Automatic insurance authentication generation system based on multi-source meteorological data fusion
The automated insurance verification generation system, which integrates multi-source meteorological data, solves the problems of low efficiency, high cost, and inconsistent results in traditional insurance verification, and achieves efficient and accurate claims processing.
Patent Information
- Application Number
- CN202510385758.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-29
- Publication Date
- 2025-11-11
AI Technical Summary
Traditional insurance assessment relies on manual on-site inspections and experience-based judgments, resulting in long claims processing cycles, high costs, inconsistent results, and difficulty in accurately assessing losses from natural disasters and extreme weather events.
An automated insurance certification generation system based on multi-source meteorological data fusion is adopted. Through the comprehensive collection and fusion of data from automatic weather stations, radar and satellites, combined with the isolated forest algorithm and cross-validation, meteorological certification products are generated and insurance certification certificates are automatically generated.
It improved the efficiency of insurance claims processing, reduced operating costs, minimized the impact of subjective factors, enhanced customer experience, and provided more accurate claims results.
Smart Images
Figure QLYQS_8 
Figure QLYQS_15
Abstract
Description
Technical Field
[0001] This application belongs to the field of insurance authentication technology, specifically, it relates to an automated insurance authentication generation system based on the fusion of multi-source meteorological data. Background Technology
[0002] Insurance attestation is a crucial step in the insurance claims process, aiming to verify whether an insured event meets the claim conditions stipulated in the insurance contract. However, the assessment and claims process for losses caused by natural meteorological disasters and extreme weather events has consistently faced numerous challenges.
[0003] Traditional insurance due diligence relies on on-site investigations by insurance claims personnel, historical experience-based judgments, and paper document review. This presents several significant problems: Manual information collection and analysis are time-consuming and labor-intensive, especially when dealing with large-scale disasters or accidents, requiring the processing of massive amounts of data, leading to lengthy claims processing cycles and poor customer experience. Human judgment is susceptible to the subjective factors of claims personnel, potentially resulting in inconsistent claims outcomes and even disputes. Insurance due diligence requires substantial human resources for information collection, analysis, and judgment, increasing the operating costs of insurance companies. Furthermore, the definitions of clauses regarding "natural disasters" and "force majeure" in insurance contracts are often ambiguous, requiring determination based on specific meteorological data (such as wind speed and precipitation thresholds). Manual processing of large amounts of information is prone to omissions and errors, potentially leading to incorrect claims results that harm the interests of insurance companies or customers.
[0004] Therefore, there is an urgent need for a system that can automatically acquire, analyze, and utilize meteorological data, and automatically generate insurance verification reports based on insurance contract terms, in order to improve insurance claims efficiency, reduce operating costs, minimize the impact of subjective factors, and enhance customer experience. Summary of the Invention
[0005] The purpose of this application is to provide an automated insurance verification generation system based on the fusion of multi-source meteorological data, which has the advantages of improving insurance claims efficiency, reducing operating costs, reducing the impact of subjective factors, enhancing customer experience, effectively integrating multi-source meteorological data, identifying and processing abnormal data, and taking into account the impact of terrain factors on meteorological conditions.
[0006] This application provides an automated insurance certification generation system based on multi-source meteorological data fusion, comprising: an input module for acquiring claim point information input by a user, wherein the claim point information includes the location information, time range, claim conditions, and at least one meteorological element of the target claim point; the meteorological element includes, but is not limited to, at least one of temperature, humidity, precipitation, and wind speed; a meteorological service data acquisition subsystem for acquiring and storing multi-source meteorological detection data collected by automatic weather stations, radar, and satellites; a meteorological data query subsystem for filtering the multi-source meteorological detection data according to the claim point information and outputting a meteorological certification product for the target claim point; an insurance meteorological parameter certification subsystem for generating an insurance meteorological certification certificate based on the meteorological certification product; and a monitoring subsystem for inputting and reviewing the insurance meteorological certification certificate; wherein the meteorological service data acquisition subsystem provides the multi-source meteorological detection data to the meteorological data query subsystem through a standardized API interface.
[0007] Furthermore, the meteorological data query subsystem includes: a multi-source data fusion module, used to perform time alignment and spatial gridding processing on the multi-source meteorological observation data based on the latitude, longitude, and time range of the target claim point; to detect outliers in the ground observation data collected by automatic weather stations using the isolated forest algorithm; and to cross-validate the outliers using the inversion results of meteorological observation data collected by at least one of radar or satellite; and a filtering module, used to filter the ground meteorological observation data of automatic weather stations within a preset radius around the target claim point based on the claim point information and the cross-validation results.
[0008] Furthermore, the multi-source data fusion module includes an anomaly detection unit and a cross-validation unit; the anomaly detection unit is used to randomly extract a subsample set from the ground observation data within the time range of the target claim point, the subsample size being... For each subset of samples, an isolation tree is recursively generated, and features are randomly selected from the subset of samples during each split. And generate segmentation thresholds. According to formula (1), the ground observation data Calculate abnormal scores ,when Data with a value not less than 0.6 is marked as candidate outlier data; (1) (2) Among them, , , The conditions are met, as shown in formula (2), where H is the harmonic number. The cross-validation unit is used to perform at least one of the following satellite or radar validations on the candidate anomaly data, wherein the satellite validation includes obtaining inversion values from satellites at the same location. ,exist An anomaly is determined if the following formula (3) is met, where The historical standard deviation of feature q; (3) The radar verification includes obtaining the radar reflectivity factor Z for the precipitation anomaly point x. When Z ≥ 40dBZ and x_precipitation < 5mm / h or when Z ≤ 35dBZ and x_precipitation > 20mm / h, it is determined to be an anomaly.
[0009] Furthermore, the meteorological data query subsystem also includes: a dynamic adjustment module, used to dynamically adjust the preset radius, wherein the preset radius is a first preset distance in plain areas; in mountainous or hilly areas, the terrain undulation is calculated based on a digital elevation model, and if the undulation exceeds a preset threshold, the radius is reduced to a second preset distance, and automatic weather stations with an altitude difference of less than a third preset distance from the target claim point are preferentially selected.
[0010] Furthermore, the meteorological service data acquisition subsystem includes: a data acquisition module for acquiring meteorological observation data from automatic weather stations within the region, wherein the meteorological observation data includes at least one of hourly data, minute data, and daily data; and a data storage module for storing the meteorological observation data into a cloud database.
[0011] Furthermore, the location information of the target claim point includes at least one of the longitude, latitude, and area code of the administrative region of the target claim point.
[0012] Furthermore, the claims conditions include: meteorological element threshold triggering conditions, including at least one of temperature threshold, cumulative precipitation threshold, and extreme wind speed threshold; time continuity conditions, used to define the cumulative duration or continuous duration for which meteorological elements reach the threshold within the time range; and compound disaster conditions, used to define that claims are triggered when two or more meteorological elements simultaneously meet the threshold.
[0013] Furthermore, the meteorological verification product includes: the name of the automatic weather station near the target claim point, the meteorological observation data of the nearby automatic weather station, and a visualization chart of the impact range of meteorological disasters.
[0014] As can be seen from the above, the insurance certification automation generation system based on multi-source meteorological data fusion provided in this application includes an input module, a meteorological service data acquisition subsystem, a meteorological data query subsystem, an insurance meteorological parameter certification subsystem, and a monitoring subsystem. By acquiring multi-source meteorological observation data, filtering and processing the data, it automatically generates meteorological certification products and insurance meteorological certification certificates, realizing automated processing of insurance certification. It has the advantages of improving insurance claims efficiency, reducing operating costs, reducing the impact of subjective factors, enhancing customer experience, effectively integrating multi-source meteorological data, identifying and processing abnormal data, and considering the impact of terrain factors on meteorological conditions. Detailed Implementation
[0015] This application addresses the challenges faced by traditional insurance verification methods in handling losses caused by natural meteorological disasters and extreme weather events, and proposes an automated insurance verification generation system based on the fusion of multi-source meteorological data.
[0016] In exploring solutions, automated systems based on a single data source were initially considered. While this approach can improve data processing efficiency, it still suffers from issues of data reliability and representativeness. For example, relying solely on ground weather station data may not accurately reflect the actual meteorological conditions in areas with complex terrain.
[0017] Secondly, a simple overlay scheme for multi-source data was considered. While this method increases the comprehensiveness of the data, it does not resolve the issues of data conflicts and inconsistencies. For example, satellite data and ground observation data may differ, and how to handle these differences becomes a critical problem.
[0018] After in-depth analysis, this application ultimately selected a comprehensive solution based on the fusion of multi-source meteorological data. This solution not only integrates multiple data sources but also introduces data fusion and anomaly detection mechanisms to improve the accuracy and reliability of the data.
[0019] Therefore, this application proposes an automated insurance certification generation system based on multi-source meteorological data fusion, comprising: an input module for acquiring claim point information input by the user, the claim point information including the location information, time range, claim conditions, and at least one meteorological element of the target claim point; the meteorological element including but not limited to at least one of temperature, humidity, precipitation, and wind speed; a meteorological service data acquisition subsystem for acquiring and storing multi-source meteorological observation data collected by automatic weather stations, radar, and satellites; a meteorological data query subsystem for filtering the multi-source meteorological observation data according to the claim point information and outputting a meteorological certification product for the target claim point; an insurance meteorological parameter certification subsystem for generating an insurance meteorological certification certificate based on the meteorological certification product; and a monitoring subsystem for inputting and reviewing the insurance meteorological certification certificate; wherein, the meteorological service data acquisition subsystem provides meteorological observation data to the meteorological data query subsystem through a standardized API interface.
[0020] The system comprises the following subsystems: The input module is used to acquire user-inputted claims information, which can be implemented using a graphical user interface or API, such as a micro-webpage on mobile devices. The meteorological service data acquisition subsystem acquires and stores multi-source meteorological observation data, which can be implemented using web crawlers, API calls, or direct data transmission. The meteorological data query subsystem filters multi-source meteorological observation data based on claims information, which can be implemented using database querying techniques. The insurance meteorological parameter verification subsystem generates insurance meteorological verification certificates based on meteorological verification products, which can be implemented using template generation and natural language processing techniques. The monitoring subsystem inputs and reviews insurance meteorological verification certificates, including workflow management and access control functions.
[0021] The core innovation of this application lies in proposing an automated insurance verification generation system based on multi-source meteorological data fusion. This system integrates multiple data sources, including automatic weather stations, radar, and satellites, to achieve comprehensive collection and fusion processing of meteorological data. The system's meteorological data query subsystem can intelligently filter multi-source data based on specific claim point information, thereby generating more accurate and reliable meteorological verification products. Furthermore, the system employs a standardized API interface to achieve data transmission between the meteorological service data collection subsystem and the meteorological data query subsystem. This design not only ensures data real-time performance but also improves the system's scalability and flexibility.
[0022] The working principle of this application can be described in detail as follows: First, the input module obtains the claim point information input by the user through the user interface or API interface, including the location information of the target claim point (such as latitude and longitude or administrative region code), time range, claim conditions, and required meteorological elements (such as temperature, humidity, precipitation, and wind speed). This information is then transmitted to other modules of the system for further processing.
[0023] Secondly, the meteorological service data acquisition subsystem is responsible for acquiring meteorological observation data from multiple data sources. Specifically, this subsystem can obtain data from publicly available meteorological websites through web scraping technology, or directly from meteorological departments' databases via API calls. It can also receive real-time data from automatic weather stations, radar, and satellites through dedicated data transmission channels. After preliminary processing, the acquired data is stored in the system's database.
[0024] Next, the meteorological data query subsystem filters the stored multi-source meteorological observation data based on the claim point information provided by the input module. The filtered data is then integrated into a meteorological verification product for the target claim point, which contains all the necessary meteorological information related to the claim.
[0025] Then, the insurance meteorological parameter verification subsystem receives the meteorological verification product and automatically generates an insurance meteorological verification certificate according to the preset template and rules.
[0026] Finally, the monitoring subsystem is responsible for entering and reviewing the generated insurance weather certificates. This process may include automated data consistency checks as well as manual review to ensure the accuracy and compliance of the certificates.
[0027] The key to the entire system lies in the data transmission between the meteorological service data acquisition subsystem and the meteorological data query subsystem. By adopting a standardized API interface, the system ensures efficient and reliable data transmission, while also facilitating future system expansion and upgrades. This design choice is based on the fact that it not only improves the system's real-time performance but also enhances its flexibility, enabling the system to easily integrate new data sources or update existing data processing algorithms.
[0028] As an example, in a real-world application scenario, an agricultural insurance company needs to verify claims in an area affected by torrential rain. When using this system, the user first inputs the claim location information via the input module: the target claim location is longitude 120.5°E, latitude 30.2°N, the time range is July 1st to July 7th, 2023, the claim condition is daily rainfall exceeding 100mm, and the required meteorological element is rainfall.
[0029] The meteorological service data acquisition subsystem obtains hourly precipitation data from all automatic weather stations within a 50-kilometer radius of the region from the National Meteorological Administration via an API interface, while also acquiring radar reflectivity data and meteorological satellite cloud imagery data covering the area. This data is transmitted in real-time to the system's cloud database for storage.
[0030] After receiving the claims location information, the meteorological data query subsystem first aligns the stored data to an hourly scale. Then, it uses inverse distance weighted interpolation to calculate the hourly precipitation at the target claims location based on data from surrounding automatic weather stations. Simultaneously, the system cross-validates the interpolation results using radar reflectivity data and satellite cloud imagery to ensure accuracy.
[0031] After processing, the system generated a meteorological verification product for the target claim location, including 24-hour daily precipitation data and cumulative precipitation. The results showed that on July 3, the daily precipitation at this location reached 127 mm, exceeding the threshold for claim settlement.
[0032] Based on this result, the insurance meteorological parameter verification subsystem automatically generated an insurance meteorological verification certificate. The certificate details the daily precipitation data for the target claim point over the entire time period and specifically marks the dates that exceeded the claim threshold.
[0033] Finally, the monitoring subsystem automatically checks the generated authentication certificates to ensure data consistency and integrity. In one implementation, the monitoring subsystem also records important user operations after logging into the platform and generates monitoring logs, such as login, logout, password change, entry of insurance weather authentication certificates, and review of insurance weather authentication certificates.
[0034] To define user permissions and operational scope, the automated insurance certification generation system also includes a user management subsystem, which is used to add, modify, and delete users; set user permissions; add, modify, and delete user groups; and set user group permissions. The default settings are an administrator group and a salesperson group, with the salesperson group further divided into an input group and an audit group.
[0035] This embodiment demonstrates how to efficiently and accurately complete the insurance verification process under complex weather conditions, greatly improving claims efficiency and reducing the possibility of human error.
[0036] In some of the embodiments described above, during the implementation of this application, there are still problems such as low processing efficiency of multi-source meteorological data and unstable data quality.
[0037] To address this, this application further proposes a meteorological data query subsystem, including a multi-source data fusion module and a filtering module. The multi-source data fusion module performs time alignment and spatial gridding processing on multi-source meteorological observation data based on the latitude, longitude, and time range of the target claim point. It detects outliers in ground observation data collected by automatic weather stations using the isolated forest algorithm and cross-validates these outliers using inversion results from meteorological observation data collected by at least one source, such as radar or satellite. The filtering module filters ground meteorological observation data from automatic weather stations within a preset radius around the target claim point based on the claim point information and the cross-validation results.
[0038] The meteorological data query subsystem of this application effectively improves the processing efficiency and quality of meteorological data through multi-source data fusion and filtering. The multi-source data fusion module first performs time alignment and spatial gridding processing on meteorological observation data from different sources to ensure data consistency and comparability. Then, it uses the isolated forest algorithm to detect outliers in ground observation data and performs cross-validation using radar or satellite data, improving data accuracy and reliability. The filtering module then accurately filters meteorological observation data around the target claim point based on claim point information and cross-validation results, providing high-quality data support for subsequent insurance verification.
[0039] Specifically, the implementation of the multi-source data fusion module may include the following steps: First, based on the latitude, longitude, and time range of the target claims point, relevant multi-source meteorological observation data is acquired from the meteorological service data acquisition subsystem. This meteorological observation data may come from different observation devices, such as automatic weather stations, radar, and satellites, and may have different temporal resolutions and spatial distributions.
[0040] Next, time alignment is performed. For example, all data can be unified to an hourly scale, minute-level data can be averaged or accumulated, and daily-level data can be interpolated. Spatial gridding can be achieved using methods such as Kriging interpolation or inverse distance weighting to interpolate observation data from different spatial locations onto a unified grid.
[0041] Then, the Isolation Forest algorithm is used to detect outliers in the ground observation data. The principle of the Isolation Forest algorithm is that outlier data points are more easily isolated. In one implementation, the above multi-source data fusion module includes an anomaly detection unit and a cross-validation unit: The anomaly detection unit is used to randomly extract a subset of samples from ground observation data within the time range of the target claim point. Randomly select features And generate segmentation thresholds. According to formula (1), the ground observation data Calculate abnormal scores ,when Data with a value not less than 0.6 is marked as candidate outlier data; (1) (2) Among them, , , The conditions are met, as shown in formula (2), where H is the harmonic number. The cross-validation unit is used to perform at least one of the following verifications on the candidate anomaly data: satellite validation or radar validation. Satellite validation includes obtaining inversion values from satellites at the same location. ,exist An anomaly is determined if the following formula (3) is met, where The historical standard deviation of feature q; (3) Radar verification includes obtaining the radar reflectivity factor Z for the precipitation anomaly point x. When Z ≥ 40dBZ and x_precipitation < 5mm / h, or when Z ≤ 35dBZ and x_precipitation > 20mm / h, it is judged as an anomaly.
[0042] Specifically, the anomaly detection unit employs the Isolation Forest algorithm, which is characterized by high computational efficiency and suitability for high-dimensional data, effectively identifying potential anomalous data points. The anomaly detection unit first randomly samples a subset of data from ground observations. This sampling method reduces computational complexity and improves algorithm efficiency. Then, for each subset, an isolation tree is recursively generated. A decision tree structure is constructed by randomly selecting features and generating a segmentation threshold. This randomness helps capture anomalous patterns in the data. The threshold setting can be adjusted according to the actual application scenario to balance the sensitivity and accuracy of detection.
[0043] The cross-validation unit further validates candidate anomaly data. Satellite validation compares the differences between ground observations and satellite inversion values, combined with historical deviation standard deviations, to determine whether the data is anomaly. This method considers the characteristics and historical variation patterns of different meteorological elements, improving the accuracy of the judgment. Radar validation targets precipitation anomalies, using the radar reflectivity factor Z for judgment. By setting different threshold conditions, it can effectively identify situations where precipitation observations do not match actual weather conditions, such as abnormally low precipitation in strong echo areas or abnormally high precipitation in weak echo areas. The filtering module filters data from automatic weather stations within a preset radius around the target claim point based on claim point information and cross-validation results. During filtering, reliable data that has undergone cross-validation is prioritized, while considering factors such as distance and altitude difference from the target claim point. In one implementation, the meteorological data query subsystem further includes a dynamic adjustment module for dynamically adjusting the preset radius. The preset radius is a first preset distance in plain areas; in mountainous or hilly areas, the terrain undulation is calculated based on a digital elevation model. If the undulation exceeds a preset threshold, the radius is reduced to a second preset distance, and automatic weather stations with an altitude difference of less than a third preset distance from the target claim point are preferentially selected.
[0044] The introduction of this dynamic adjustment module is primarily to address the issue of representativeness and accuracy of meteorological data under varying terrain conditions. In plains areas, meteorological conditions are relatively uniform, allowing for the use of a larger preset radius to select relevant automatic weather stations. However, in mountainous or hilly areas, the complexity of the terrain leads to significant differences in local meteorological conditions. Therefore, a smaller preset radius is needed, taking into account altitude differences, to ensure that the data from the selected weather stations more accurately reflects the actual meteorological conditions at the target claims location. Preferably, the first preset distance is 50 kilometers, and the second preset distance is 30 kilometers.
[0045] In practice, the dynamic adjustment module first determines the terrain type of the target claim point based on its location information. For plains areas, a preset radius of 50 kilometers is used directly. For mountainous or hilly areas, the module uses a digital elevation model (DEM) to calculate the terrain undulation around the target claim point. The calculation method can use the standard deviation method or the relative elevation difference method, etc. If the calculated undulation exceeds a preset third preset distance (for example, a relative elevation difference exceeding 300 meters or a standard deviation exceeding 100 meters), the preset radius is reduced to 30 kilometers.
[0046] After determining the preset radius, the dynamic adjustment module further filters automatic weather stations. For mountainous or hilly areas, in addition to considering distance, it prioritizes automatic weather stations with an elevation difference of less than 200 meters from the target claim point. This is because in complex terrain, elevation differences can lead to significant changes in weather conditions. By limiting the elevation difference, it ensures that the data from the selected weather stations are more representative.
[0047] For example, suppose the target claim point is located in a mountainous area with coordinates (30.5°N, 114.3°E) and an altitude of 800 meters. The dynamic adjustment module first calculates the terrain undulation within a 50-kilometer radius of this point using a digital elevation model, finding a relative elevation difference of 500 meters, which exceeds the preset threshold of 300 meters. Therefore, the module reduces the preset radius to 30 kilometers. Within this 30-kilometer range, the module identifies five automatic weather stations, three of which have altitudes of 750 meters, 820 meters, and 1050 meters. Based on the condition that the altitude difference is less than 200 meters, the module ultimately selects the two weather stations at altitudes of 750 meters and 820 meters as data sources.
[0048] By introducing a dynamic adjustment module, the meteorological data query subsystem of this application can flexibly adjust the data collection range and selection criteria according to different terrain conditions. This method significantly improves the representativeness and accuracy of the acquired meteorological data, especially in mountainous or hilly areas with complex terrain. Compared with the traditional method using a fixed radius, the dynamic adjustment module can better adapt to different geographical environments, thus providing more reliable meteorological data support for insurance verification.
[0049] In some of the embodiments described above, during the implementation of this application, there are still problems such as low efficiency in meteorological service data collection and inconvenience in data storage for subsequent processing.
[0050] To address this, this application further proposes a meteorological service data acquisition subsystem, including a data acquisition module and a data storage module. The data acquisition module is used to collect meteorological observation data from automatic weather stations within the region, including at least one type of meteorological observation data: hourly, minute, and daily data. The data storage module is used to store the meteorological observation data in a cloud database.
[0051] The meteorological service data acquisition subsystem provides meteorological observation data to the meteorological data query subsystem through a standardized API interface, realizing efficient data acquisition and storage, and laying the foundation for subsequent data processing and analysis.
[0052] Specifically, the data acquisition module can collect data from automatic weather stations in various ways. For example, it can use a timed polling method to periodically send data requests to the automatic weather station; or it can use a push mechanism, where the automatic weather station actively pushes data to the acquisition module. The frequency of data acquisition can be adjusted according to actual needs; for example, minute-level data can be collected once per minute, hourly data once per hour, and daily data once per day.
[0053] The data storage module is responsible for storing the collected data into a cloud database. The choice of cloud database can be optimized based on the data volume and access requirements. For example, for minute-level data with high real-time requirements, an in-memory database such as Redis can be used to improve data retrieval speed; for historical data, a distributed file system can be used to achieve efficient management of large-scale data.
[0054] As a preferred implementation, the meteorological service data acquisition subsystem of this application can adopt a distributed architecture design. Specifically, a data acquisition node can be deployed in each region to collect data from automatic weather stations within that region. These acquisition nodes can be deployed using containerization technology, such as Docker containers, to facilitate system expansion and maintenance.
[0055] The data storage module can employ a distributed database cluster, such as Apache Cassandra or MongoDB, to support the storage and rapid querying of massive amounts of meteorological data. The database can be sharded according to time and geographic location to optimize query performance.
[0056] In some of the above embodiments, during the implementation of this application, there is still a problem that the input of target claim point location information is inconsistent, making it difficult for the system to accurately locate and process the claim.
[0057] In this regard, this application further proposes that the location information of the target claim point includes at least one of the longitude, latitude and longitude of the target claim point and the area code of the administrative region to which the target claim point belongs.
[0058] This technical solution increases the system's flexibility and applicability by providing multiple location information input methods. Users can choose the most convenient or accurate location information input method according to their actual situation, thereby improving the system's user-friendliness and data processing efficiency.
[0059] Specifically, the location information input methods in this application include the following implementation methods: 1. Latitude and Longitude Coordinates: Users can directly input the precise latitude and longitude coordinates of the target claim point. This method is suitable when the accurate geographical coordinates are known, and can provide the most accurate location information. For example, the longitude can be entered as "116.3972°E" and the latitude can be entered as "39.9075°N".
[0060] 2. Administrative Region Code: For users who do not know the exact latitude and longitude but are familiar with the administrative divisions, they can enter the region code of the administrative region where the target claim point is located. This method is suitable for claims covering a large area. For example, the administrative region code for Beijing can be entered as "110000".
[0061] 3. Address Conversion: The system can integrate geocoding services, allowing users to input detailed addresses and automatically convert them into latitude and longitude coordinates. This provides users with a more intuitive input method.
[0062] These different input methods can be converted and validated to each other. For example, when a user enters an administrative region code, the system can automatically obtain the center coordinates of that region as the initial location point. When latitude and longitude are entered, the system can look up the corresponding administrative region information to provide more contextual information.
[0063] In practical applications, this diverse method of inputting location information can be implemented as follows: The system provides a unified location information input interface with three input options: latitude and longitude coordinates, administrative region code, and detailed address. Users can choose one of these methods to input the information.
[0064] When a user selects to input latitude and longitude coordinates, the system provides two input boxes for entering longitude and latitude, respectively. The system will perform a validity check on the input values to ensure they are within the valid range (longitude -180° to 180°, latitude -90° to 90°).
[0065] If the user chooses to enter an administrative region code, the system will provide an input box and a drop-down menu. Users can directly enter a 6-digit region code, or have the region code automatically generated by selecting the province, city, or district / county level.
[0066] For detailed address input, the system provides a text box allowing users to enter the complete address information. The system then calls the geocoding service API to convert the address into latitude and longitude coordinates.
[0067] Regardless of the input method chosen by the user, the system will convert the final location information into latitude and longitude coordinates for subsequent data processing and analysis. Simultaneously, the system will mark the location on a map and display the distribution of nearby automatic weather stations, allowing users to intuitively confirm the accuracy of the location.
[0068] This implementation not only solves the problem of diverse location information input but also improves the system's user-friendliness and data processing accuracy. Compared to traditional single-input methods, this approach is more adaptable and practical, better meeting the needs of different users in different scenarios.
[0069] In some of the embodiments described above, during the implementation of this application, there are still problems such as the lack of flexibility in the definition of claims conditions, which cannot meet the needs of complex insurance products.
[0070] In response, this application further proposes an automated insurance verification generation system based on the fusion of multi-source meteorological data, wherein the claim conditions include meteorological element threshold triggering conditions, time continuity conditions, and complex disaster conditions.
[0071] Meteorological element threshold triggering conditions include at least one of temperature threshold, cumulative precipitation threshold, and extreme wind speed threshold. Temporal continuity conditions define the cumulative or continuous duration for which a meteorological element reaches a threshold within a given time frame. Composite disaster conditions define when two or more meteorological elements simultaneously meet their thresholds, triggering a claim.
[0072] This design of claim conditions allows for flexible adaptation to the needs of different types of insurance products. Meteorological threshold trigger conditions allow for setting reasonable claim thresholds based on the climate characteristics of different regions and seasons. For example, for agricultural insurance, a continuous temperature below 0°C for 8 hours can be set as a claim trigger condition for frost damage. Temporal continuity conditions can distinguish between short-term extreme weather and persistent hazardous weather, more accurately determining whether the claim criteria have been met. For example, cumulative precipitation exceeding 50mm within 24 hours can be set as a claim condition for rainstorm damage. Composite disaster conditions consider disasters caused by the combined effects of multiple meteorological factors. For instance, typhoon disasters typically involve both heavy rainfall and strong winds; therefore, a claim condition for typhoon disasters can be set as 1-hour precipitation exceeding 30mm and instantaneous wind speed exceeding 17.2m / s.
[0073] By setting these multi-dimensional claim conditions, the system can more accurately determine whether claim criteria are met. Meteorological threshold triggering conditions provide a basis for basic judgment, while time continuity and complex disaster conditions further refine the judgment criteria, making the definition of claim conditions closer to actual disaster occurrences. This design greatly improves the accuracy and flexibility of claim judgment, adapting to the needs of various complex insurance products.
[0074] In practice, the system allows insurance companies to flexibly set claim conditions based on the characteristics of different insurance products through a user interface. For example, for a crop frost damage insurance product, the following claim conditions can be set: 1. Meteorological element threshold triggering condition: temperature below -2°C 2. Time continuity condition: The temperature threshold must be reached continuously for 6 hours. 3. Complex disaster conditions: Simultaneous wind speed greater than 5 m / s When the system receives a claim application, it retrieves relevant meteorological data from the meteorological data query subsystem based on the input claim location information (including location and time range). Then, the system assesses the claim based on the set claim conditions: first, it checks if the temperature is below -2°C; if so, it further checks if the temperature has remained below this level for six consecutive hours; finally, it checks if the wind speed during these six hours is greater than 5 m / s. Only when all conditions are met will the system determine that the claim meets the criteria.
[0075] Compared to traditional single-threshold judgment methods, this design can more accurately reflect the actual disaster situation. For example, short-term low temperatures may not cause serious damage to crops, while sustained low temperatures are more likely to cause frost damage. At the same time, considering wind speed can better simulate the actual perceived temperature, because strong winds will exacerbate the damage of low temperatures to crops.
[0076] In this way, the system of this application can flexibly set claims conditions according to the characteristics of different insurance products and the features of actual meteorological disasters, greatly improving the accuracy and objectivity of claims judgment. This not only reduces the subjectivity and inconsistency of human judgment, but also better balances the interests of insurance companies and policyholders, reduces claims disputes, and improves the reliability and attractiveness of insurance products.
[0077] In some of the embodiments described above, during the implementation of this application, there is still the issue of how to provide more detailed and intuitive meteorological verification product information.
[0078] In this regard, this application further proposes that the meteorological verification products include: the names of automatic weather stations near the target claim point, meteorological observation data from the nearby automatic weather stations, and a visualization chart of the impact range of meteorological disasters.
[0079] This improvement helps insurance companies and customers better understand and assess weather disasters by providing more comprehensive and visualized information on weather verification products, thereby enhancing the accuracy and transparency of claims decisions.
[0080] Specifically, meteorological verification products consist of three main components: First, the names of nearby automatic weather stations are included in the weather verification product. This information helps insurance companies and customers determine the exact location of the data source, increasing the credibility of the verification results. For example, if the target claim point is located in a suburban area of a city, the weather verification product may list the names of nearby automatic weather stations such as "Suburban Station A" and "Suburban Station B".
[0081] Secondly, meteorological data from nearby automatic weather stations are incorporated into the meteorological verification product. This data may include specific values for key meteorological elements such as temperature, humidity, precipitation, and wind speed. For example, it might provide detailed information such as, "From 14:00 to 16:00 on July 1, 2023, station A in the urban suburbs recorded a maximum wind speed of 25 m / s and a cumulative precipitation of 80 mm." This specific data provides an objective basis for claims decisions.
[0082] Finally, as part of meteorological verification products, the visualization chart of the impact range of meteorological disasters visually displays the spatial distribution and intensity of meteorological disasters. This visualization can take the form of contour maps, heat maps, or color-coded maps. For example, for a rainstorm event, a color map displaying different rainfall levels can be generated, clearly marking the location of the target claim point and the rainfall intensity distribution in the surrounding area.
[0083] These three components complement each other, forming a comprehensive meteorological verification product. By providing the names of nearby automatic weather stations, insurance companies can verify the reliability of the data source. Specific meteorological observation data provides quantitative evidence to support claims decisions. And visual charts help non-experts quickly understand the overall situation and spatial distribution of meteorological hazards.
[0084] In practical applications, this weather verification product can be presented as follows: Suppose there is a claim case involving crop damage caused by heavy rain, and the target claim location is located in an agricultural area of a certain county. The weather verification product may include the following: 1. Names of nearby automatic weather stations: County Agricultural Weather Station A, County Agricultural Weather Station B 2. Meteorological observation data: – County Agricultural Meteorological Station A: From 08:00 to 20:00 on August 15, 2023, the cumulative precipitation was 180 mm, and the maximum hourly precipitation was 45 mm (occurring from 14:00 to 15:00). – County Agricultural Meteorological Station B: From 08:00 to 20:00 on August 15, 2023, the cumulative precipitation was 165 mm, and the maximum hourly precipitation was 42 mm (occurring from 13:00 to 14:00). 3. Visualization chart of the impact range of meteorological disasters: A color precipitation distribution map covering the entire county, using different colors to represent different precipitation levels, and clearly marking the location of the target claims point.
[0085] By providing such detailed meteorological verification products, insurance companies can more accurately assess the actual impact of meteorological disasters on target claim points. For example, they can determine whether rainfall data meets the criteria for a rainstorm as stipulated in the claim conditions, and use visual charts to understand the overall rainfall situation in the surrounding area of the target claim point, thus assessing the reasonableness of crop damage.
[0086] The above description is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined in the claims, they should all fall within the protection scope of the present invention.
[0087] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.
[0088] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to specific implementations. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.
Claims
1. An automated insurance authentication generation system based on multi-source meteorological data fusion, characterized in that, include: The input module is used to obtain the claim point information input by the user. The claim point information includes the location information of the target claim point, the time range, the claim conditions, and at least one meteorological element. The meteorological element includes, but is not limited to, at least one of temperature, humidity, precipitation, and wind speed. The meteorological service data acquisition subsystem is used to acquire and store multi-source meteorological observation data collected by automatic weather stations, radar, and satellites; The meteorological data query subsystem is used to filter the multi-source meteorological detection data based on the claim point information and output the meteorological verification product for the target claim point. The insurance meteorological parameter verification subsystem is used to generate insurance meteorological verification certificates based on the meteorological verification products. The monitoring subsystem is used to input and review the insurance weather certification. The meteorological service data acquisition subsystem provides the multi-source meteorological detection data to the meteorological data query subsystem through a standardized API interface.
2. The automated insurance verification generation system as described in claim 1, characterized in that, The meteorological data query subsystem includes: The multi-source data fusion module is used to perform time alignment and spatial gridding processing on the multi-source meteorological observation data based on the latitude, longitude and time range of the target claim point, detect outliers in the ground observation data collected by automatic weather stations through the isolated forest algorithm, and cross-validate the outliers through the inversion results of meteorological observation data collected by at least one of radar or satellite. The filtering module is used to filter the ground meteorological observation data of automatic weather stations within a preset radius around the target claim point based on the claim point information and cross-validation results.
3. The automated insurance verification generation system as described in claim 2, characterized in that, The multi-source data fusion module includes an anomaly detection unit and a cross-validation unit; The anomaly detection unit is used to randomly extract a subsample set from the ground observation data within the time range of the target claim point, with a subsample size of [missing information]. For each subset of samples, an isolation tree is recursively generated, and features are randomly selected from the subset of samples during each split. And generate segmentation thresholds. According to formula (1), the ground observation data Calculate abnormal scores ,when Data with a value not less than 0.6 is marked as candidate outlier data; (1) (2) Among them, , , It satisfies formula (2), where H is the harmonic number; The cross-validation unit is used to perform at least one of the following verifications on the candidate anomaly data: satellite verification or radar verification. The satellite verification includes obtaining inversion values from satellites at the same location. ,exist An anomaly is determined if the following formula (3) is met, where The historical standard deviation of feature q; (3) The radar verification includes obtaining the radar reflectivity factor Z for the precipitation anomaly point x. When Z ≥ 40dBZ and x_precipitation < 5mm / h or when Z ≤ 35dBZ and x_precipitation > 20mm / h, it is determined to be an anomaly.
4. The automated insurance verification generation system as described in claim 2, characterized in that, The meteorological data query subsystem also includes: The dynamic adjustment module is used to dynamically adjust the preset radius. The preset radius is a first preset distance in plain areas; in mountainous or hilly areas, the terrain undulation is calculated based on the digital elevation model. If the undulation exceeds a preset threshold, the radius is reduced to a second preset distance, and an automatic weather station with an altitude difference of less than a third preset distance from the target claim point is preferentially selected.
5. The automated insurance verification generation system as described in claim 1, characterized in that, The meteorological service data acquisition subsystem includes: The data acquisition module is used to collect meteorological observation data from automatic weather stations within the area, wherein the meteorological observation data includes at least one of hourly data, minute data, and daily data; The data storage module is used to store the meteorological detection data into a cloud database.
6. The automated insurance verification generation system as described in claim 1, characterized in that, The location information of the target claim point includes at least one of the longitude, latitude, and area code of the administrative region to which the target claim point belongs.
7. The automated insurance verification generation system as described in claim 1, characterized in that, The claim conditions include: Meteorological element threshold triggering conditions include at least one of temperature threshold, cumulative precipitation threshold, and extreme wind speed threshold; The time continuity condition is used to define the cumulative or continuous duration for which meteorological elements reach a threshold within the time range. Composite disaster conditions are used to define when two or more meteorological elements simultaneously meet a threshold to trigger a claim.
8. The automated insurance verification generation system as described in claim 1, characterized in that, The meteorological verification product includes: the name of the automatic weather station near the target claim point, the meteorological observation data of the nearby automatic weather station, and a visualization chart of the impact range of meteorological disasters.