-occurrence frequency between any two water accumulation point location data as an edge weight to construct the semantic network for water accumulation points.
[0051] In this embodiment, the weight of the edge in the water accumulation point semantic network can be determined based on the frequency of occurrence of any two water accumulation point positions that have a co-occurrence relationship; that is, the higher the frequency, the greater the weight of the edge.
[0052] In this embodiment, location words are extracted sequentially from the set of extracted location words for water accumulation points. Each positional word (corresponding to the location data of each water accumulation point) is used as a node vector; and edge vectors are set according to the co-occurrence relationship between each node vector.
[0053] Next, calculate the different positional words that have a co-occurrence relationship (nodes have edge vectors). and The co-occurrence frequency between edges is used to set edge weights.
[0054] Specifically, determine the water accumulation points. node degree ,as well as and Previous edge weights The calculation formula for determining the co-occurrence frequency among various disaster themes is as described above; finally, the semantic network of water accumulation points is formed through the above node vectors, edge vectors and edge weights.
[0055] The disaster analysis method for flood disasters provided in this invention constructs a semantic network of water accumulation points by using the location data of each water accumulation point as node vectors, the co-occurrence relationship between any two water accumulation point location data as edge vectors, and the co-occurrence frequency between any two water accumulation point location data as edge weights. This method can fully explore the spatial distribution relationship of each water accumulation point location in the disaster-stricken area and provide reliable data support for subsequent disaster analysis combined with different disaster themes.
[0056] In some embodiments, constructing a hybrid semantic network based on the co-occurrence relationship between target disaster theme data in the disaster theme semantic network and target water accumulation point location data in the water accumulation point semantic network includes: using the intersection nodes between the disaster theme semantic network and the water accumulation point semantic network as node vectors, the co-occurrence relationship between any two intersection nodes as edge vectors, and the co-occurrence frequency between any two intersection nodes as edge weights to construct the hybrid semantic network.
[0057] In this embodiment, the weights of edges in the hybrid semantic network represent the frequency of simultaneous occurrence of disasters and water accumulation points.
[0058] In this embodiment, the intersection of the water accumulation point semantic network and the disaster topic semantic network is sought, i.e. and The co-occurrence situation between them, and then according to and The co-occurrence situation, establish and The connections between them are identified, and the edge weights between them are determined based on their specific co-occurrence frequencies, ultimately determining the form of the hybrid disaster semantic network.
[0059] In this embodiment, after obtaining the hybrid disaster semantic network, the spatial correlation of water accumulation points and the semantic relationship of disaster are analyzed by combining the features of the hybrid semantic network to provide decision support for disaster reduction and emergency response. Specifically, this includes: (1) Visualize the relationship between water accumulation points and analyze the spatial distribution pattern of water accumulation points.
[0060] (2) Analyze the disaster situation of related waterlogging points by combining the semantic features of the disaster situation.
[0061] The disaster analysis method for flood disasters provided in this embodiment of the invention constructs a hybrid semantic network by using the intersection nodes between the disaster theme semantic network and the water accumulation point semantic network as node vectors, the co-occurrence relationship between any two intersection nodes as edge vectors, and the co-occurrence frequency between any two intersection nodes as edge weights. This method can improve the accuracy of disaster analysis by combining the spatial distribution characteristics and corresponding semantic distribution characteristics of water accumulation points in the disaster-stricken area.
[0062] Figure 2 This is a schematic diagram of the structure of the hybrid semantic network provided by the present invention. Figure 2 In the illustrated embodiment, the community formed by M4 is a waterlogging point semantic network, and its node P is a waterlogging point location word. The community formed by M1, M2, and M3 is a disaster semantic network, and its nodes represent disaster-affected keywords. P1, P2, P3, and P4 are the intersection points of different networks, and together they constitute a disaster hybrid semantic network.
[0063] In some embodiments, multiple disaster-themed data are obtained through the following steps: performing word segmentation, stop word removal, and keyword extraction on the multimedia information in sequence to obtain a disaster dataset; and filtering the disaster dataset based on a deep learning classification model to obtain multiple disaster-themed data.
[0064] In this embodiment, users can use disaster-related keywords, such as "City A; rainstorm; flood", to obtain text data containing various disaster information from multiple social media platforms.
[0065] In this embodiment, a word segmentation tool can be used to perform sentence splitting and word segmentation operations on the text data, and a standardized stop word list is used to filter out meaningless words such as "de" and "le". Finally, nouns / verbs are selected through词性标注 (such as "collapse" and "rescue") to improve the subsequent analysis efficiency.
[0066] In this embodiment, the TF-IDF algorithm is used to calculate the term frequency-inverse document frequency weights in the text data, extract keywords with high discrimination, and calculate the importance of words based on a graph model, which is suitable for the extraction of core concepts in long texts.
[0067] In this embodiment, a deep learning classification model is used to filter out data related to the disaster situation from it, and multiple disaster theme data are obtained.
[0068] For example, a BiLSTM-CRF model is used to identify disaster-related entities (such as place names and building names), word vectors are generated through Word2Vec / GloVe, and the keyword library is expanded by combining semantic similarity.
[0069] The disaster situation analysis method for flood disasters provided by the embodiments of the present invention obtains a disaster situation data set by sequentially performing word segmentation, stop word removal, and keyword extraction operations on multimedia information; based on a deep learning classification model, the disaster situation data set is subject to theme filtering to obtain multiple disaster theme data, realizing high-precision semantic parsing of multiple disaster themes, and can also provide dynamic decision support for scenarios such as disaster warning and resource scheduling.
[0070] In some embodiments, the waterlogging point location data is obtained through the following steps: sequentially performing word segmentation, stop word removal, and keyword extraction operations on multimedia information to obtain waterlogging point location words; based on a sequence prediction model, the waterlogging point location words are processed to obtain waterlogging point location data.
[0071] In this embodiment, a word segmentation tool can be used to perform sentence splitting and word segmentation operations on the text data, first identify disaster domain entities, and then combine a custom stop word list (filtering out function words such as "de" and "le") and disaster scenario extended stop words (such as irrelevant symbols) to retain geographical location keywords; finally, the text in the image (such as road signs and water level gauges) is recognized through OCR technology, location descriptions such as "low-lying section of XX Road" and "water accumulation depth of 30 cm" are extracted, and high-frequency words (such as "water accumulation" and "waterlogging") are selected using the TF-IDF algorithm, and semantic related words are extracted by combining the TextRank algorithm to obtain the corresponding waterlogging point location words.
[0072] In this embodiment, a sequence prediction model including ENIRE and BERT is used to predict the waterlogging point location words to obtain waterlogging point location data, such as a waterlogging point location sequence.
[0073] The disaster analysis method for flood disasters provided in this invention involves performing word segmentation, stop word removal, and keyword extraction on multimedia information to obtain location words of waterlogged points; then processing these location words based on a sequence prediction model to obtain location data of waterlogged points, thereby enabling the accurate extraction of location data of waterlogged points from multimedia information and providing decision support for urban flooding early warning and emergency dispatch.
[0074] Figure 3 This invention provides a disaster situation display map of different water accumulation points based on a hybrid semantic network. Figure 3 In the illustrated embodiment, the disaster-related hybrid semantic network constructed from three nodes—Town D, City A, and Station B—reveals the situation of post-disaster social relief and the plight of the people. Co-occurring keywords include terms such as "missing," "workers," "sick," "safe," "emergency," "cars," "survival," "missing person notices," "rescue," "heavy rain," "skills," and "elderly." These keywords reflect the urgent need for relief and the pressing issue of life safety for the people after the disaster. These co-occurrences reflect the common focus of post-disaster society, especially emergency relief and the life-or-death issues of the people. Town D and City A co-occurred keywords such as "missing," "workers," and "sick," reflecting another predicament faced by the people during the disaster: the problem of people being trapped and losing contact, particularly workers. The emergency medical needs of groups and sick individuals revealed a strong shared concern between the two regions after the disaster, particularly regarding missing persons and medical assistance. The non-co-occurring keywords revealed the specific characteristics of the disaster. For example, the non-co-occurring keywords in City A, such as "electrical leakage," "rescue," and "trapped," likely reflected the damage to infrastructure and the need for rescue of trapped individuals in the area. Meanwhile, Station B emphasized issues like "breaking train windows," "rescue," and "inside the train," reflecting the challenges Station B faced in rescuing people trapped inside vehicles. This indicates that although the three disaster-stricken areas share common needs in post-disaster relief, differences in the extent of infrastructure damage and emergency response strategies may lead to different relief priorities in each region; notably, Henan corresponds to the area south of the Yangtze River.
[0075] Figure 4 This is the second flowchart of the disaster analysis method for flood disasters provided by this invention. Figure 4In the illustrated embodiment, the method includes the following steps: acquiring social media data; filtering disaster-related data to obtain disaster-related social media text; extracting location words of waterlogging points to obtain a set of location words for waterlogging points; performing word segmentation, stop word removal, and keyword extraction on the disaster-related social media text, and calculating the degree of co-occurrence between disaster-themed data and between waterlogging point locations, respectively, and constructing a disaster semantic network (corresponding to the disaster-themed semantic network) and a waterlogging point semantic network based on the co-occurrence frequency; constructing a disaster hybrid semantic network (corresponding to the hybrid semantic network) by connecting the intersection points of the disaster semantic network and the waterlogging point semantic network; finally, obtaining the spatial distribution pattern of waterlogging points based on the disaster hybrid semantic network, and analyzing the disaster situation of associated waterlogging points.
[0076] The disaster analysis device for flood disasters provided by the present invention will be described below. The disaster analysis device for flood disasters described below can be referred to in correspondence with the disaster analysis method for flood disasters described above.
[0077] Figure 5 This is a schematic diagram of the disaster analysis device for flood disasters provided by the present invention, as shown below. Figure 5 As shown, the disaster analysis device for flood disasters includes: a data acquisition module 510, a first network construction module 520, a second network construction module 530, and a disaster analysis module 540.
[0078] The data acquisition module 510 acquires multiple disaster-themed data and multiple water accumulation point location data from the multimedia information corresponding to the disaster-stricken area; The first network construction module 520 is used to construct a disaster theme semantic network based on multiple disaster theme data and the co-occurrence relationship between each disaster theme data; and to construct a water accumulation point semantic network based on multiple water accumulation point location data and the co-occurrence relationship between each water accumulation point location data. The second network construction module 530 is used to construct a hybrid semantic network based on the co-occurrence relationship between the target disaster theme data in the disaster theme semantic network and the target water accumulation point location data in the water accumulation point semantic network; wherein, the target disaster theme data belongs to multiple disaster theme data and the target water accumulation point location data belongs to multiple water accumulation point location data. The disaster analysis module 540 is used to analyze the spatial correlation and disaster semantics of water accumulation points in disaster-stricken areas based on real-time disaster data using a hybrid semantic network, and obtain disaster analysis results.
[0079] The disaster analysis device for flood disasters provided in this invention obtains multiple disaster theme data and multiple waterlogging point location data from multimedia information corresponding to the disaster-stricken area. It then constructs a disaster theme semantic network based on the co-occurrence relationships between the multiple disaster theme data and the data on each disaster theme; constructs a waterlogging point semantic network based on the co-occurrence relationships between the multiple waterlogging point location data and the data on each waterlogging point location; and constructs a hybrid semantic network based on the co-occurrence relationships between the target disaster theme data in the disaster theme semantic network and the target waterlogging point location data in the waterlogging point semantic network. Finally, based on the hybrid semantic network, it analyzes the spatial association and disaster semantics of waterlogging points in the disaster-stricken area using real-time disaster data to obtain disaster analysis results. By combining the spatial commonalities and semantic characteristics of different affected waterlogging points, the accuracy of flood disaster analysis is improved, and disaster relief efficiency is increased.
[0080] Figure 6 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6 As shown, the electronic device may include a processor 610, a communications interface 620, a memory 630, and a communication bus 640. The processor 610, communications interface 620, and memory 630 communicate with each other via the communication bus 640. The processor 610 can call logical instructions in the memory 630 to execute a disaster analysis method for flooding. This method includes: acquiring multiple disaster theme data and multiple waterlogging point location data from multimedia information corresponding to the disaster-stricken area; constructing a disaster theme semantic network based on the multiple disaster theme data and the co-occurrence relationships between them; constructing a waterlogging point semantic network based on the multiple waterlogging point location data and the co-occurrence relationships between them; and constructing a hybrid semantic network based on the co-occurrence relationships between target disaster theme data in the disaster theme semantic network and target waterlogging point location data in the waterlogging point semantic network. The target disaster theme data belongs to multiple disaster theme data, and the target waterlogging point location data belongs to multiple waterlogging point location data.
[0081] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0082] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the disaster analysis method for flood disasters provided by the above methods. The method includes: obtaining multiple disaster theme data and multiple waterlogging point location data from multimedia information corresponding to the disaster-stricken area; constructing a disaster theme semantic network based on the multiple disaster theme data and the co-occurrence relationship between each disaster theme data; constructing a waterlogging point semantic network based on the multiple waterlogging point location data and the co-occurrence relationship between each waterlogging point location data; and constructing a hybrid semantic network based on the co-occurrence relationship between the target disaster theme data in the disaster theme semantic network and the target waterlogging point location data in the waterlogging point semantic network; wherein the target disaster theme data belongs to multiple disaster theme data, and the target waterlogging point location data belongs to multiple waterlogging point location data.
[0083] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, is implemented to perform the disaster analysis method for flood disasters provided by the above methods. This method includes: acquiring multiple disaster theme data and multiple waterlogging point location data from multimedia information corresponding to the disaster-stricken area; constructing a disaster theme semantic network based on the multiple disaster theme data and the co-occurrence relationships between them; constructing a waterlogging point semantic network based on the multiple waterlogging point location data and the co-occurrence relationships between them; and constructing a hybrid semantic network based on the co-occurrence relationships between target disaster theme data in the disaster theme semantic network and target waterlogging point location data in the waterlogging point semantic network; wherein the target disaster theme data belongs to multiple disaster theme data, and the target waterlogging point location data belongs to multiple waterlogging point location data.
[0084] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.
[0085] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.
[0086] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A disaster analysis method for flood disasters, characterized in that, include: Multiple disaster-themed data and multiple waterlogging point location data were obtained from multimedia information corresponding to the disaster-stricken areas; A disaster theme semantic network is constructed based on the multiple disaster theme data and the co-occurrence relationships among them; a water accumulation point semantic network is constructed based on the multiple water accumulation point location data and the co-occurrence relationships among them. A hybrid semantic network is constructed based on the co-occurrence relationship between the target disaster theme data in the disaster theme semantic network and the target water accumulation point location data in the water accumulation point semantic network; wherein, the target disaster theme data belongs to the multiple disaster theme data, and the target water accumulation point location data belongs to the multiple water accumulation point location data; Based on the hybrid semantic network, the spatial correlation and disaster semantics of water accumulation points in the disaster-stricken area are analyzed according to real-time disaster data to obtain disaster analysis results.
2. The disaster analysis method for flood disasters according to claim 1, characterized in that, The step of constructing a disaster theme semantic network based on the multiple disaster theme data and the co-occurrence relationships among the disaster theme data includes: The disaster theme semantic network is constructed using each disaster theme data as node vectors, the co-occurrence relationship between any two disaster theme data as edge vectors, and the co-occurrence frequency between any two disaster theme data as edge weights.
3. The disaster analysis method for flood disasters according to claim 1, characterized in that, The step of constructing a semantic network for water accumulation points based on the multiple water accumulation point location data and the co-occurrence relationships among the water accumulation point location data includes: The semantic network of water accumulation points is constructed using the location data of each water accumulation point as node vectors, the co-occurrence relationship between any two water accumulation point location data as edge vectors, and the co-occurrence frequency between any two water accumulation point location data as edge weights.
4. The disaster analysis method for flood disasters according to claim 1, characterized in that, The step of constructing a hybrid semantic network based on the co-occurrence relationship between target disaster theme data in the disaster theme semantic network and target water accumulation point location data in the water accumulation point semantic network includes: The hybrid semantic network is constructed using the intersection nodes between the disaster-themed semantic network and the water accumulation point semantic network as node vectors, the co-occurrence relationship between any two intersection nodes as edge vectors, and the co-occurrence frequency between any two intersection nodes as edge weights.
5. The disaster analysis method for flood disasters according to claim 1, characterized in that, The multiple disaster-related data were obtained through the following steps: The multimedia information is then subjected to word segmentation, stop word removal, and keyword extraction to obtain a disaster data set. The disaster dataset is filtered by topic using a deep learning classification model to obtain multiple disaster-themed data.
6. The disaster analysis method for flood disasters according to claim 1, characterized in that, The location data of the water accumulation points is obtained through the following steps: The multimedia information is sequentially segmented, stop words are removed, and keywords are extracted to obtain the location words of the water accumulation point; The location words of the water accumulation point are processed based on the sequence prediction model to obtain the location data of the water accumulation point.
7. A disaster analysis device for flood disasters, characterized in that, include: The data acquisition module obtains multiple disaster-themed data and multiple water accumulation point location data from the multimedia information corresponding to the disaster-stricken areas; The first network construction module is used to construct a disaster theme semantic network based on the multiple disaster theme data and the co-occurrence relationship between each disaster theme data; and to construct a water accumulation point semantic network based on the multiple water accumulation point location data and the co-occurrence relationship between each water accumulation point location data. The second network construction module is used to construct a hybrid semantic network based on the co-occurrence relationship between the target disaster theme data in the disaster theme semantic network and the target water accumulation point location data in the water accumulation point semantic network; wherein, the target disaster theme data belongs to the plurality of disaster theme data, and the target water accumulation point location data belongs to the plurality of water accumulation point location data; The disaster analysis module is used to analyze the spatial correlation and disaster semantics of water accumulation points in the disaster-stricken area based on the hybrid semantic network and real-time disaster data, and obtain disaster analysis results.
8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the disaster analysis method for flood disasters as described in any one of claims 1 to 6.
9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the disaster analysis method for flood disasters as described in any one of claims 1 to 6.
10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the disaster analysis method for flood disasters as described in any one of claims 1 to 6.
