Meteorological risk pre-evaluation triggered intelligent meteorological service method and system

Abstract

The invention provides an intelligent meteorological service method and system triggered by meteorological risk pre-assessment, and relates to the technical field of meteorological disaster early warning, the method constructs a description vector library of various historical meteorological disasters, provides rich and high-quality data support for the training of a general large model, and improves the early warning efficiency of the meteorological disasters. A disaster risk pre-assessment large model is obtained through fine tuning and retraining of a general large model, so that various meteorological disaster risks can be identified and assessed more accurately. After a user inputs the multi-modal prompt data, the multi-modal prompt data comprises at least one of the following data: position data, scene image data and scene video data, and the model can quickly obtain a meteorological disaster risk assessment result, so that the method gets rid of the technical problem that a meteorological risk estimation area is limited. In addition, under the condition that the assessment result is determined to have the meteorological disaster risk, the method can also trigger the intelligent meteorological service based on the assessment result, and the intelligent meteorological service can effectively improve the pertinence of meteorological disaster early warning.

Description

Technical Field

[0001] The present invention relates to the technical field of meteorological disaster early warning, and in particular to a smart meteorological service method and system triggered by pre-assessment of meteorological risks. Background Art

[0002] In order to identify and early warn of disaster weather, the method in the prior art is an artificial intelligence method for disaster weather identification and early warning service based on video image data. Specifically, it determines the category of disaster weather through real-time video data, and discovers and optimizes the early warning description content through semantic quality control. This method has the following problems: 1) The data source is relatively single, and the risk types that can be identified from the video are limited. 2) It is mainly for areas covered by cameras, and areas without camera coverage cannot be applied, and risk early warning services cannot be provided.

[0003] In summary, the meteorological disaster early warning method in the prior art has technical problems of poor accuracy of meteorological risk estimation results and limited meteorological risk estimation areas. Summary of the Invention

[0004] The purpose of the present invention is to provide a smart meteorological service method and system triggered by pre-assessment of meteorological risks, so as to alleviate the technical problems of poor accuracy of meteorological risk estimation results and limited meteorological risk estimation areas existing in the prior meteorological disaster early warning methods.

[0005] In a first aspect, the present invention provides a smart meteorological service method triggered by pre-assessment of meteorological risks, including: constructing a description vector library of various historical meteorological disasters; wherein, the description vector of each historical meteorological disaster includes: a meteorological data description vector, a meteorological disaster risk scenario description vector, and a meteorological disaster knowledge description vector; training a general large model based on the description vector library to obtain a disaster risk pre-assessment large model; receiving multi-modal prompt data input by a user; wherein, the multi-modal prompt data includes at least one of the following: location data, scene image data, scene video data; using the disaster risk pre-assessment large model to perform meteorological disaster risk assessment on the multi-modal prompt data to obtain a meteorological disaster risk assessment result; in the case where it is determined that the meteorological disaster risk assessment result indicates the existence of meteorological disaster risks, triggering a smart meteorological service based on the meteorological disaster risk assessment result.

[0006] Optionally, construct a description vector library for multiple historical meteorological disasters, including: obtaining multimodal disaster situation data of historical meteorological disasters, meteorological disaster-related literature, meteorological-related encyclopedia data, location data corresponding to the target historical meteorological disaster, meteorological data, satellite images, and video images; where the target historical meteorological disaster represents any one of the multiple historical meteorological disasters; construct a meteorological data description vector for the target historical meteorological disaster based on the location data and meteorological data corresponding to the target historical meteorological disaster; construct a meteorological disaster risk scenario description vector for the target historical meteorological disaster based on the satellite images and video images corresponding to the target historical meteorological disaster; construct a meteorological disaster knowledge description vector library based on the multimodal disaster situation data, meteorological disaster-related literature, and meteorological-related encyclopedia data of historical meteorological disasters; calculate the similarity between the meteorological data description vector and the meteorological disaster risk scenario description vector of the target historical meteorological disaster and each vector in the meteorological disaster knowledge description vector library respectively; construct a meteorological disaster knowledge description vector for the target historical meteorological disaster based on the meteorological disaster knowledge description vectors whose similarity calculation results exceed the preset threshold; construct a description vector library for multiple historical meteorological disasters based on the description vectors of multiple historical meteorological disasters.

[0007] Optionally, construct a meteorological data description vector for the target historical meteorological disaster based on the location data and meteorological data corresponding to the target historical meteorological disaster, including: obtaining the observation data of meteorological stations around the location and satellite radar meteorological observation data based on the location data; using a preset high-impact weather diagnosis algorithm model library to process the meteorological data, the observation data of surrounding meteorological stations, and the satellite radar meteorological observation data to obtain a high-impact weather diagnosis result; construct a meteorological data description vector for the target historical meteorological disaster based on the meteorological data and the high-impact weather diagnosis result; where the meteorological data description vector includes: basic meteorological elements, high-impact weather elements, and high-impact weather diagnosis elements.

[0008] Optionally, construct a meteorological disaster risk scenario description vector for the target historical meteorological disaster based on the satellite images and video images corresponding to the target historical meteorological disaster, including: using a preset meteorological disaster scenario recognition model to process the satellite images and video images of the target historical meteorological disaster to obtain a scenario recognition result of the target historical meteorological disaster; where the scenario recognition result includes: disaster risk type, affected object, and damage level; using a general large model to process the scenario recognition result to output a meteorological disaster risk scenario description vector for the target historical meteorological disaster; where the meteorological disaster risk scenario description vector is a multi-dimensional vector that comprehensively represents the risk scenario characteristics and information of the target historical meteorological disaster.

[0009] Optionally, based on the multi-modal disaster situation data of historical meteorological disasters, meteorological disaster-related literature, and meteorological-related encyclopedia data, construct a meteorological disaster knowledge description vector library, including: performing semantic segmentation and word embedding processing on the multi-modal disaster situation data of historical meteorological disasters and meteorological disaster-related literature to obtain an initial meteorological disaster knowledge description vector library; constructing a meteorological knowledge graph based on meteorological-related encyclopedia data; using graph-based knowledge retrieval enhancement technology to process the meteorological knowledge graph to obtain supplementary meteorological disaster knowledge description vectors; and constructing a meteorological disaster knowledge description vector library based on the initial meteorological disaster knowledge description vector library and the supplementary meteorological disaster knowledge description vectors.

[0010] Optionally, trigger intelligent meteorological services based on the meteorological disaster risk assessment results, including: obtaining a fuzzy matching rule library; wherein, the fuzzy matching rule library defines the corresponding relationships between various meteorological disaster risk assessment results and intelligent meteorological services; wherein, the intelligent meteorological services include: meteorological disaster warning information and meteorological disaster avoidance strategies in various display forms; and querying in the fuzzy matching rule library using the meteorological disaster risk assessment results to determine the corresponding intelligent meteorological services.

[0011] Optionally, after obtaining the meteorological disaster risk assessment results, it further includes: displaying the evaluation and reasoning process by which the disaster risk pre-assessment large model determines the meteorological disaster risk assessment results.

[0012] In a second aspect, the present invention provides an intelligent meteorological service system triggered by meteorological risk pre-assessment, including: a construction module for constructing a description vector library of various historical meteorological disasters; wherein, the description vector of each historical meteorological disaster includes: a meteorological data description vector, a meteorological disaster risk scenario description vector, and a meteorological disaster knowledge description vector; a training module for training a general large model based on the description vector library to obtain a disaster risk pre-assessment large model; a receiving module for receiving multi-modal prompt data input by a user; wherein, the multi-modal prompt data includes at least one of the following: location data, scene image data, scene video data; an evaluation module for using the disaster risk pre-assessment large model to perform meteorological disaster risk assessment on the multi-modal prompt data to obtain a meteorological disaster risk assessment result; and a triggering module for triggering intelligent meteorological services based on the meteorological disaster risk assessment results when it is determined that the meteorological disaster risk assessment result indicates the existence of a meteorological disaster risk.

[0013] In a third aspect, the present invention provides an electronic device, including a memory and a processor, where a computer program that can run on the processor is stored on the memory, and when the processor executes the computer program, it implements the method for intelligent meteorological services triggered by meteorological risk pre-assessment according to any one of the foregoing embodiments.

[0014] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions, which, when executed by a processor, implement the intelligent meteorological service method triggered by meteorological risk pre-assessment described in any one of the foregoing embodiments.

[0015] The present invention provides an intelligent meteorological service method triggered by meteorological risk pre-assessment. By constructing a description vector library of various historical meteorological disasters, a comprehensive and in-depth data set is obtained, providing rich and high-quality data support for the training of general large models. Utilizing the powerful learning ability and generalization ability of general large models, a disaster risk pre-assessment large model is obtained through fine-tuning and retraining, enabling it to more accurately identify and evaluate various meteorological disaster risks. On this basis, after the user inputs multi-modal prompt data, where the multi-modal prompt data includes at least one of the following: location data, scene image data, scene video data, the disaster risk pre-assessment large model can quickly obtain the meteorological disaster risk assessment result. Thus, it can be seen that this method solves the technical problem of limited meteorological risk estimation area. Further, when it is determined that the assessment result indicates the existence of meteorological disaster risk, this method can also trigger intelligent meteorological services based on the assessment result, and the intelligent meteorological services can effectively improve the pertinence of meteorological disaster warnings. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the specific embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0017] Figure 1 It is a flowchart of an intelligent meteorological service method triggered by meteorological risk pre-assessment provided by an embodiment of the present invention; Figure 2 It is a flowchart of constructing a description vector library of various historical meteorological disasters provided by an embodiment of the present invention; Figure 3 It is a functional module diagram of an intelligent meteorological service system triggered by meteorological risk pre-assessment provided by an embodiment of the present invention; Figure 4 It is a schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Components of the embodiments of the present invention usually described and illustrated in the drawings here can be arranged and designed in various different configurations.

[0019] Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed present invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0020] The following will describe in detail some embodiments of the present invention with reference to the accompanying drawings. Without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0021] Embodiment 1 Figure 1 is a flowchart of a smart meteorological service method triggered by pre-assessment of meteorological risks provided by an embodiment of the present invention. As Figure 1 shown, the method specifically includes the following steps: Step S102, construct a description vector library of various historical meteorological disasters.

[0022] Among them, the description vector of each historical meteorological disaster includes: a meteorological data description vector, a meteorological disaster risk scenario description vector, and a meteorological disaster knowledge description vector.

[0023] Step S104, train a general large model based on the description vector library to obtain a disaster risk pre-assessment large model.

[0024] To improve the accuracy of meteorological disaster risk assessment, the embodiments of the present invention collect a large amount of data related to meteorological disasters to construct a description vector library of various historical meteorological disasters, and then use this vector library to train the general large model in the prior art. By fine-tuning the model parameters, based on the powerful learning ability and generalization ability of the general large model, a disaster risk pre-assessment large model that can accurately assess meteorological disaster risks can be obtained. The embodiments of the present invention do not specifically specify the general large model, and users can select according to actual needs. For example, deepseek, Tongyi Qianwen, etc.

[0025] In the embodiments of the present invention, the description vector library of multiple historical meteorological disasters is a database composed of description vectors of multiple historical meteorological disasters. The description vector of each historical meteorological disaster includes: a meteorological data description vector, a meteorological disaster risk scenario description vector, and a meteorological disaster knowledge description vector. Among them, the meteorological data description vector is a structured feature set, which consists of three parts: the observed values of basic meteorological elements (temperature, pressure, humidity, wind, precipitation, etc.), the marks of high-impact weather phenomena (such as thunderstorms, hailstorms, heavy precipitation, strong winds, lightning, low visibility, typhoons, etc.), and the high-impact weather diagnosis indexes (such as CAPE index, hail index, etc.). It aims to comprehensively characterize the weather conditions and their potential extremities. By quantifying the basic meteorological data, high-impact weather events, and professional diagnosis parameters, it provides multi-dimensional information support for meteorological analysis, forecasting and early warning, and impact assessment.

[0026] The meteorological disaster risk scenario description vector is a structured feature expression formed by extracting the disaster risk type, affected objects, and damage levels by a general large model based on satellite image and video image analysis. It is used to quantitatively describe the impact range and severity of meteorological disasters, and can support meteorological disaster risk assessment and emergency response decision-making.

[0027] The meteorological disaster knowledge description vector is a multi-dimensional feature (such as disaster type, spatio-temporal distribution, associated keywords, affected objects, causal mechanism, etc.) extracted based on multi-modal disaster situation data (text, images, etc.), literature, and encyclopedia data of historical meteorological disasters. Through semantic similarity calculation or feature matching (such as cosine similarity, topic model weight), a vector set that highly matches the meteorological disaster in terms of disaster characteristics, evolution laws, or social impact dimensions is selected, which is used to support disaster-related knowledge mining and risk reasoning.

[0028] Based on the explanations of each description vector in the description vector library above, the description vector library of multiple historical meteorological disasters is a comprehensive and in-depth data set. Therefore, training the general large model based on this data set can ensure the accuracy of the disaster risk pre-assessment model in identifying multiple meteorological disaster risks.

[0029] Step S106: Receive the multi-modal prompt data input by the user.

[0030] Among them, the multi-modal prompt data includes at least one of the following: location data, scene image data, and scene video data.

[0031] Step S108: Use the disaster risk pre-assessment model to conduct meteorological disaster risk assessment on the multi-modal prompt data, and obtain the meteorological disaster risk assessment result.

[0032] In the embodiments of the present invention, the disaster risk pre-assessment large model supports users to input various types of prompt words, that is, the model supports the input of multi-modal prompt data. If the multi-modal prompt data is text-based location data, the disaster risk pre-assessment large model is used to perform meteorological disaster risk assessment on the multi-modal prompt data, including: First, key features are extracted from the location data through natural language processing technology; among them, the key features include: geographical location coordinates and administrative divisions. It should be noted that the above location information can be a definite location data or route information, including: starting location information and ending location information.

[0033] Next, meteorological data with the highest correlation with the above key features is queried through a preset meteorological data query channel to construct a target meteorological data description vector based on the meteorological data; among them, the preset meteorological data query channel can be an official meteorological agency platform (such as a meteorological bureau), a mobile application, a smart device, etc. The meteorological data with the highest correlation with the key features can be understood as the meteorological data of the location data input by the user and its surrounding areas. The construction of the target meteorological data description vector refers to the structure and definition of the meteorological data description vector in the above text, and the target meteorological data description vector is a description vector constructed according to the currently queried meteorological data.

[0034] Then, the similarity between the target meteorological data description vector and each vector in the meteorological disaster knowledge description vector library is calculated, and the meteorological disaster knowledge description vector whose similarity calculation result exceeds the preset threshold is used as the first target meteorological disaster knowledge description vector; among them, the meteorological disaster knowledge description vector library is a vector library composed of multi-dimensional features extracted from historical meteorological disaster multi-modal disaster situation data, literature, and encyclopedia data.

[0035] Finally, in the case where the meteorological disaster risk scenario description vector is missing, the disaster risk pre-assessment large model performs disaster risk assessment based on the target meteorological data description vector and the first target meteorological disaster knowledge description vector to obtain an assessment result.

[0036] If the multi-modal prompt data is scenario image data and / or scenario video data, the disaster risk pre-assessment large model is used to perform meteorological disaster risk assessment on the multi-modal prompt data, including: First, use a preset meteorological disaster scenario recognition model to process the scenario image data and / or scenario video data to obtain the corresponding scenario recognition results. Among them, the scenario recognition results include: disaster risk type, affected object, and damage level. The construction process of the preset meteorological disaster scenario recognition model can refer to the following process: collect satellite remote sensing images and video images of various meteorological disasters, and label the meteorological disaster risk type, affected object, and damage level corresponding to each image to form a meteorological disaster scenario dataset. Based on this dataset, fine-tune the LingMou large model or train models such as YOLO V10 to obtain the above-mentioned preset meteorological disaster scenario recognition model.

[0037] Then, perform a structured feature expression on the above scenario recognition results to construct a target meteorological disaster risk scenario description vector. The construction of the target meteorological disaster risk scenario description vector refers to the structure and definition of the meteorological disaster risk scenario description vector in the above text. The target meteorological disaster risk scenario description vector is a description vector constructed based on the current scenario recognition results.

[0038] Next, calculate the similarity between the target meteorological disaster risk scenario description vector and each vector in the meteorological disaster knowledge description vector library, and use the meteorological disaster knowledge description vector whose similarity calculation result exceeds the preset threshold as the second target meteorological disaster knowledge description vector.

[0039] Finally, in the case where the meteorological data description vector is missing, the disaster risk pre-assessment large model performs a disaster risk assessment based on the target meteorological disaster risk scenario description vector and the second target meteorological disaster knowledge description vector to obtain an assessment result.

[0040] If the multi-modal prompt data includes: location data, scenario image data, and scenario video data, then the disaster risk pre-assessment large model can perform a disaster risk assessment based on the target meteorological data description vector, the target meteorological disaster risk scenario description vector, and the target meteorological disaster knowledge description vector to obtain a more accurate assessment result. In the embodiments of the present invention, the assessment result includes one of the following: there is a certain meteorological disaster risk, and there is no meteorological disaster risk.

[0041] Step S110, in the case where it is determined that the meteorological disaster risk assessment result is that there is a meteorological disaster risk, trigger the intelligent meteorological service based on the meteorological disaster risk assessment result.

[0042] Specifically, if the meteorological disaster risk assessment result indicates the absence of meteorological disaster risk, the disaster risk pre-assessment large model can directly feedback this assessment result to the user. However, if the meteorological disaster risk assessment result indicates the presence of meteorological disaster risk, in order to prevent the user from ignoring this meteorological disaster risk, the embodiment of the present invention will further trigger intelligent meteorological services based on the meteorological disaster risk assessment result, aiming to achieve the precision, automation, and proactivity of disaster warning and emergency services.

[0043] For example, intelligent meteorological services automatically match the emergency plan library based on meteorological disaster risk scenarios (such as evacuation route planning, measures for suspending work and classes), and output guiding texts through natural language generation technology; use user portraits (location, occupation, scenario) to achieve hierarchical push (such as sending typhoon shelter reminders to fishermen, pushing road control plans to the transportation department), etc. The functions of the above intelligent meteorological services can be achieved by having a reinforcement learning model learn a large number of push strategies, and then the reinforcement learning model triggers different push strategies according to different meteorological disaster risk assessment results. This reinforcement learning model can be integrated into the disaster risk pre-assessment large model or exist independently. Intelligent meteorological services can also be to use generative artificial intelligence based on large language models to generate meteorological disaster service information according to meteorological disaster risk assessment results.

[0044] The embodiment of the present invention provides a method for intelligent meteorological services triggered by meteorological risk pre-assessment. By constructing a description vector library of various historical meteorological disasters, a comprehensive and in-depth data set is obtained, providing rich and high-quality data support for the training of the general large model. Utilizing the powerful learning ability and generalization ability of the general large model, a disaster risk pre-assessment large model is obtained through fine-tuning and retraining, enabling it to more accurately identify and evaluate various meteorological disaster risks. On this basis, after the user inputs multi-modal prompt data, where the multi-modal prompt data includes at least one of the following: location data, scenario image data, scenario video data, the disaster risk pre-assessment large model can quickly obtain the meteorological disaster risk assessment result. It can be seen that this method solves the technical problem of limited meteorological risk estimation area. Further, in the case where the assessment result indicates the presence of meteorological disaster risk, this method can also trigger intelligent meteorological services based on the assessment result, and the intelligent meteorological services can effectively improve the pertinence of meteorological disaster warnings.

[0045] In an optional implementation manner, as Figure 2 shown, the above step S102, constructing a description vector library of various historical meteorological disasters, specifically includes the following steps: Step S1021, obtain multi-modal disaster situation data of historical meteorological disasters, meteorological disaster-related literature, meteorological-related encyclopedia data, location data corresponding to the target historical meteorological disaster, meteorological data, satellite images, and video images.

[0046] Among them, the target historical meteorological disaster represents any one of a variety of historical meteorological disasters.

[0047] After a certain meteorological disaster occurs, relevant information about the meteorological disaster will be spread on the Internet. Therefore, multi-modal disaster situation data of historical meteorological disasters can be obtained through channels such as the APIs or RSS subscriptions of various news websites (such as Xinhua News Agency), and crawler programs of social media platforms (such as Weibo, Twitter, WeChat official accounts). Meteorological disaster-related literature can be obtained through academic databases (such as CNKI, PubMed, Web of Science). Meteorological-related encyclopedia data can be obtained through encyclopedias (such as Wikipedia, Baidu Encyclopedia) and the official websites of meteorological agencies (such as the China Meteorological Administration, NOAA).

[0048] After the target historical meteorological disaster occurs, in order to construct its description vector in the embodiments of the present invention, it is necessary to obtain the location data, meteorological data, satellite images, and video images corresponding to the meteorological disaster. Among them, the meteorological data includes at least one of the following data: ERA5 meteorological reanalysis data, CLDAS (China Meteorological Administration Data Distribution Service System) historical weather data, real-time weather data, numerical model forecast data (such as European Centre for Medium-Range Weather Forecasts (EC) numerical model forecasts, Central Meteorological Observatory intelligent grid weather forecasts, etc.).

[0049] Step S1022: Based on the location data and meteorological data corresponding to the target historical meteorological disaster, construct a meteorological data description vector of the target historical meteorological disaster.

[0050] Specifically, according to the location data where the target historical meteorological disaster occurs, the meteorological data around it can be obtained. Combining the meteorological data at this location obtained in the previous step, through high-impact weather analysis of all the data, a meteorological data description vector of the target historical meteorological disaster can be formed.

[0051] Step S1023: Based on the satellite images and video images corresponding to the target historical meteorological disaster, construct a meteorological disaster risk scenario description vector of the target historical meteorological disaster.

[0052] After obtaining the satellite images and video images corresponding to the target historical meteorological disaster, data such as disaster risk types, affected objects, and damage levels can be identified from them, so as to construct a meteorological disaster risk scenario description vector of the target historical meteorological disaster based on the above data.

[0053] Step S1024: Based on the multi-modal disaster situation data of historical meteorological disasters, meteorological disaster-related literature, and meteorological-related encyclopedia data, construct a meteorological disaster knowledge description vector library.

[0054] In an alternative implementation, the steps of constructing a meteorological disaster knowledge description vector library can be divided into three parts: multi-source data integration, semantic feature extraction, and vectorization modeling. First, by crawling multi-modal disaster situation data, literature, and encyclopedia data of historical meteorological disasters, cleaning and standardization are carried out (such as unifying disaster naming and aligning spatio-temporal tags); secondly, natural language processing technologies (such as BERT and knowledge graph entity recognition) are used to extract multi-dimensional features, including: disaster type, spatio-temporal pattern, disaster-causing factors, affected body attributes (such as population density, economic vulnerability), causal chain (such as typhoon → heavy rain → mountain flood), keywords in social public information flow (such as "power outage", "rescue"), etc.; finally, unstructured text is converted into high-dimensional vectors through feature fusion (such as weighted splicing or graph embedding), and a vector library is constructed based on similarity measurement (such as contrast learning).

[0055] Step S1025, calculate the similarity between the meteorological data description vector and the meteorological disaster risk scenario description vector of the target historical meteorological disaster and each vector in the meteorological disaster knowledge description vector library respectively.

[0056] Step S1026, construct the meteorological disaster knowledge description vector of the target historical meteorological disaster based on the meteorological disaster knowledge description vectors whose similarity calculation results exceed the preset threshold.

[0057] That is to say, not only the similarity between the meteorological data description vector of the target historical meteorological disaster and each vector in the meteorological disaster knowledge description vector library needs to be calculated, but also the similarity between the meteorological disaster risk scenario description vector of the target historical meteorological disaster and each vector in the meteorological disaster knowledge description vector library needs to be calculated. In this embodiment, the preset threshold of similarity is used as the screening condition to screen vectors from the meteorological disaster knowledge description vector library to construct the meteorological disaster knowledge description vector of the target historical meteorological disaster. In another implementation, the above two groups of similarity calculation results can also be sorted in descending order first, and then the top N vectors are intercepted as the meteorological disaster knowledge description vector of the target historical meteorological disaster. The user can set the value of N according to actual needs.

[0058] Step S1027, construct a description vector library of multiple historical meteorological disasters based on the description vectors of multiple historical meteorological disasters.

[0059] Referring to the determination method of the description vector of the target historical meteorological disaster in the above text, the position data, meteorological data, satellite images, and video images corresponding to each historical meteorological disaster are processed, and the description vector of each historical meteorological disaster can be obtained, and then a description vector library of multiple historical meteorological disasters is constructed.

[0060] In an alternative embodiment, in step S1022, based on the location data and meteorological data corresponding to the target historical meteorological disaster, a meteorological data description vector of the target historical meteorological disaster is constructed, which specifically includes the following steps: Step S10221, based on the location data, obtain the observation data of meteorological stations around the location and satellite radar meteorological observation data.

[0061] Step S10222, use the preset high-impact weather diagnostic algorithm model library to process the meteorological data, the observation data of meteorological stations around, and the satellite radar meteorological observation data to obtain a high-impact weather diagnostic result.

[0062] Specifically, high-impact weather refers to weather phenomena or events that may have a significant impact on human activities, economy, society, or the natural environment. These weather events usually have a high intensity, frequency, or duration, and may cause serious consequences such as natural disasters, infrastructure damage, casualties, and economic losses. In the embodiments of the present invention, each type of high-impact weather corresponds to a high-impact weather diagnostic algorithm model. The target high-impact weather diagnostic algorithm model is a model for diagnosing whether there is target high-impact weather based on multi-source meteorological data. Target high-impact weather represents any type of high-impact weather, such as extreme weather like thunderstorms, hailstorms, heavy precipitation, strong winds, lightning, low visibility, typhoons, etc.

[0063] Therefore, in order to accurately identify various high-impact weathers, the embodiments of the present invention need to use various high-impact weather diagnostic algorithm models (i.e., the preset high-impact weather diagnostic algorithm model library) to process the meteorological data corresponding to the target historical meteorological disaster, the observation data of meteorological stations around, and the satellite radar meteorological observation data that have been obtained, so as to obtain the results of whether various high-impact weathers exist. In the embodiments of the present invention, the above high-impact weather diagnostic results are used to characterize the cause of the target historical meteorological disaster.

[0064] Step S10223, based on the meteorological data and the high-impact weather diagnostic result, construct a meteorological data description vector of the target historical meteorological disaster; wherein, the meteorological data description vector includes: basic meteorological elements, high-impact weather elements, and high-impact weather diagnostic elements.

[0065] Optionally, input the meteorological data and the high-impact weather diagnostic result into an existing general large model (such as DeepSeek), and through prompt words, require the large model to output a fused meteorological data description vector, and require this vector to be composed of the following parts: basic meteorological elements (temperature, pressure, humidity, wind, precipitation, etc.), high-impact weather elements (for example, thunderstorms, hailstorms, heavy precipitation, strong winds, lightning, low visibility, typhoons, etc.), high-impact weather diagnostic elements (such as CAPE index, hail index, etc.).

[0066] In an alternative implementation, in step S1023 above, based on the satellite images and video images corresponding to the target historical meteorological disaster, a meteorological disaster risk scenario description vector of the target historical meteorological disaster is constructed, which specifically includes the following steps: Step S10231: Use a preset meteorological disaster scenario recognition model to process the satellite images and video images of the target historical meteorological disaster to obtain the scenario recognition result of the target historical meteorological disaster; among them, the scenario recognition result includes: disaster risk type, affected object, and damage level.

[0067] As can be seen from the above description, the preset meteorological disaster scenario recognition model is obtained by fine-tuning and training a large model in the prior art with a large amount of labeled sample data. After obtaining the satellite images and video images corresponding to the target historical meteorological disaster, input them into the preset meteorological disaster scenario recognition model, and the output of the model is the scenario recognition result of the target historical meteorological disaster.

[0068] Step S10232: Use a general large model to process the scenario recognition result to output a meteorological disaster risk scenario description vector of the target historical meteorological disaster; among them, the meteorological disaster risk scenario description vector is a multi-dimensional vector that comprehensively represents the risk scenario characteristics and information of the target historical meteorological disaster.

[0069] In this step, the general large model is mainly used to process the scenario recognition result of the target historical meteorological disaster according to the preset data standardization processing method, and construct a multi-dimensional vector after forming a structured feature expression.

[0070] A method for constructing a meteorological disaster knowledge description vector library has been described above. Now, another method for constructing a meteorological disaster knowledge description vector library is introduced. In an alternative implementation, in step S1024 above, based on the multi-modal disaster situation data of historical meteorological disasters, meteorological disaster-related literature, and meteorological-related encyclopedia data, a meteorological disaster knowledge description vector library is constructed, which specifically includes the following steps: Step S10241: Perform semantic segmentation and word embedding processing on the multi-modal disaster situation data of historical meteorological disasters and meteorological disaster-related literature to obtain an initial meteorological disaster knowledge description vector library.

[0071] Step S10242: Construct a meteorological knowledge graph based on meteorological-related encyclopedia data.

[0072] Step S10243: Use graph-based knowledge retrieval enhancement technology to process the meteorological knowledge graph to obtain supplementary meteorological disaster knowledge description vectors.

[0073] Step S10244: Construct a meteorological disaster knowledge description vector library based on the initial meteorological disaster knowledge description vector library and the supplementary meteorological disaster knowledge description vectors.

[0074] That is to say, in this implementation, the processing method of meteorological-related encyclopedia data is different from the processing methods of multi-modal disaster situation data of historical meteorological disasters and meteorological disaster-related literature. First, an initial meteorological disaster knowledge description vector library is constructed based on the multi-modal disaster situation data of historical meteorological disasters and meteorological disaster-related literature. Then, a meteorological knowledge graph is constructed based on the meteorological-related encyclopedia data. Furthermore, based on the knowledge retrieval enhancement technology of the graph, such as methods similar to GRAPH_RAG, the meteorological knowledge graph is analyzed and processed to further enrich the initial meteorological disaster knowledge description vector library based on the processing results, and finally a meteorological disaster knowledge description vector library is obtained.

[0075] In an optional implementation, step S110, triggering intelligent meteorological services based on the meteorological disaster risk assessment results, specifically includes the following steps: Step S1101, obtain a fuzzy matching rule library; wherein, the fuzzy matching rule library defines the corresponding relationships between various meteorological disaster risk assessment results and intelligent meteorological services; wherein, the intelligent meteorological services include: meteorological disaster warning information and meteorological disaster avoidance strategies in various display forms.

[0076] Step S1102, query in the fuzzy matching rule library using the meteorological disaster risk assessment results to determine the corresponding intelligent meteorological services.

[0077] The embodiments of the present invention do not specifically limit the form of the intelligent meteorological services, and their display forms can be in forms such as text, voice, or decision-making service special materials. By maintaining the fuzzy matching rule library, the corresponding intelligent meteorological services can be quickly matched after the meteorological disaster risk assessment results are determined.

[0078] For the convenience of understanding, the intelligent meteorological services are illustrated by examples below: Suppose a user inputs "The current location is XX Beach, and there are abnormal weather changes" as a prompt at the seaside. The disaster risk pre-assessment large model determines through assessment that there is a typhoon risk at this beach, and then typhoon warning information and avoidance strategies will be pushed to the user.

[0079] Suppose a user uploads a video of a heavy rain scene in the mountains. The disaster risk pre-assessment large model extracts features through image recognition and combines the mountain location information to assess that there is a flash flood risk at this location, and then an emergency evacuation strategy and rescue contact information will be sent to the user.

[0080] Suppose the user is located in a mountainous area and is hiking. The user inputs the current location, hiking route, and photos taken along the way into the disaster risk pre-assessment large model through the model interaction interface. Then, the disaster risk pre-assessment large model can conduct a meteorological risk assessment based on the information input by the user. If it is predicted that heavy rainfall may occur in the area within the next few hours and there is a risk of flash floods, an early warning message will be immediately sent to the user to enable the user to evacuate to a safe area as soon as possible. The emergency response plan can also be activated to notify the local rescue department to make rescue preparations.

[0081] In an optional implementation manner, after obtaining the meteorological disaster risk assessment result, the method of the present invention further includes the following: displaying the evaluation reasoning process by which the disaster risk pre-assessment large model determines the meteorological disaster risk assessment result.

[0082] Specifically, after the disaster risk pre-assessment large model outputs the meteorological disaster risk assessment result, in order to enhance the credibility of the above assessment result, the evaluation reasoning process (i.e., the disaster risk assessment chain) by which the disaster risk pre-assessment large model determines the meteorological disaster risk assessment result can be further displayed, thereby enhancing the user's attention to the meteorological disaster early warning information and meteorological disaster avoidance strategies.

[0083] In summary, the intelligent meteorological service method triggered by the meteorological risk pre-assessment provided by the embodiments of the present invention can quickly and accurately identify meteorological disasters based on the location data and / or scenario data provided by the user, and can also realize the personalized, intelligent, and automated generation of disaster service information (meteorological disaster early warning information and meteorological disaster avoidance strategies). Compared with the prior art, the embodiments of the present invention can effectively improve the efficiency of meteorological disaster risk services and provide personalized meteorological services.

[0084] Embodiment 2 The embodiments of the present invention further provide an intelligent meteorological service system triggered by meteorological risk pre-assessment. This system is mainly used to execute the intelligent meteorological service method triggered by meteorological risk pre-assessment provided in the above Embodiment 1. The following is a specific introduction to the intelligent meteorological service system triggered by meteorological risk pre-assessment provided by the embodiments of the present invention.

[0085] Figure 3 It is a functional module diagram of an intelligent meteorological service system triggered by meteorological risk pre-assessment provided by the embodiments of the present invention. As Figure 3 shown, this system mainly includes: a construction module 10, a training module 20, a receiving module 30, an evaluation module 40, and a triggering module 50, where: The construction module 10 is used to construct a description vector library of various historical meteorological disasters; among them, the description vector of each historical meteorological disaster includes: a meteorological data description vector, a meteorological disaster risk scenario description vector, and a meteorological disaster knowledge description vector.

[0086] A training module 20 for training a general large model based on a description vector library to obtain a large model for pre-assessing disaster risks.

[0087] A receiving module 30 for receiving multi-modal prompt data input by a user; wherein, the multi-modal prompt data includes at least one of the following: location data, scene image data, and scene video data.

[0088] An evaluation module 40 for using the large model for pre-assessing disaster risks to perform meteorological disaster risk assessment on the multi-modal prompt data to obtain a meteorological disaster risk assessment result.

[0089] A triggering module 50 for triggering intelligent meteorological services based on the meteorological disaster risk assessment result when it is determined that the meteorological disaster risk assessment result indicates the existence of meteorological disaster risks.

[0090] An embodiment of the present invention provides an intelligent meteorological service system triggered by meteorological risk pre-assessment. By constructing a description vector library of various historical meteorological disasters, a comprehensive and in-depth data set is obtained, providing rich and high-quality data support for the training of the general large model. Utilizing the powerful learning ability and generalization ability of the general large model, a large model for pre-assessing disaster risks is obtained through fine-tuning and retraining, enabling it to more accurately identify and evaluate various meteorological disaster risks. On this basis, after the user inputs multi-modal prompt data, wherein the multi-modal prompt data includes at least one of the following: location data, scene image data, and scene video data, the large model for pre-assessing disaster risks can quickly obtain a meteorological disaster risk assessment result. Thus, it can be seen that this system overcomes the technical problem of limited meteorological risk prediction areas. Further, when it is determined that the assessment result indicates the existence of meteorological disaster risks, this system can also trigger intelligent meteorological services based on the assessment result, and the intelligent meteorological services can effectively improve the pertinence of meteorological disaster warnings.

[0091] Optionally, the construction module 10 includes: An acquisition unit for acquiring multi-modal disaster situation data of historical meteorological disasters, meteorological disaster-related literature, meteorological-related encyclopedia data, location data corresponding to the target historical meteorological disaster, meteorological data, satellite images, and video images; wherein, the target historical meteorological disaster represents any one of the various historical meteorological disasters.

[0092] A first construction unit for constructing a meteorological data description vector of the target historical meteorological disaster based on the location data and meteorological data corresponding to the target historical meteorological disaster.

[0093] A second construction unit for constructing a meteorological disaster risk scenario description vector of the target historical meteorological disaster based on the satellite images and video images corresponding to the target historical meteorological disaster.

[0094] A third construction unit, configured to construct a meteorological disaster knowledge description vector library based on multi-modal disaster situation data of historical meteorological disasters, meteorological disaster-related literature, and meteorological-related encyclopedia data.

[0095] A calculation unit, configured to calculate the similarity between the meteorological data description vector and the meteorological disaster risk scenario description vector of the target historical meteorological disaster and each vector in the meteorological disaster knowledge description vector library respectively.

[0096] A fourth construction unit, configured to construct a meteorological disaster knowledge description vector of the target historical meteorological disaster based on the meteorological disaster knowledge description vectors whose similarity calculation results exceed a preset threshold.

[0097] A fifth construction unit, configured to construct a description vector library of multiple historical meteorological disasters based on the description vectors of multiple historical meteorological disasters.

[0098] Optionally, the first construction unit is specifically configured to: Obtain the observation data of meteorological stations around the location and satellite radar meteorological observation data based on the location data.

[0099] Process the meteorological data, the observation data of surrounding meteorological stations, and the satellite radar meteorological observation data by using a preset high-impact weather diagnosis algorithm model library to obtain a high-impact weather diagnosis result.

[0100] Construct a meteorological data description vector of the target historical meteorological disaster based on the meteorological data and the high-impact weather diagnosis result; wherein, the meteorological data description vector includes: basic meteorological elements, high-impact weather elements, and high-impact weather diagnosis elements.

[0101] Optionally, the second construction unit is specifically configured to: Process the satellite images and video images of the target historical meteorological disaster by using a preset meteorological disaster scenario recognition model to obtain a scenario recognition result of the target historical meteorological disaster; wherein, the scenario recognition result includes: disaster risk type, affected objects, and damage levels.

[0102] Process the scenario recognition result by using a general large model to output a meteorological disaster risk scenario description vector of the target historical meteorological disaster; wherein, the meteorological disaster risk scenario description vector is a multi-dimensional vector comprehensively representing the risk scenario characteristics and information of the target historical meteorological disaster.

[0103] Optionally, the third construction unit is specifically configured to: Perform semantic segmentation and word embedding processing on the multi-modal disaster situation data of historical meteorological disasters and meteorological disaster-related literature to obtain an initial meteorological disaster knowledge description vector library.

[0104] Construct a meteorological knowledge graph based on meteorological-related encyclopedia data.

[0105] The meteorological knowledge graph is processed using graph-based knowledge retrieval enhancement technology to obtain supplementary meteorological disaster knowledge description vectors.

[0106] Based on the initial meteorological disaster knowledge description vector library and the supplementary meteorological disaster knowledge description vectors, a meteorological disaster knowledge description vector library is constructed.

[0107] Optionally, the trigger module 50 is specifically configured to: Obtain a fuzzy matching rule library; wherein, the fuzzy matching rule library defines the corresponding relationships between various meteorological disaster risk assessment results and intelligent meteorological services; wherein, the intelligent meteorological services include: meteorological disaster warning information and meteorological disaster avoidance strategies in various display forms.

[0108] Use the meteorological disaster risk assessment result to query in the fuzzy matching rule library to determine the corresponding intelligent meteorological service.

[0109] Optionally, the system is also used for: Display the evaluation and reasoning process of the disaster risk pre-assessment large model to determine the meteorological disaster risk assessment result.

[0110] Embodiment III Refer to Figure 4 , this embodiment of the present invention provides an electronic device, which includes: a processor 60, a memory 61, a bus 62, and a communication interface 63, and the processor 60, the communication interface 63, and the memory 61 are connected through the bus 62; the processor 60 is used to execute an executable module stored in the memory 61, such as a computer program.

[0111] Among them, the memory 61 may include a high-speed random access memory (RAM, Random Access Memory), and may also include a non-volatile memory, such as at least one disk memory. Through at least one communication interface 63 (which can be wired or wireless), a communication connection is realized between this system network element and at least one other network element, and the Internet, wide area network, local area network, metropolitan area network, etc. can be used.

[0112] The bus 62 can be an ISA bus, a PCI bus, an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of representation, Figure 4 only a two-way arrow is used in [description], but it does not mean that there is only one bus or one type of bus.

[0113] Among them, the memory 61 is used to store a program. After receiving an execution instruction, the processor 60 executes the program. The method executed by the device defined by the process disclosed in any embodiment of the foregoing embodiments of the present invention can be applied to the processor 60 or implemented by the processor 60.

[0114] The processor 60 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit in the hardware of the processor 60 or by instructions in software form. The above-mentioned processor 60 may be a general-purpose processor, including a central processing unit (CPU for short), a network processor (NP for short), etc.; it may also be a digital signal processor (DSP for short), an application specific integrated circuit (ASIC for short), a field-programmable gate array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present invention can be directly embodied as being executed and completed by a hardware decoding processor, or executed and completed by a combination of hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory 61, and the processor 60 reads the information in the memory 61 and combines its hardware to complete the steps of the above method.

[0115] A computer program product of a smart meteorological service method and system triggered by meteorological risk pre-assessment provided by an embodiment of the present invention includes a computer-readable storage medium storing non-volatile program code executable by a processor. The instructions included in the program code can be used to execute the method described in the foregoing method embodiments. For specific implementation, reference can be made to the method embodiments and will not be elaborated here.

[0116] In addition, in each embodiment of the present invention, each functional unit may be integrated in a processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit.

[0117] When the above-mentioned function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile computer-readable storage medium executable by a processor. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this 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 for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0118] It should be noted that: similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0119] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the product of this invention is usually placed during use. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation to the present invention. In addition, the terms "first", "second", "third", etc. are only used for descriptive distinction and cannot be understood as indicating or implying relative importance.

[0120] In addition, the terms "horizontal", "vertical", "overhanging", etc. do not mean that the components are required to be absolutely horizontal or overhanging, but can be slightly inclined. For example, "horizontal" only means that its direction is more horizontal relative to "vertical", and does not mean that the structure must be completely horizontal, but can be slightly inclined.

[0121] In the description of the present invention, it should also be noted that unless otherwise clearly specified and defined, the terms "set", "install", "connect", "couple" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, an electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, and can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0122] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A smart meteorological service method triggered by meteorological risk pre-assessment, characterized in that: include: Constructing a description vector library of various historical meteorological disasters; wherein each description vector of the historical meteorological disasters includes: a meteorological data description vector, a meteorological disaster risk scenario description vector and a meteorological disaster knowledge description vector; Training the general large model based on the description vector library to obtain a disaster risk pre-assessment large model; Receiving multimodal prompt data input by a user; wherein the multimodal prompt data includes at least one of the following: position data, scene image data, and scene video data; Using the disaster risk pre-assessment large model to conduct meteorological disaster risk assessment on the multimodal prompt data to obtain a meteorological disaster risk assessment result; When it is determined that the meteorological disaster risk assessment result indicates that there is a meteorological disaster risk, triggering a smart meteorological service based on the meteorological disaster risk assessment result; The smart meteorological service is triggered based on the meteorological disaster risk assessment result, including: Obtaining a fuzzy matching rule library; wherein the fuzzy matching rule library defines a correspondence between a plurality of meteorological disaster risk assessment results and smart meteorological services; wherein the smart meteorological services include: meteorological disaster warning information and meteorological disaster risk avoidance strategies in a plurality of display forms; The meteorological disaster risk assessment result is used to query the fuzzy matching rule library to determine the corresponding smart meteorological service.

2. The intelligent meteorological service method triggered by meteorological risk pre-assessment according to claim 1 is characterized in that: Construct a description vector library of various historical meteorological disasters, including: Acquire multimodal disaster data of historical meteorological disasters, meteorological disaster-related literature, meteorological-related encyclopedia data, location data corresponding to target historical meteorological disasters, meteorological data, satellite images, and video images; wherein the target historical meteorological disaster represents any one of the multiple historical meteorological disasters; Based on the location data and meteorological data corresponding to the target historical meteorological disaster, construct a meteorological data description vector of the target historical meteorological disaster; Based on the satellite images and video images corresponding to the target historical meteorological disaster, construct a meteorological disaster risk scenario description vector of the target historical meteorological disaster; Based on the multimodal disaster data of historical meteorological disasters, meteorological disaster-related literature and meteorological-related encyclopedia data, a meteorological disaster knowledge description vector library is constructed; Respectively calculating the similarity between the meteorological data description vector and the meteorological disaster risk scenario description vector of the target historical meteorological disaster and each vector in the meteorological disaster knowledge description vector library; Based on the meteorological disaster knowledge description vectors whose similarity calculation results exceed a preset threshold, constructing a meteorological disaster knowledge description vector of the target historical meteorological disaster; A description vector library of various historical meteorological disasters is constructed based on the description vectors of various historical meteorological disasters.

3. The intelligent meteorological service method triggered by meteorological risk pre-assessment according to claim 2 is characterized in that: Based on the location data and meteorological data corresponding to the target historical meteorological disaster, a meteorological data description vector of the target historical meteorological disaster is constructed, including: Based on the location data, the observation data of the meteorological stations around the location and the satellite radar meteorological observation data are obtained; The meteorological data, the observation data of the surrounding meteorological stations and the satellite radar meteorological observation data are processed using a preset high-impact weather diagnosis algorithm model library to obtain a high-impact weather diagnosis result; Based on the meteorological data and the high-impact weather diagnosis results, a meteorological data description vector of the target historical meteorological disaster is constructed; wherein the meteorological data description vector includes: basic meteorological elements, high-impact weather elements and high-impact weather diagnosis elements.

4. The intelligent meteorological service method triggered by meteorological risk pre-assessment according to claim 2 is characterized in that: Based on the satellite images and video images corresponding to the target historical meteorological disaster, a meteorological disaster risk scenario description vector of the target historical meteorological disaster is constructed, including: Using a preset meteorological disaster scene recognition model to process the satellite images and video images of the target historical meteorological disaster, a scene recognition result of the target historical meteorological disaster is obtained; wherein the scene recognition result includes: disaster risk type, disaster-affected objects and damage level; The scene recognition result is processed using a general large model to output a meteorological disaster risk scenario description vector of the target historical meteorological disaster; wherein the meteorological disaster risk scenario description vector is a multidimensional vector that comprehensively characterizes the risk scenario characteristics and information of the target historical meteorological disaster.

5. The intelligent meteorological service method triggered by meteorological risk pre-assessment according to claim 2 is characterized in that: Based on the multimodal disaster data of historical meteorological disasters, meteorological disaster-related literature and meteorological-related encyclopedia data, a meteorological disaster knowledge description vector library is constructed, including: Perform semantic segmentation and word embedding on the multimodal disaster data of historical meteorological disasters and relevant literature on meteorological disasters to obtain the initial meteorological disaster knowledge description vector library; Constructing a meteorological knowledge graph based on the meteorological-related encyclopedia data; The meteorological knowledge graph is processed by using graph-based knowledge retrieval enhancement technology to obtain a supplementary meteorological disaster knowledge description vector; The meteorological disaster knowledge description vector library is constructed based on the initial meteorological disaster knowledge description vector library and the supplementary meteorological disaster knowledge description vector.

6. The intelligent meteorological service method triggered by meteorological risk pre-assessment according to claim 1 is characterized in that: After obtaining the meteorological disaster risk assessment results, it also includes: The evaluation reasoning process of the disaster risk pre-assessment model to determine the meteorological disaster risk assessment results is demonstrated.

7. A smart meteorological service system triggered by meteorological risk pre-assessment, characterized in that: include: A construction module is used to construct a description vector library of various historical meteorological disasters; wherein each description vector of the historical meteorological disasters includes: a meteorological data description vector, a meteorological disaster risk scenario description vector and a meteorological disaster knowledge description vector; A training module, used for training the general large model based on the description vector library to obtain a disaster risk pre-assessment large model; A receiving module, configured to receive multimodal prompt data input by a user; wherein the multimodal prompt data includes at least one of the following: position data, scene image data, and scene video data; An assessment module, used to use the disaster risk pre-assessment large model to perform meteorological disaster risk assessment on the multimodal prompt data to obtain a meteorological disaster risk assessment result; A trigger module, for triggering a smart meteorological service based on the meteorological disaster risk assessment result when it is determined that the meteorological disaster risk assessment result indicates that there is a meteorological disaster risk; The trigger module is specifically used for: Obtaining a fuzzy matching rule library; wherein the fuzzy matching rule library defines a correspondence between a plurality of meteorological disaster risk assessment results and smart meteorological services; wherein the smart meteorological services include: meteorological disaster warning information and meteorological disaster risk avoidance strategies in a plurality of display forms; The meteorological disaster risk assessment result is used to query the fuzzy matching rule library to determine the corresponding smart meteorological service.

8. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, characterized in that: When the processor executes the computer program, the intelligent meteorological service method triggered by meteorological risk pre-assessment according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, which, when executed by a processor, implement the intelligent meteorological service method triggered by meteorological risk pre-assessment according to any one of claims 1 to 6.

Images