Maize planting income insurance double-precision double-reduction technology integrated intelligent system based on air-space-ground multi-source data

Abstract

The invention relates to the crossing field of intelligent agricultural insurance and intelligent agriculture, in particular to a corn planting income insurance double-precision double-reduction technology integrated intelligent system based on space-space-ground multi-source data, which comprises a space-space-ground multi-source data acquisition module, a corn planting income insurance management module, a corn planting income insurance management module and a corn planting income insurance management module, the multi-dimensional data acquisition module is used for acquiring multi-dimensional data such as satellite remote sensing, unmanned aerial vehicle aerial survey fine monitoring, ground Internet of Things intelligent sensing, meteorological monitoring, government affairs and markets in the whole corn planting period; according to the method, according to the CGCS2000 national geodetic coordinate system standard, the space, growth state and environment monitoring of a corn planting area are collected, space-air-ground multi-source data and a DSSAT corn growth model are fused, a standardized data set is output, the variety identification accuracy is larger than or equal to 95%, the planting area underwriting error is smaller than or equal to 2%, the yield prediction accuracy is larger than or equal to 92%, and the method is suitable for large-scale popularization and application. According to the method, yield loss-price fluctuation double risk association is integrated, yield and price data are dynamically updated, the insurance claim settlement efficiency and precision are greatly improved, the pain points of single data and inaccurate accounting of traditional agricultural insurance are solved, and a solid data support is provided for double-precision and double-reduction.

Description

TECHNICAL FIELD

[0001] The application relates to the field of agricultural insurance and intelligent agriculture, in particular to a corn planting income insurance "double-precision two-reduction" technology integrated intelligent system based on space-air-ground multi-source data. BACKGROUND

[0002] Corn is the first grain crop in China, and income insurance is a core means to protect farmers' planting income, which needs to cover both "yield loss" and "price fluctuation" risks. However, traditional corn income insurance has many technical bottlenecks:

[0003] Insufficient precision: relying on manual survey of plots and yield estimation, with large errors; corn variety differences lead to different disaster resistance and yield, but traditional insurance does not differentiate; price collection is lagging, making it difficult to dynamically reflect market changes, and income accounting is not comprehensive.

[0004] Low efficiency and accuracy of claims: disaster loss assessment relies on manual investigation, which is greatly affected by terrain and growth period, and has limited coverage; it does not consider the disaster sensitivity differences of corn at different growth stages (such as the disaster loss of large horn period being much higher than that of seedling period), and the loss assessment result is highly subjective; the claims process is complicated, with a cycle of 15-30 days, and the farmer's disaster reduction and reproduction funds are slow to arrive.

[0005] Poor targeting of risk prevention and control: the occurrence and prevention of common corn disasters (lodging, corn borer, large spot disease, drought) have significant growth period specificity, but traditional risk early warning lacks targeting; risk prevention and control are disconnected from insurance business, and do not form a "early warning-intervention-loss assessment-claims" closed loop, with limited disaster prevention and loss reduction effect.

[0006] Low data credibility and process transparency: yield and price data are stored in centralized platforms and are easily tampered with; business processes rely on manual verification by multiple parties, which has moral risk and high trust cost for farmers and insurance companies.

[0007] Existing technologies mainly focus on single yield insurance or single data source application, and do not achieve "deep integration of space-air-ground multi-source data + adaptation to corn growth characteristics + double-factor income protection + "double-precision two-reduction" technology integration" design, which cannot fundamentally solve the core problems of traditional corn income insurance. Therefore, a corn planting income insurance "double-precision two-reduction" technology integrated intelligent system based on space-air-ground multi-source data is proposed.

[0008] At present, there are many pain points in the corn cost insurance mode: first, the underwriting link relies on manual verification, the land ownership is not clear, the planting area is over-reported, the data error rate is high, and the supervision of the financial subsidy fund is difficult; second, the efficiency of surveying and loss is low, the artificial on-site measurement is long and low in precision due to the limitation of terrain, and the farmer's compensation is slow; third, the risk control is in a passive state, there is a lack of whole-process risk management mechanism for the growth cycle of corn, the pre-disaster warning is lagging and the intervention is insufficient in the disaster, which is difficult to effectively reduce the disaster loss; fourth, there are serious barriers between departments, and the data of agricultural and rural, meteorological and financial departments cannot be connected, so it is difficult to form a synergistic effect of supervision and risk prevention and control.

[0009] Although the existing agricultural insurance system introduces some digital technologies, there are still limitations: multi-source data integration is insufficient, a three-dimensional collection network has not been formed, the intelligent degree of underwriting and claim settlement is low, the compensation standard has not been dynamically adjusted according to the growth cycle of corn, and the risk reduction and disaster prevention and loss reduction are disconnected, and there is no precise prevention and control scheme for the specific risks of corn frost damage and pest damage. Therefore, it is urgent to build an integrated intelligent system based on multi-source data of space, air and ground, to break through the drawbacks of the traditional mode and help the high-quality development of corn planting full cost insurance, with "precise underwriting and precise claim settlement" as the core result and "risk reduction and disaster prevention and loss reduction" as the key process. SUMMARY

[0010] Therefore, the present application provides a corn planting income insurance "double-precision two-reduction" technology integrated intelligent system based on multi-source data of space, air and ground, to solve or alleviate the technical problems in the prior art, and at least provide a beneficial choice.

[0011] The technical solution of the present application is as follows: a corn planting income insurance "double-precision two-reduction" technology integrated intelligent system based on multi-source data of space, air and ground, comprising:

[0012] (1) Mode I, comprising the following modules:

[0013] Space-air-ground multi-source data acquisition module: integrating satellite remote sensing, unmanned aerial vehicle surveying and mapping, and ground Internet of Things equipment, collecting spatial data, growth state data and environmental monitoring data of corn planting areas according to the CGCS2000 national geodetic coordinate system standard, and outputting standardized data sets;

[0014] Precise underwriting module: based on multi-source data to realize insurance information verification, ownership verification, intelligent mark verification, risk rating and underwriting publicity, to meet the technical requirements of online insurance rate ≥ 90%, biological identification identity verification coverage 100%, and data error rate ≤ 3%;

[0015] Accurate claim settlement module: through intelligent case reporting, space-time coordination survey, AI dynamic damage assessment, automatic calculation and payment of compensation, 24-hour survey response, damage assessment within 10 working days, compensation within 3 days, damage assessment accuracy ≥95%;

[0016] Risk reduction module: based on meteorological, plant protection and other multi-source data, a three-level early warning system is established to carry out pre-disaster prediction and warning, disaster prevention technology training and risk investigation, with an early warning accuracy of ≥90% and a high-risk area inspection cycle of ≤48 hours;

[0017] Disaster prevention and loss reduction module: after the occurrence of disaster, a coordinated response is started to realize disaster reduction through unmanned aerial vehicle inspection and agricultural technology guidance, and to realize post-disaster loss reduction through loss accurate determination and secondary disaster monitoring, with a loss expansion reduction of ≥20%;

[0018] Data sharing and supervision module: realize data interconnection with agricultural and rural, meteorological, financial, gold monitoring and other departments, establish cross verification mechanism of insurance and claim data, support financial subsidy supervision, reinsurance arrangement and large disaster risk reserve management;

[0019] User interaction module: provides online insurance, reporting, inquiry, consultation and complaint functions, integrates AI intelligent consultation and manual service, and responds to complaints within 7 working days.

[0020] Further preferably, the space-time-ground multi-source data acquisition module comprises:

[0021] Potential monitoring and disaster range preliminary judgment (such as flood and drought affected areas).

[0022] Workflow: automatically obtain satellite images every day, extract land boundaries, crop coverage and other information through image interpretation algorithm, compare and calibrate with land right data, generate standardized spatial data set, update cycle ≤24 hours.

[0023] Unmanned aerial vehicle aerial survey sub-module:

[0024] Technical parameters: aerial survey image resolution ≤2 meters, land vector map format Shapefile or GeoJSON, boundary error within ±0.5 meters, supporting manual / automatic route planning.

[0025] Workflow: for planting households of more than 30 mu or sampled land, automatically generate aerial survey tasks, unmanned aerial vehicle flies according to the route to collect images, draws land boundaries and verifies planting scale through AI algorithm, synchronously extracts crop growth status (such as seedling emergence rate in seedling stage, plant height in jointing stage), and uploads image data to the system in real time.

[0026] Ground Internet of Things sub-module:

[0027] Technical parameters: Deployment density ≥ 1 unit / 200 mu, including soil moisture meter (monitoring soil moisture and temperature), small weather station (monitoring precipitation, wind speed and air temperature), data transmission delay ≤ 30 seconds, sampling frequency is 1 time / hour.

[0028] Workflow: The equipment collects environmental data in real time and uploads it to the system via 4G / 5G network. After data cleaning and standardization, it is stored in the database to provide real-time data support for risk warning and loss assessment.

[0029] Furthermore, the precise underwriting module incorporates a corn planting rights database, which integrates land rights confirmation data from the agricultural and rural affairs departments with data from the Third National Land Survey ("Third Survey"). It uses facial recognition technology to verify the insured's identity and implements graded verification based on the insured area.

[0030] When the insured area is less than 10 mu, the verification rate shall be ≥2% and the sample size shall be no less than 3 households, and satellite image verification shall be combined with random on-site inspection.

[0031] When the insured area is between 10 mu and 30 mu, the verification rate is ≥5%, and the top 3 households with the largest insured area are subject to mandatory verification.

[0032] When the insured area is ≥30 mu, drone aerial surveying and marker verification shall be mandatory, with a verification rate of 100%.

[0033] In the case of individual investors applying for insurance, the verification efficiency is ≤3 working days / household. The verification images must include a distant view of the insured object (with geographical identifier), a close-up view (corn crop characteristics), and a group photo of the insured, and be uploaded to the system for archiving in a timely manner.

[0034] Furthermore, the precise claims module incorporates a dynamic claims model for the corn growth period, which divides the corn growth cycle into four stages: seedling stage, jointing stage, tasseling and silking stage, and grain filling and maturity stage.

[0035] For seedlings facing risks of drought and continuous rain, a loss rate threshold of ≤30% is set, with a compensation coefficient of 0.5. Soil moisture data combined with visible light images is used to verify the loss.

[0036] During the jointing stage, in response to the risks of waterlogging and lodging, and drought, a loss rate threshold of 20%-50% is set, with a compensation coefficient of 0.6, and multispectral remote sensing data is used to verify the loss.

[0037] During the male-eating and silking period, in response to the risks of high temperature and drought, a loss rate threshold of ≥20% is set, with a compensation coefficient of 0.8, and the loss is verified by using drone inspection data.

[0038] During the grain filling and ripening period, in response to the risks of continuous rain and ear sprouting, a loss rate threshold of ≥10% is set, with a compensation coefficient of 1.0. Losses are verified by combining IoT humidity sensor data with satellite thermal infrared imagery.

[0039] After receiving the data from the air-space-ground collaborative survey, the AI ​​model generates a corn loss assessment report within 4 hours; for total loss cases (loss rate ≥ 80%), the assessment is completed within 10 days, and for partial loss cases (20% ≤ loss rate < 80%), the assessment is completed within 20 days.

[0040] More preferably, the risk reduction module divides the early warning into three levels according to the probability of disaster occurrence:

[0041] Blue alert: The probability of disaster occurrence is ≥30%. Initiate the risk assessment process and complete the rectification of potential hazards in corn planting areas within 72 hours.

[0042] Yellow alert: The probability of disaster occurrence is ≥50%. Develop a targeted disaster prevention and control plan within 48 hours, including corn field management measures and emergency preparedness.

[0043] Red Alert: The probability of a disaster occurring is ≥70%. The emergency response plan for major disasters will be activated, and a 24-hour duty response will be implemented.

[0044] For severe convective weather, the advance warning time is ≥40 minutes. Warning information and corn disaster prevention guidance are pushed to policyholders via SMS and system APP. No less than 2 disaster prevention technology training sessions are conducted every year, with a farmer training coverage rate of ≥90%. The investment in pre-disaster prevention funds is ≥1.5% of the annual premium income.

[0045] Further preferably, the data sharing and supervision module uses JSON format to transmit cross-departmental data and encrypts sensitive information such as policyholder identity information and fiscal subsidy data using the SM4 national cryptographic algorithm; it achieves real-time sharing of 12 types of core metadata, including corn planting area, land ownership information, disaster early warning data, insured list, claims list, fiscal subsidy payment records, reinsurance agreement information, major disaster risk reserve provision data, corn growth monitoring data, disaster loss assessment reports, farmer complaint records, and system operation logs; the cross-departmental data query response speed is ≤30 seconds / time, and an automatic cross-verification mechanism for underwriting and claims data is established, with verification dimensions including consistency between insured area and confirmed area, matching of claims loss with disaster scope, and suitability of fiscal subsidy amount with insured scale.

[0046] (2) Mode 2, which includes the following modules:

[0047] Multi-source data acquisition module (space, air, and ground): used to acquire satellite remote sensing data, UAV refined monitoring data, ground intelligent sensing data, meteorological monitoring data, and multi-dimensional government and market data throughout the entire corn planting cycle;

[0048] Multi-source data preprocessing and maize growth coupling and fusion module: Cleans, denoises, and aligns the collected data in time and space. It achieves deep coupling of heterogeneous data by fusing maize growth model (DSSAT) and extracts key features related to maize growth and income.

[0049] The dual-precision core business modules include a precise underwriting sub-module and a precise claims settlement sub-module, which respectively realize corn variety identification, land ownership confirmation, dynamic income accounting, risk level pre-assessment, as well as precise disaster location, differentiated loss assessment during the growing season, and income loss accounting;

[0050] The two-reduction coordinated intervention module includes a risk reduction module and a disaster prevention and loss reduction sub-module, which provides graded early warning and intervention guidance for risks (diseases, lodging, drought, etc.) during the key growth period of maize, so as to achieve risk reduction and disaster prevention and loss reduction.

[0051] Digital Twin and Blockchain Trust Assurance Module: Construct a digital twin of corn planting, store key data (yield forecast, price data, loss assessment report) on the blockchain, and realize automated underwriting and claims settlement through smart contracts;

[0052] Multi-terminal collaborative interaction module: Provides customized interactive interfaces for farmers, insurance institutions, and government departments, supporting business processing, early warning reception, data query, and decision analysis.

[0053] More preferably, the air-space-ground multi-source data acquisition module includes:

[0054] Satellite remote sensing submodule: Acquires multispectral and SAR data from Gaofen satellites (Gaofen-6 and Sentinel-2) with a resolution ≤10m, for cornfield boundary identification, crop growth monitoring (NDVI / EVI2 index), and large-scale disaster (flood, drought) identification;

[0055] The UAV-based precision monitoring submodule is equipped with a hyperspectral camera, thermal imager, and lidar to acquire centimeter-level images and identify corn varieties, ear development status, lodging degree, and localized disease and pest patches (great leaf spot, corn borer).

[0056] Ground-based intelligent sensing submodule: Sensor nodes are deployed at a density of 30 mu / node, integrating sensors for soil temperature and humidity, nitrogen, phosphorus and potassium nutrients, plant stem thickness, and field cameras, focusing on collecting data during the corn jointing stage, large trumpet stage, and grain filling stage;

[0057] Meteorological monitoring submodule: Integrates data from national meteorological stations and regional micro meteorological stations to obtain key meteorological factors for maize growth (temperature, precipitation, relative humidity, sunshine duration) and meteorological conditions for the prevalence of pests and diseases;

[0058] The government affairs and market multi-dimensional sub-module connects to data on corn planting registration, variety registration, and land rights confirmation from agricultural and rural departments, as well as real-time corn purchase prices and agricultural input prices from the National Grain Trade Center, regional purchasing markets, and e-commerce platforms.

[0059] More preferably, the multi-source data preprocessing and maize growth coupling and fusion module adopts a four-level architecture of "preprocessing - spatiotemporal alignment - growth model coupling - feature fusion":

[0060] Preprocessing: Outlier removal (3σ principle), image denoising (bilateral filtering), and data completion (GA-LSTM prediction) ensure data integrity ≥98%;

[0061] Spatiotemporal alignment: Based on the GIS latitude and longitude coordinate system and the maize growth period time axis, multi-source data are mapped to a unified spatiotemporal dimension;

[0062] Growth model coupling: The preprocessed data is input into the DSSAT maize growth model to calibrate key parameters such as leaf area index, biomass, and yield, and output dynamic growth status data.

[0063] Feature fusion: An attention-driven multimodal fusion algorithm is used to weight and fuse remote sensing growth features, sensing environment features, model growth features, and market price features to output the core feature vector of corn income insurance.

[0064] More preferably, the precise underwriting submodule includes:

[0065] Accurate land ownership determination for corn varieties and plots: Corn varieties are identified using UAV hyperspectral imagery and CNN algorithms (accuracy ≥ 95%). Combined with satellite remote sensing and land ownership data, plot boundaries are automatically extracted and planting area is calculated (error ≤ 2%).

[0066] Dynamic Revenue Accounting: Construct a revenue model of "production forecast × real-time price integration". Production forecast is based on the fused feature vector + XGBoost algorithm (prediction accuracy ≥92%). Price integrates real-time data at the national, regional and market levels and dynamically updates the insured amount benchmark.

[0067] Risk level pre-assessment: Integrating the disaster resistance of corn varieties, historical disaster data, plot soil conditions, and meteorological risk thresholds, the random forest algorithm is used to classify risks into three levels: high, medium, and low, and insurance rates are set differently (15% higher for high-risk areas and 10% lower for low-risk areas).

[0068] More preferably, the precise claims processing submodule includes:

[0069] Precise disaster location: Through satellite remote sensing change detection (CVA algorithm) + drone field patrol, the system automatically identifies the type of corn disaster (lodging, drought, flood, pests and diseases), its location and scope of impact, with a response time of ≤24 hours;

[0070] Differentiated loss assessment based on growth period: Based on the disaster impact coefficient of maize growth period (0.3 for seedling stage, 0.6 for jointing stage, 1.0 for large trumpet stage, 1.2 for grain filling stage, and 0.8 for maturity stage), the U-Net semantic segmentation algorithm is used to calculate the damaged area, and the yield reduction rate is calculated in combination with the yield loss model (error ≤ 4%).

[0071] Revenue loss accounting: Covers both "production loss type" and "price decline type" losses. The formula is "Revenue loss = (forecasted production - actual production) × average price + max(agreement price - actual price, 0) × actual production", and automatically generates a claim report.

[0072] More preferably, the two-reduction synergistic intervention module includes:

[0073] Risk reduction module: Based on risk level and corn growth stage, push differentiated prevention and control plans: high-risk areas push early warning once every 5 days, focusing on corn borer, large leaf spot and lodging risk; medium and low-risk areas push routine management suggestions according to growth stage;

[0074] Disaster prevention and mitigation submodule: Before a disaster, emergency prevention and control plans are pushed out (such as chemical control to prevent lodging before a typhoon, and water-saving irrigation before a drought); during a disaster, precise intervention measures are pushed out (such as targeted application of pesticides to pests and diseases); after a disaster, production recovery plans are pushed out (such as supporting seedlings and applying fertilizer after lodging, and foliar fertilizer spraying after drought during the grain-filling period), with a disaster prevention and mitigation rate of ≥18%.

[0075] More preferably, the digital twin and blockchain trust assurance module includes:

[0076] Digital twin: Construct a digital twin model of the entire corn planting cycle to map the plot status, growth progress, and risk situation in real time, and support disaster loss simulation and intervention effect prediction;

[0077] Blockchain evidence storage: Adopting a consortium blockchain architecture (insurance institutions + agricultural and rural departments + grain purchasing enterprises + farmers), the results of variety identification, yield forecast data, price data, and loss assessment reports are uploaded to the blockchain to ensure immutability and traceability;

[0078] Smart contracts: Preset rules for underwriting effectiveness, loss assessment triggering, and claims payment, supporting automatic claims triggering for price decline losses, shortening the claims cycle to within 2 business days.

[0079] More preferably, the multi-terminal collaborative interaction module includes a farmer's APP, an insurance institution's intelligent platform, and a government supervision and decision-making platform:

[0080] Farmer APP: Supports variety registration, insurance payment, risk warning reception, field management video guidance, claims application and progress inquiry, and is compatible with voice interaction;

[0081] Insurance institution intelligent platform: supports underwriting review, loss assessment review, claims calculation, risk data analysis, policy management, and supports automatic generation of business reports;

[0082] Government regulatory and decision-making platform: Supports compliance supervision of insurance business, analysis of the corn industry situation (output, price, and disaster impact), and assessment of food security.

[0083] The embodiments of the present invention have the following advantages due to the adoption of the above technical solutions:

[0084] I. This invention integrates multi-source data from air, space, and ground with the DSSAT maize growth model to achieve a variety identification accuracy rate of ≥95%, a planting area calculation error of ≤2%, and a yield prediction accuracy rate of ≥92%. It addresses the pain points of traditional insurance, such as "single data and inaccurate calculations," and innovatively integrates the dual risk calculations of "yield loss" and "price fluctuations," dynamically updating yield and price data to ensure comprehensive income protection, which is in line with the characteristics of large fluctuations in maize planting income.

[0085] Third, this invention designs differentiated loss assessment algorithms and intervention schemes to address the disaster sensitivity and growth needs of corn at different growth stages, achieving a disaster prevention and loss reduction rate of ≥18%. It shifts from passive claims processing to proactive prevention and control. Digital twins support risk simulation and effect prediction, blockchain ensures data immutability and process traceability, and smart contracts shorten the claims cycle to within 2 working days, enhancing system credibility and operational efficiency. The farmer's APP provides convenient operation, insurance institutions' platforms improve operational efficiency, and government platforms support macro-level decision-making, achieving a win-win situation for farmers, insurance institutions, and the government, thus contributing to the stable development of the corn industry and ensuring food security. Therefore, this invention focuses on the unique risks of corn, establishing a closed loop of "pre-disaster early warning - in-disaster intervention - post-disaster loss reduction," achieving both risk reduction (disaster prevention) and disaster prevention and loss reduction (loss reduction + prevention of secondary disasters), proactively reducing disaster losses, establishing a standardized data sharing mechanism, achieving multi-departmental data interoperability and collaborative supervision, ensuring the safety of fiscal subsidies, and improving regulatory efficiency.

[0086] The above overview is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the invention will become readily apparent from the accompanying drawings and the following detailed description. Attached Figure Description

[0087] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0088] Figure 1 This is a diagram of the overall system architecture of the present invention;

[0089] Figure 2 This is a logical diagram of the technical process of the present invention;

[0090] Figure 3 This is a flowchart of the precise underwriting process of the present invention;

[0091] Figure 4 This is a flowchart illustrating the risk reduction process of the present invention. Detailed Implementation

[0092] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0093] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0094] (1) As Figure 1 As shown, this embodiment of the invention provides an integrated intelligent system for corn planting income insurance based on multi-source data from air, space, and ground, incorporating "dual precision and two reductions" technologies, including:

[0095] Multi-source data acquisition module: integrates satellite remote sensing, UAV aerial surveying, and ground IoT equipment, and collects spatial data, growth status data and environmental monitoring data of maize planting area according to the CGCS2000 national geodetic coordinate system standard, and outputs standardized datasets;

[0096] Precise underwriting module: Based on multi-source data, it realizes the verification of insurance information, ownership verification, intelligent label verification, risk rating and underwriting disclosure, meeting the technical requirements of online insurance rate ≥90%, biometric identity verification coverage 100%, and data error rate ≤3%.

[0097] Precise claims module: Through intelligent reporting, air-ground collaborative investigation, AI dynamic damage assessment, automated calculation and payment, it achieves 24-hour investigation response, damage assessment within 10 working days, and claims payment within 3 days, with an assessment accuracy of ≥95%.

[0098] Risk reduction module: Based on multi-source data such as meteorology and plant protection, a three-level early warning system is constructed to carry out pre-disaster forecasting and early warning, disaster prevention technology training and risk investigation. The early warning accuracy rate is ≥90%, and the inspection cycle of high-risk areas is ≤48 hours.

[0099] Disaster prevention and mitigation module: After a disaster occurs, a collaborative response is initiated to reduce losses during the disaster through drone inspections and agricultural technical guidance, and to reduce losses after the disaster through accurate loss assessment and secondary disaster monitoring, thereby reducing the proportion of losses from expansion by ≥20%;

[0100] Data sharing and supervision module: Enables data exchange with departments such as agriculture and rural affairs, meteorology, finance, and financial supervision, establishes a cross-verification mechanism for underwriting and claims data, and supports the supervision of fiscal subsidies, reinsurance arrangements, and management of disaster risk reserves;

[0101] User interaction module: Provides online insurance application, claim reporting, inquiry, consultation and complaint functions, integrates AI intelligent consultation and human service, and the complaint response time is ≤7 working days.

[0102] In one embodiment, the air-space-ground multi-source data acquisition module includes:

[0103] Monitoring the situation and making preliminary assessments of the scope of disasters (such as areas affected by floods, droughts, and hail).

[0104] Workflow: Satellite imagery is automatically acquired daily, and information such as plot boundaries and crop coverage is extracted through image interpretation algorithms. This information is then compared and calibrated with land title confirmation data to generate a standardized spatial dataset. The update cycle is ≤24 hours.

[0105] UAV aerial survey submodule:

[0106] Technical parameters: Aerial survey image resolution ≤ 2 meters, plot vector map format is Shapefile or GeoJSON, boundary error within ±0.5 meters, supports manual / automatic flight path planning.

[0107] Workflow: For growers with more than 30 mu or randomized plots, aerial survey tasks are automatically generated. Drones fly along the flight path to collect images. AI algorithms are used to delineate plot boundaries, verify planting scale, and simultaneously extract crop growth status (such as seedling emergence rate and plant height during the jointing stage). Image data is uploaded to the system in real time.

[0108] Terrestrial IoT Sub-module:

[0109] Technical parameters: Deployment density ≥ 1 unit / 200 mu, including soil moisture meter (monitoring soil moisture and temperature), small weather station (monitoring precipitation, wind speed and air temperature), data transmission delay ≤ 30 seconds, sampling frequency is 1 time / hour.

[0110] Workflow: The equipment collects environmental data in real time and uploads it to the system via 4G / 5G network. After data cleaning and standardization, it is stored in the database to provide real-time data support for risk warning and loss assessment.

[0111] In one embodiment, the precise underwriting module incorporates a corn planting rights database, which integrates land ownership confirmation data from the agricultural and rural affairs departments with data from the Third National Land Survey ("Third Survey"). It uses facial recognition technology to verify the insured's identity and implements tiered verification based on the insured area.

[0112] When the insured area is less than 10 mu, the verification rate shall be ≥2% and the sample size shall be no less than 3 households, and satellite image verification shall be combined with random on-site inspection.

[0113] When the insured area is between 10 mu and 30 mu, the verification rate is ≥5%, and the top 3 households with the largest insured area are subject to mandatory verification.

[0114] When the insured area is ≥30 mu, drone aerial surveying and marker verification shall be mandatory, with a verification rate of 100%.

[0115] In the case of individual investors applying for insurance, the verification efficiency is ≤3 working days / household. The verification images must include a distant view of the insured object (with geographical identifier), a close-up view (corn crop characteristics), and a group photo of the insured, and be uploaded to the system for archiving in a timely manner.

[0116] In one embodiment, the precise claims module incorporates a dynamic claims model for the corn growth period, which divides the corn growth cycle into four stages: seedling stage, jointing stage, tasseling and silking stage, and grain-filling and maturity stage.

[0117] For seedlings facing risks of drought and continuous rain, a loss rate threshold of ≤30% is set, with a compensation coefficient of 0.5. Soil moisture data combined with visible light images is used to verify the loss.

[0118] During the jointing stage, in response to the risks of waterlogging and lodging, and drought, a loss rate threshold of 20%-50% is set, with a compensation coefficient of 0.6, and multispectral remote sensing data is used to verify the loss.

[0119] During the male-eating and silking period, in response to the risks of high temperature and drought, a loss rate threshold of ≥20% is set, with a compensation coefficient of 0.8, and the loss is verified by using drone inspection data.

[0120] During the grain filling and ripening period, in response to the risks of continuous rain and ear sprouting, a loss rate threshold of ≥10% is set, with a compensation coefficient of 1.0. Losses are verified by combining IoT humidity sensor data with satellite thermal infrared imagery.

[0121] After receiving the air-ground collaborative survey data, the AI ​​model generates a corn loss assessment report within 4 hours; for total loss cases (loss rate ≥ 80%), the assessment is completed within 10 days, and for partial loss cases (20% ≤ loss rate < 80%), the assessment is completed within 20 days. The compensation model parameters are shown in the table below.

[0122]

[0123] In one embodiment, the risk reduction module classifies early warning into three levels based on the probability of disaster occurrence:

[0124] Blue alert: The probability of disaster occurrence is ≥30%. Initiate the risk assessment process and complete the rectification of potential hazards in corn planting areas within 72 hours.

[0125] Yellow alert: The probability of disaster occurrence is ≥50%. Develop a targeted disaster prevention and control plan within 48 hours, including corn field management measures and emergency preparedness.

[0126] Red Alert: The probability of a disaster occurring is ≥70%. The emergency response plan for major disasters will be activated, and a 24-hour duty response will be implemented.

[0127] The warning standards and response guidelines are shown in the table below:

[0128]

[0129] For severe convective weather, the advance warning time is ≥40 minutes. Warning information and corn disaster prevention guidance are pushed to policyholders via SMS and system APP. No less than 2 disaster prevention technology training sessions are conducted every year, with a farmer training coverage rate of ≥90%. The investment in pre-disaster prevention funds is ≥1.5% of the annual premium income.

[0130] In one embodiment, the data sharing and supervision module uses JSON format to transmit cross-departmental data and encrypts sensitive information such as policyholder identity information and fiscal subsidy data using the SM4 national cryptographic algorithm. It achieves real-time sharing of 12 types of core metadata, including corn planting area, land ownership information, disaster early warning data, insured list, claims list, fiscal subsidy payment records, reinsurance agreement information, major disaster risk reserve provision data, corn growth monitoring data, disaster loss assessment reports, farmer complaint records, and system operation logs. The cross-departmental data query response speed is ≤30 seconds / time, and an automatic cross-verification mechanism for underwriting and claims data is established. The verification dimensions include consistency between insured area and land ownership area, matching of claims loss with disaster scope, and suitability of fiscal subsidy amount with insured scale.

[0131] Implementation steps

[0132] System deployment and data integration:

[0133] Deploy multi-source data acquisition equipment from air, space, and ground, and complete satellite remote sensing interface docking, drone deployment, and IoT device installation and commissioning (density ≥ 1 unit / 200 mu);

[0134] Data interfaces with departments such as agriculture and rural affairs, meteorology, finance, and natural resources have been established to achieve real-time sharing of 12 types of core data and complete data standardization processing.

[0135] Precise underwriting implementation:

[0136] Farmers can purchase insurance online through an app, uploading their identity information and land parcel information. The system verifies their identity through biometrics and automatically matches the land ownership data.

[0137] The system generates verification tasks based on the insured area, and drones conduct aerial surveys and land surveys for farmers with more than 30 mu of land, automatically completing verification and risk assessment.

[0138] For collective insurance cases, a list of individual households is generated, and a 7-day public notice period is conducted. After no objections are raised, the insurance is underwritten and the policy is issued, with the electronic policy automatically pushed to the farmers.

[0139] Risk reduction implementation:

[0140] The system monitors meteorological and IoT data in real time, issues three-level warnings (blue, yellow, and red), and pushes corn disaster prevention guidance (such as frost protection and stripe rust prevention) via SMS.

[0141] We organize two disaster prevention training sessions each year, covering ≥90% of insured farmers, and provide targeted field management advice.

[0142] Accurate claims processing and disaster prevention and mitigation implementation:

[0143] After farmers are affected by the disaster, they report the incident through an app. The system responds within one hour and initiates preliminary satellite remote sensing and precise drone surveys.

[0144] The AI ​​model combines multi-source data to generate a loss assessment report, which is publicized for 3 days and the compensation is paid to the farmer's social security card account within 3 days.

[0145] Agricultural experts were dispatched to the disaster site to guide farmers in self-rescue efforts such as drainage and pest and disease control, and to monitor secondary disasters after the disaster to prevent further losses.

[0146] Regulation and optimization:

[0147] Government departments monitor the entire underwriting and claims process through the system and verify the use of government subsidies;

[0148] At the end of the year, the system automatically compiles disaster loss and business data, optimizes risk maps and compensation models, and adjusts rate mechanisms.

[0149] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in the present invention, and these should all be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0150] (2) Figure 2 As shown, this embodiment of the invention provides an integrated intelligent system for corn planting income insurance based on multi-source data from air, space, and ground, incorporating the following modules:

[0151] Multi-source data acquisition module (space, air, and ground): used to acquire satellite remote sensing data, UAV refined monitoring data, ground intelligent sensing data, meteorological monitoring data, and multi-dimensional government and market data throughout the entire corn planting cycle;

[0152] Multi-source data preprocessing and maize growth coupling and fusion module: Cleans, denoises, and aligns the collected data in time and space. It achieves deep coupling of heterogeneous data by fusing maize growth model (DSSAT) and extracts key features related to maize growth and income.

[0153] The dual-precision core business modules include a precise underwriting sub-module and a precise claims settlement sub-module, which respectively realize corn variety identification, land ownership confirmation, dynamic income accounting, risk level pre-assessment, as well as precise disaster location, differentiated loss assessment during the growing season, and income loss accounting;

[0154] The two-reduction coordinated intervention module includes a risk reduction module and a disaster prevention and loss reduction sub-module, which provides graded early warning and intervention guidance for risks (diseases, lodging, drought, etc.) during the key growth period of maize, so as to achieve risk reduction and disaster prevention and loss reduction.

[0155] Digital Twin and Blockchain Trust Assurance Module: Construct a digital twin of corn planting, store key data (yield forecast, price data, loss assessment report) on the blockchain, and realize automated underwriting and claims settlement through smart contracts;

[0156] Multi-terminal collaborative interaction module: Provides customized interactive interfaces for farmers, insurance institutions, and government departments, supporting business processing, early warning reception, data query, and decision analysis.

[0157] In one embodiment, the air-space-ground multi-source data acquisition module includes:

[0158] Satellite remote sensing submodule: Acquires multispectral and SAR data from Gaofen satellites (Gaofen-6 and Sentinel-2) with a resolution ≤10m, for cornfield boundary identification, crop growth monitoring (NDVI / EVI2 index), and large-scale disaster (flood, drought) identification;

[0159] The UAV-based precision monitoring submodule is equipped with a hyperspectral camera, thermal imager, and lidar to acquire centimeter-level images and identify corn varieties, ear development status, lodging degree, and localized disease and pest patches (great leaf spot, corn borer).

[0160] Ground-based intelligent sensing submodule: Sensor nodes are deployed at a density of 30 mu / node, integrating sensors for soil temperature and humidity, nitrogen, phosphorus and potassium nutrients, plant stem thickness, and field cameras, focusing on collecting data during the corn jointing stage, large trumpet stage, and grain filling stage;

[0161] Meteorological monitoring submodule: Integrates data from national meteorological stations and regional micro meteorological stations to obtain key meteorological factors for maize growth (temperature, precipitation, relative humidity, sunshine duration) and meteorological conditions for the prevalence of pests and diseases;

[0162] The government affairs and market multi-dimensional sub-module connects to data on corn planting registration, variety registration, and land rights confirmation from agricultural and rural departments, as well as real-time corn purchase prices and agricultural input prices from the National Grain Trade Center, regional purchasing markets, and e-commerce platforms.

[0163] In one embodiment, the multi-source data preprocessing and maize growth coupling and fusion module adopts a four-level architecture of "preprocessing - spatiotemporal alignment - growth model coupling - feature fusion":

[0164] Preprocessing: Outlier removal (3σ principle), image denoising (bilateral filtering), and data completion (GA-LSTM prediction) ensure data integrity ≥98%;

[0165] Spatiotemporal alignment: Based on the GIS latitude and longitude coordinate system and the maize growth period time axis, multi-source data are mapped to a unified spatiotemporal dimension;

[0166] Growth model coupling: The preprocessed data is input into the DSSAT maize growth model to calibrate key parameters such as leaf area index, biomass, and yield, and output dynamic growth status data.

[0167] Feature fusion: An attention-driven multimodal fusion algorithm is used to weight and fuse remote sensing growth features, sensing environment features, model growth features, and market price features to output the core feature vector of corn income insurance.

[0168] In one embodiment, the precise underwriting submodule includes:

[0169] Accurate land ownership determination for corn varieties and plots: Corn varieties are identified using UAV hyperspectral imagery and CNN algorithms (accuracy ≥ 95%). Combined with satellite remote sensing and land ownership data, plot boundaries are automatically extracted and planting area is calculated (error ≤ 2%).

[0170] Dynamic Revenue Accounting: Construct a revenue model of "production forecast × real-time price integration". Production forecast is based on the fused feature vector + XGBoost algorithm (prediction accuracy ≥92%). Price integrates real-time data at the national, regional and market levels and dynamically updates the insured amount benchmark.

[0171] Risk level pre-assessment: Integrating the disaster resistance of corn varieties, historical disaster data, plot soil conditions, and meteorological risk thresholds, the random forest algorithm is used to classify risks into three levels: high, medium, and low, and insurance rates are set differently (15% higher for high-risk areas and 10% lower for low-risk areas).

[0172] In one embodiment, the accurate claims processing submodule includes:

[0173] Precise disaster location: Through satellite remote sensing change detection (CVA algorithm) + drone field patrol, the system automatically identifies the type of corn disaster (lodging, drought, flood, pests and diseases), its location and scope of impact, with a response time of ≤24 hours;

[0174] Differentiated loss assessment based on growth period: Based on the disaster impact coefficient of maize growth period (0.3 for seedling stage, 0.6 for jointing stage, 1.0 for large trumpet stage, 1.2 for grain filling stage, and 0.8 for maturity stage), the U-Net semantic segmentation algorithm is used to calculate the damaged area, and the yield reduction rate is calculated in combination with the yield loss model (error ≤ 4%).

[0175] Revenue loss accounting: Covers both "production loss type" and "price decline type" losses. The formula is "Revenue loss = (forecasted production - actual production) × average price + max(agreement price - actual price, 0) × actual production", and automatically generates a claim report.

[0176] In one embodiment, the two-reduction coordinated intervention module includes:

[0177] Risk reduction module: Based on risk level and corn growth stage, push differentiated prevention and control plans: high-risk areas push early warning once every 5 days, focusing on corn borer, large leaf spot and lodging risk; medium and low-risk areas push routine management suggestions according to growth stage;

[0178] Disaster prevention and mitigation submodule: Before a disaster, emergency prevention and control plans are pushed out (such as chemical control to prevent lodging before a typhoon, and water-saving irrigation before a drought); during a disaster, precise intervention measures are pushed out (such as targeted application of pesticides to pests and diseases); after a disaster, production recovery plans are pushed out (such as supporting seedlings and applying fertilizer after lodging, and foliar fertilizer spraying after drought during the grain-filling period), with a disaster prevention and mitigation rate of ≥18%.

[0179] In one embodiment, the digital twin and blockchain trust assurance module includes:

[0180] Digital twin: Construct a digital twin model of the entire corn planting cycle to map the plot status, growth progress, and risk situation in real time, and support disaster loss simulation and intervention effect prediction;

[0181] Blockchain evidence storage: Adopting a consortium blockchain architecture (insurance institutions + agricultural and rural departments + grain purchasing enterprises + farmers), the results of variety identification, yield forecast data, price data, and loss assessment reports are uploaded to the blockchain to ensure immutability and traceability;

[0182] Smart contracts: Preset rules for underwriting effectiveness, loss assessment triggering, and claims payment, supporting automatic claims triggering for price decline losses, shortening the claims cycle to within 2 business days.

[0183] In one embodiment, the multi-terminal collaborative interaction module includes a farmer's APP, an insurance institution's intelligent platform, and a government regulatory and decision-making platform.

[0184] Farmer APP: Supports variety registration, insurance payment, risk warning reception, field management video guidance, claims application and progress inquiry, and is compatible with voice interaction;

[0185] Insurance institution intelligent platform: supports underwriting review, loss assessment review, claims calculation, risk data analysis, policy management, and supports automatic generation of business reports;

[0186] Government regulatory and decision-making platform: Supports compliance supervision of insurance business, analysis of the corn industry situation (output, price, and disaster impact), and assessment of food security.

[0187] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in the present invention, and these should all be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A space-ground multi-source data-based corn planting income insurance "double-precision two-reduction" technology integrated intelligent system, comprising the following modules: Space-ground multi-source data acquisition module: used for acquiring satellite remote sensing data, unmanned aerial vehicle fine monitoring data, ground intelligent sensing data, meteorological monitoring data, government affairs and market multi-dimensional data of the whole cycle of corn planting; Multi-source data preprocessing and corn growth coupling fusion module: the collected data is cleaned, denoised and time and space aligned, the deep coupling of heterogeneous data is realized through the fusion of corn growth model (DSSAT), and the key features related to corn growth and income are extracted; Double-precision core business module: containing precision underwriting sub-module and precision claim settlement sub-module, respectively realizing corn variety identification, land right confirmation, income dynamic accounting, risk grade pre-evaluation, and disaster point precise positioning, growth period differentiation loss determination, income loss accounting; Two-reduction collaborative intervention module: containing risk reduction sub-module and disaster prevention and loss reduction sub-module, providing hierarchical warning and intervention guidance for corn key growth period risks (disease and pest, lodging, drought, etc.), realizing risk reduction and disaster prevention and loss reduction; Digital twin and blockchain trusted guarantee module: constructing corn planting digital twin, storing key data (yield prediction, price data, loss determination report) on chain, and realizing insurance, claim settlement automation through smart contract; Multi-terminal collaborative interaction module: providing customized interaction interface for farmers, insurance agencies and government departments, supporting business handling, early warning receiving, data query and decision analysis. Its characteristics are: Space-ground multi-source data acquisition module: integrating satellite remote sensing, unmanned aerial vehicle surveying and mapping, ground Internet of Things equipment, collecting corn planting area spatial data, growth state data and environmental monitoring data according to the CGCS2000 national geodetic coordinate system standard, and outputting standardized data set; Precision underwriting module: realizing insurance information verification, right verification, intelligent mark verification, risk rating and underwriting publicity based on multi-source data, meeting the technical requirements of online insurance rate ≥ 90%, biological identification identity verification coverage 100%, data error rate ≤ 3%; Precision claim settlement module: through intelligent case receiving, space-ground collaborative investigation, AI dynamic loss determination, automatic calculation and payment of compensation, realizing 24-hour investigation response, loss determination within 10 working days, compensation payment within 3 days, loss determination accuracy ≥ 95%; Risk reduction module: based on multi-source data such as meteorology and plant protection, a three-level early warning system is constructed to carry out pre-disaster prediction and warning, disaster prevention technology training and risk investigation, with a warning accuracy rate of ≥ 90% and a high-risk area inspection cycle of ≤ 48 hours; Disaster prevention and loss reduction module: after the disaster occurs, the collaborative response is started, the unmanned aerial vehicle inspection and agricultural technology guidance are realized to reduce the loss in the disaster, the loss is accurately determined and the secondary disaster is monitored to realize the post-disaster loss reduction, and the loss expansion ratio is reduced by ≥ 20%; Data sharing and supervision module: realizing data interconnection with departments of agriculture and rural areas, meteorology, finance, gold supervision, etc., establishing cross-verification mechanism of insurance and claim settlement data, supporting financial subsidy supervision, reinsurance arrangement and management of large disaster risk reserve; User interaction module: provides online insurance, report, inquiry, consultation and complaint functions, integrates AI intelligent consultation and manual service, and the complaint response time is ≤7 working days.

2. The "double precision two reduction" technical integrated intelligent system based on space-earth multi-source data for corn planting income insurance according to claim 1, the space-earth multi-source data acquisition module comprises: Satellite remote sensing sub-module: obtain multispectral and SAR data of high-resolution satellites (Gao Fen No. 6 and Sentinel No. 2) with a resolution of ≤10 m, which are used for corn plot boundary identification, growth monitoring (NDVI / EVI2 index), and large-scale disaster (flood, drought) identification; Unmanned aerial vehicle fine monitoring sub-module: carry high-spectral cameras, thermal imagers, and laser radars to obtain centimeter-level images and identify corn varieties, ear development state, lodging degree, and local disease and pest (large spot disease, corn borer) patches; Ground intelligent sensing sub-module: deploy sensor nodes at a density of 30 mu per node, integrate soil temperature and humidity, nitrogen, phosphorus, and potassium nutrient sensors, and plant stem thickness sensors and field cameras, and focus on collecting data at the corn jointing stage, large trumpet stage, and grain filling stage; Weather monitoring sub-module: integrate national and regional miniature weather station data to obtain key weather factors (temperature, precipitation, relative humidity, and sunshine duration) for corn growth and disease and pest prevalence weather conditions; Government and market multi-dimensional sub-module: connect corn planting records, variety registration, and land right data of the agricultural and rural departments, and corn real-time purchase prices and agricultural material prices of the national grain trading center, regional purchase markets, and e-commerce platforms. Its characteristics are that the space-earth multi-source data acquisition module comprises: Potential monitoring and disaster range preliminary judgment (such as flood and drought affected areas). Workflow: automatically obtain satellite images every day, extract plot boundaries and crop coverage through image interpretation algorithms, compare and calibrate with land right data, generate standardized spatial data sets, and update every 24 hours. Unmanned aerial vehicle aerial survey sub-module: Technical parameters: aerial survey image resolution ≤2 meters, plot vector graph format Shapefile or GeoJSON, boundary error within ±0.5 meters, and support for manual / automatic route planning. Workflow: automatically generate aerial survey tasks for corn growers with more than 30 mu or randomly selected plots, unmanned aerial vehicles fly along the route to collect images, and AI algorithms are used to draw plot boundaries, verify planting scale, and synchronously extract crop growth state (such as seedling emergence rate in the seedling stage and plant height in the jointing stage), and image data is uploaded to the system in real time. Ground Internet of Things sub-module: Technical parameters: deployment density ≥1 per 200 mu, including soil moisture and temperature sensors and small weather stations (monitoring precipitation, wind speed, and temperature), data transmission delay ≤30 seconds, and sampling frequency 1 per hour. Workflow: devices collect environmental data in real time, upload them to the system through 4G / 5G network, and after data cleaning and standardization, store them in the database to provide real-time data support for risk early warning and loss assessment.

3. The "double precision and two reduction" technology integrated intelligent system for corn planting income insurance based on space, sky, and multi-source data according to claim 1, wherein the multi-source data preprocessing and corn growth coupling fusion module adopts a four-level architecture of "preprocessing-time and space alignment-growth model coupling-feature fusion": Preprocessing: ensure data integrity ≥98% through outlier rejection (3σ principle), image denoising (bilateral filtering), and data completion (GA-LSTM prediction); Time and space alignment: map multi-source data to a unified time and space dimension based on GIS latitude and longitude coordinate system and corn growth period time axis; Growth model coupling: input the preprocessed data into the DSSAT corn growth model, calibrate key parameters such as leaf area index, biomass, and yield, and output dynamic growth state data; Feature fusion: use attention mechanism driven multi-modal fusion algorithm to weight and fuse remote sensing growth features, environmental features, model growth features, and market price features, and output corn income insurance core feature vector. characterized in that The precise underwriting module has a corn planting ownership database, which integrates land right data from the agricultural and rural departments and the third national land survey ("three surveys") data; the identity verification of the insured is completed through face recognition technology, and the verification of the insured is implemented according to the insured area classification: When the insured area is less than 10 mu, the verification ratio is ≥2% and the sample size is not less than 3 households, satellite image verification combined with random on-site inspection is adopted; When the insured area is 10-30 mu, the verification ratio is ≥5%, and the top 3 households with the largest insured area are forced to verify the mark; When the insured area is ≥30 mu, unmanned aerial vehicle aerial survey is used to verify the land, and the verification ratio is 100%; In the case of individual insured, the verification efficiency is ≤3 working days / household, the verification image needs to include the mark long shot (with geographical identification), close-up (corn crop features), and insured photo, and needs to be uploaded to the system for archiving in real time.

4. The "double precision and two reduction" technology integrated intelligent system for corn planting income insurance based on space, sky, and multi-source data according to claim 1, wherein the precise underwriting submodule comprises: Corn variety and plot precise right: identify corn varieties through unmanned aerial vehicle hyperspectral image + CNN algorithm (identification accuracy ≥95%), combine satellite remote sensing and land right data, automatically extract plot boundaries, and calculate planting area (error ≤2%); Income dynamic calculation: build a "yield prediction × price real-time integration" income model, yield prediction is based on fused feature vector + XGBoost algorithm (prediction accuracy ≥92%), price integration is based on real-time data at national, regional, and market levels, and insurance amount benchmark is dynamically updated; Risk level pre-evaluation: fuse corn variety disaster resistance, historical disaster data, plot soil conditions, and meteorological risk threshold, and use random forest algorithm to divide high, medium, and low risk levels, and set insurance rates differently (high risk area increases by 15%, low risk area decreases by 10%). The precise claim settlement module has a corn growth period dynamic settlement model, which is divided into four stages of seedling stage, jointing stage, tasseling and silk stage, and filling and maturation stage according to the corn growth cycle. For drought and continuous rain risk in the seedling stage, the loss rate threshold is set to ≤30%, the compensation coefficient is 0.5, and the soil moisture content instrument data combined with visible light image are used to verify the loss; For waterlogging and drought risk in the jointing stage, the loss rate threshold is set to 20%-50%, the compensation coefficient is 0.6, and multispectral remote sensing data are used to verify the loss; For high temperature and drought risk in the silk stage, the loss rate threshold is set to ≥20%, the compensation coefficient is 0.8, and unmanned aerial vehicle inspection data are used to verify the loss; For continuous rain and ear germination risk in the filling and maturing stage, the loss rate threshold is set to ≥10%, the compensation coefficient is 1.0, and the Internet of Things humidity sensor data combined with satellite thermal infrared image are used to verify the loss; After receiving the space-ground-land collaborative survey data, the AI model generates a corn loss assessment report within 4 hours; the full loss case (loss rate ≥80%) is completed within 10 days, and the partial loss case (20%≤loss rate<80%) is completed within 20 days.

5. The "double precision two reduction" technical integrated intelligent system for corn planting income insurance based on space-ground-land multi-source data according to claim 1, wherein the precise claim settlement submodule comprises: Precise disaster point positioning: through satellite remote sensing change detection (CVA algorithm) + unmanned aerial vehicle field inspection, automatically identify the type (lodging, drought, flood, disease and pest), location and influence range of corn disasters, and the response time is ≤24 hours; Differentiated loss determination in growth stages: based on the corn growth stage disaster influence coefficient (0.3 in seedling stage, 0.6 in jointing stage, 1.0 in large trumpet stage, 1.2 in filling stage, and 0.8 in maturing stage), the U-Net semantic segmentation algorithm is used to calculate the damaged area, and the yield loss model is used to calculate the yield reduction rate (error ≤4%); Income loss accounting: covering both "yield loss type" and "price drop type" double losses, the formula is "income loss=(forecast yield-actual yield)×average price+max (contract price-actual price, 0)×actual yield", and the claim settlement report is automatically generated. Its characteristics are that the risk reduction module divides the disaster occurrence probability into three levels of early warning: Blue warning: disaster occurrence probability ≥30%, start risk investigation process, complete corn planting area hidden danger rectification within 72 hours; Yellow warning: disaster occurrence probability ≥50%, develop targeted disaster prevention countermeasures within 48 hours, including corn field management measures and emergency preparation; Red warning: disaster occurrence probability ≥70%, start disaster emergency plan, implement 24-hour on-duty response; For strong convective weather, the early warning time is ≥40 minutes, and the warning information and corn disaster prevention guidance are pushed to the policyholder through SMS and system APP, not less than 2 times of disaster prevention technical training is carried out every year, the farmer training coverage rate is ≥90%, and the disaster prevention fund input is ≥1.5% of the premium income of the current year.

6. The "double precision two reduction" technical integrated intelligent system for corn planting income insurance based on space-ground-land multi-source data according to claim 1, wherein the two reduction collaborative intervention module comprises: Risk reduction sub-module: Based on the risk level and the growth period of corn, push differentiated prevention and control scheme: high-risk area push warning every 5 days, focus on corn borer, large spot disease, lodging risk; Medium and low risk areas push routine management recommendations according to growth period; Disaster prevention and reduction sub-module: Push emergency prevention and control scheme before disaster (such as typhoon prevention and control before control, water-saving irrigation before drought); Push precise intervention measures during disaster (such as targeted pesticide application); Push recovery scheme after disaster (such as post-lodging seedling and fertilizer application, drought after grain filling period), and the disaster prevention and reduction rate is ≥18%. Its characteristics are: the data sharing and supervision module transmits cross-department data in JSON format, and encrypts sensitive information such as the identity information of the insured and financial subsidy data through SM4 national encryption algorithm; Real-time sharing of 12 types of core metadata, including corn planting area, land right information, disaster warning data, insurance list, claim list, financial subsidy disbursement record, reinsurance agreement information, large disaster risk reserve data, corn growth monitoring data, disaster loss assessment report, farmer complaint record, and system operation log; The cross-department data query response speed is ≤30 seconds / time, an automatic cross-verification mechanism for insurance and claim data is established, and the verification dimensions include the consistency of insured area and right area, the matching of claim loss and disaster range, and the adaptability of financial subsidy amount and insurance scale.

7. The "double-precision two-reduction" technology integrated intelligent system for corn planting income insurance based on space-air-ground multi-source data according to claim 1, characterized in that: The digital twin and blockchain trusted security module includes: Digital twin: Construct a digital twin model of the whole cycle of corn planting, which can real-time map the state of the land, the growth progress and the risk situation, and support disaster loss simulation and intervention effect prediction; Blockchain storage: Using a consortium chain architecture (insurance agencies + agricultural and rural departments + grain purchasing enterprises + farmers), the variety identification results, yield prediction data, price data, and loss assessment reports are uploaded to the chain to ensure that they are tamper-proof and traceable; Smart contract: Pre-set insurance effectiveness, loss assessment triggering, and claim payment rules, support automatic triggering of claims for price decline type loss, and shorten the claim period to within 2 working days.

8. The "double-precision two-reduction" technology integrated intelligent system for corn planting income insurance based on space-air-ground multi-source data according to claim 1, characterized in that: The multi-terminal collaborative interaction module includes farmer APP, insurance agency intelligent platform, government supervision and decision-making platform: Farmer APP: Supports variety registration, insurance payment, risk warning reception, field management video guidance, claim application and progress inquiry, and adapts to voice interaction; Insurance agency intelligent platform: Supports insurance audit, loss assessment review, claim accounting, risk data analysis, and policy management, and supports automatic generation of business reports; Government supervision and decision-making platform: Supports insurance business compliance supervision, corn industry situation analysis (yield, price, disaster impact), and food security research and judgment.

Images