An intelligent management system for agro-pastoral ecological hydrology based on cloud edge-end cooperation

Abstract

The application discloses a kind of based on cloud edge end coordination's agricultural and pastoral ecological hydrology intelligent management system, it is related to ecological hydrology management technical field, and perception layer is collected multi-source data by space-based, air-based and ground-based;Transmission layer transmits multi-source data to edge layer by data transmission and priority scheduling mechanism;Edge layer realizes local data analysis and rapid response based on multi-source data, and multi-source data is sent to cloud platform;Cloud platform integrates digital twin engine, model library and blockchain fusion module, provides comprehensive data processing, simulation prediction and intelligent decision support;Decision layer is based on the deep analysis result of cloud platform, adopts multi-objective decision optimization algorithm to optimize the balance between the income of animal husbandry, ecological water demand and net ecosystem productivity, and the execution of trigger ecological compensation mechanism, generates multi-dimensional decision instruction, realizes the perception-simulation-decision closed-loop control of ecological hydrology whole process.

Description

Technical Field

[0001] This invention relates to the field of ecological hydrological management technology, and more specifically to an intelligent management system for agricultural and pastoral ecological hydrology based on cloud-edge-device collaboration. Background Technology

[0002] With the intensification of global climate change and human activities, the coupling relationship between hydrological processes and ecosystems in the agro-pastoral ecotone, as an ecologically fragile and resource-sensitive area, is becoming increasingly complex. Traditional eco-hydrological management models are no longer sufficient to meet the needs of efficient water resource allocation, ecological protection, and coordinated production development in this region. Currently, the information-based management of eco-hydrology in the agro-pastoral ecotone faces the following key bottlenecks:

[0003] First, there is a serious "data silo" problem in the data acquisition stage. Existing monitoring methods mainly include satellite remote sensing, UAV patrols, and ground sensors. However, due to the diverse sources of equipment, heterogeneous data formats, and inconsistent communication protocols, it is difficult to achieve standardized access and fusion analysis of multi-source data, which seriously restricts information sharing and system integration.

[0004] Secondly, the computing processing method suffers from high response latency. Currently, most platforms still rely on centralized cloud computing centers for data processing and model calculations, lacking the localized intelligent processing capabilities of edge nodes. In scenarios requiring rapid response, such as flood warnings (e.g., response time less than 2 hours), traditional architectures often fail to meet real-time requirements, impacting decision-making timeliness and emergency response capabilities.

[0005] Secondly, there are issues of "fragmentation and reliance on experience" in decision support. There is a lack of scientific and quantitative balancing mechanisms between ecological protection and livestock production, and a lack of intelligent decision-making models based on multi-objective optimization. This makes it difficult for management strategies to balance ecological and economic benefits, resulting in a highly subjective and poorly adaptable decision-making process.

[0006] Therefore, how to provide an intelligent management system for agricultural and pastoral ecological hydrology based on cloud-edge-device collaboration is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0007] In view of this, the present invention provides an intelligent management system for agricultural and pastoral ecological hydrology based on cloud-edge-device collaboration. By constructing a multi-level data fusion mechanism, dynamic resource scheduling strategy and intelligent model inference system under the "cloud-edge-device" collaborative architecture, it realizes closed-loop control of perception-simulation-decision for the entire process of ecological hydrology, thereby improving the efficiency of water resource utilization and the ability to ensure ecological security in arid / semi-arid regions.

[0008] To achieve the above objectives, the present invention adopts the following technical solution:

[0009] A cloud-edge-device collaborative intelligent management system for agricultural and pastoral ecological hydrology includes: an ecological hydrological information management module and an ecological hydrological source tracing and early warning application module; the ecological hydrological information management module includes a sensing layer, a transmission layer and an edge layer; the ecological hydrological source tracing and early warning application module includes a cloud platform and a decision-making layer;

[0010] The sensing layer collects multi-source data through space-based, air-based, and ground-based systems;

[0011] The transport layer transmits multi-source data to the edge layer through data transmission and priority scheduling mechanisms;

[0012] The edge layer enables local data analysis and rapid response based on multi-source data, and sends multi-source data to the cloud platform;

[0013] The cloud platform integrates a digital twin engine, model library, and blockchain fusion module, providing comprehensive data processing, simulation prediction, and intelligent decision support.

[0014] Based on the in-depth analysis results from the cloud platform, the decision-making level uses a multi-objective decision optimization algorithm to optimize the balance between livestock income, ecological water demand, and net ecosystem productivity, as well as trigger the execution of the ecological compensation mechanism, and generate multi-dimensional decision instructions.

[0015] Preferably, the perception layer includes:

[0016] Space-based: Integrates different types of satellites to acquire daily data on Normalized Difference Vegetation Index (NDVI), Enhanced Difference Vegetation Index (EVI), groundwater storage, and precipitation forecasts;

[0017] Airborne: A drone swarm is configured, equipped with multispectral sensors, LiDAR, and thermal infrared sensors. Among them, the multispectral sensors are used to invert and process biomass density, the LiDAR is used to obtain runoff paths, waterlogged areas, and gully development through point cloud analysis, and the thermal infrared sensors are used to obtain grazing activity index and grazing intensity index. Based on the data obtained above, a high-precision vegetation cover heat map is generated in real time to optimize grazing paths.

[0018] Foundation: Deploy IoT terminals, including a soil multi-parameter instrument and a smart collar for livestock. The soil multi-parameter instrument is used to collect soil moisture content and monitor surface parameters, while the smart collar for livestock is used to monitor livestock feeding behavior and GPS movement trajectory.

[0019] Preferably, the edge layer includes:

[0020] At the edge of the region, a high-performance server cluster is set up to run the SWAT-HS hydrological model, simulate watershed-scale water cycle and generate ecological security index.

[0021] The edge gateway of the site runs a TensorFlow Lite model to calculate the carrying capacity of grassland in real time.

[0022] Preferably, the cloud platform layer includes:

[0023] The digital twin engine integrates multi-source data through 3D modeling, supports dynamic parameter injection and real-time rendering, and is used to simulate ecological responses under different scenarios.

[0024] The model library contains hydrological models, ecological models, and AI models for use by the edge layer and decision-making layer.

[0025] The blockchain integration module utilizes Hyperledger Fabric to build a consortium blockchain for blockchain notarization and supports an ecosystem compensation mechanism.

[0026] Preferably, the hydrological models include the SWAT-HS hydrological model and the HEC-RAS flood simulation model;

[0027] Ecological models include the CENTURY grassland carbon cycle model and the TensorFlow Lite model;

[0028] AI models are used to optimize grazing routes based on high-precision vegetation cover heat maps.

[0029] Preferably, the Ecological Security Index (ESI) is calculated using the SWAT-HS hydrological model:

[0030]

[0031] in, For soil moisture content, For soil moisture critical threshold, Normalized Difference Vegetation Index (NDVI) The normalized vegetation index baseline value, For groundwater reserves, Average groundwater storage over many years, weighting coefficient ;

[0032] Run a TensorFlow Lite model to calculate the pasture carrying capacity C in real time:

[0033]

[0034] Where A represents grassland area and BD represents biomass density. The coefficient of recovery, This refers to the daily food intake. The number of days for grazing.

[0035] The preferred ecological compensation mechanism is as follows:

[0036] When the Ecological Security Index (ESI) of a certain grassland is greater than 0.8 for three consecutive months, compensation is triggered.

[0037]

[0038] in, For grassland ecological compensation funds. Let A be a coefficient, and A be the grassland area. This refers to the price of carbon sinks.

[0039] The preferred multi-objective optimization function is:

[0040]

[0041] in, For the benefit of the livestock economy, For ecological water demand, For net ecosystem productivity, For ecological security index, The carrying capacity of grassland. The maximum carrying capacity of the grassland. Soil moisture content, This represents the critical threshold for soil moisture. , , , , The weights are denoted by , and ESI stands for Ecological Security Index.

[0042] Preferably, the decision-making level also includes an AR interactive terminal.

[0043] As can be seen from the above technical solution, compared with the prior art, the present invention discloses a cloud-edge-device collaborative intelligent management system for agricultural and pastoral ecological hydrology, which has the following advantages:

[0044] 1) Data integration and breaking down "data silos":

[0045] By integrating multi-source data, including space-based (satellite imagery), air-based (LiDAR point cloud), and ground-based (sensor data), the problem of heterogeneous data formats and lack of unified access standards among traditional monitoring equipment has been solved. This allows data from different sources to be analyzed on the same platform, improving the comprehensiveness and accuracy of data analysis.

[0046] 2) Real-time performance and low latency:

[0047] By adopting 5G slicing network technology and a hierarchical processing architecture, dedicated bandwidth is provided for emergency data (such as flood warnings and ecological disasters). This means that edge-layer local computing avoids the network jumps of traditional cloud center processing, reduces the data transmission distance, and thus ensures the rapid transmission of information. The response time is less than 50ms, which is far lower than the response time of traditional cloud center processing mode (>2 hours), meeting the needs of real-time decision-making.

[0048] 3) Intelligent model inference and optimization decision support:

[0049] Integrating various advanced eco-hydrological analysis models (such as SWAT-HS, HEC-RAS, CENTURY, TensorFlowLite, etc.) and AI models (deep reinforcement learning, LSTM, etc.), it can simulate and predict eco-hydrological processes and generate optimization suggestions or early warning information. Furthermore, the application of the NSGA-II multi-objective genetic algorithm helps resolve the conflict between ecological protection and livestock production, achieving an optimal balance between economic and ecological benefits.

[0050] 4) Blockchain technology ensures data credibility and transparency:

[0051] Built on Hyperledger Fabric, this consortium blockchain uses smart contracts to automatically verify data integrity, ensuring data authenticity and immutability. Simultaneously, blockchain technology enables an automated ecological compensation mechanism, enhancing the transparency and efficiency of ecological compensation and incentivizing herders to participate in ecological protection.

[0052] 5) Flexibility and scalability:

[0053] The system design employs a modular structure, facilitating the addition of new sensors, models, or functional modules according to actual needs, demonstrating strong flexibility and scalability. Simultaneously, the "cloud-edge-device" collaborative architecture allows local computing tasks to be executed on edge nodes, reducing cloud load and improving overall system performance. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0055] Figure 1 This is a schematic diagram of the structure of an intelligent management system for agricultural and pastoral ecological hydrology based on cloud-edge-device collaboration, provided by the present invention.

[0056] Figure 2 The data processing flowchart provided for this invention. Detailed Implementation

[0057] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] This invention discloses an intelligent management system for agricultural and pastoral ecological hydrology based on cloud-edge-device collaboration, such as... Figure 1 and Figure 2 As shown, it includes: an eco-hydrological information management module and an eco-hydrological source tracing and early warning application module;

[0059] The eco-hydrological information management module includes a sensing layer, a transmission layer, and an edge layer; the eco-hydrological source tracing and early warning application module includes a cloud platform and a decision-making layer.

[0060] The sensing layer collects multi-source data through space-based, air-based, and ground-based systems;

[0061] The transport layer transmits multi-source data to the edge layer through data transmission and priority scheduling mechanisms;

[0062] The edge layer enables local data analysis and rapid response based on multi-source data, and sends multi-source data to the cloud platform;

[0063] The cloud platform integrates a digital twin engine, model library, and blockchain fusion module, providing comprehensive data processing, simulation prediction, and intelligent decision support.

[0064] At the decision-making level, based on the in-depth analysis results from the cloud platform, a multi-objective decision optimization algorithm is used to optimize the balance between livestock income, ecological water demand, and net ecosystem productivity, as well as to trigger the execution of the ecological compensation mechanism, and generate multi-dimensional decision instructions.

[0065] To facilitate a further understanding of the technical solution of this invention, a detailed description is provided below:

[0066] I. Perception Layer (End Device Layer)

[0067] ① Space-based sensing:

[0068] [1] Satellite network: integrating Sentinel-2 (optical), GRACE-FO (gravity), and FY-4A (weather), acquiring data daily:

[0069] 10m resolution vegetation indices (Normalized Difference Vegetation Index NDVI and Enhanced Difference Vegetation Index EVI).

[0070] Groundwater reserves (GWSA);

[0071] 15-minute precipitation forecast data (spatial resolution 4km);

[0072] [2] Data preprocessing: The vegetation index, groundwater storage and precipitation forecast data were radiometrically corrected and geometrically registered at the satellite ground station, and then transmitted directly to the cloud platform after the preliminary processing was completed at the satellite ground station.

[0073] ②Empty-based sensing:

[0074] [1] Drone swarm:

[0075] Aircraft configuration: Multi-rotor (DJI M300) for fixed-point scanning, fixed-wing (eBee X) for area surveying;

[0076] Sensors: Multispectral (RedEdge-MX) inversion processing acquires biomass density (BD) for real-time calculation of grassland carrying capacity (C) in the TensorFlowLite model and drives the CENTURY model to predict grassland carbon cycle simulation; LiDAR (Velodyne Puck) point cloud analysis acquires runoff paths, waterlogged areas, and gully development for running a simplified version of the HEC-RAS model to quickly simulate flood inundation range; Thermal infrared (FLIR Vue Pro) and multispectral fusion processing acquire grazing activity index (GAI) and grazing intensity index (GII) for AR interactive terminals to help herders intuitively identify compliant areas;

[0077] Edge processing: Real-time generation of vegetation cover heatmaps (5cm resolution) on the NVIDIA Jetson AGX mounted on the drone, used for deep reinforcement learning (AI model) to optimize grazing paths;

[0078] ③Ground sensing:

[0079] [1] Internet of Things (IoT) terminals:

[0080] Soil multi-parameter instrument: measures soil moisture content (FDR principle), temperature, and electrical conductivity (accuracy ±2%) from 0-100cm, and uploads data every 15 minutes;

[0081] Smart collar for livestock: integrates GPS / BeiDou dual-mode positioning, a three-axis accelerometer, and RFID tags to monitor livestock feeding behavior and GPS movement trajectory;

[0082] [2] Communication protocol: LoRaWAN+MQTT protocol is adopted, supporting low power wide area transmission (node ​​battery life > 3 years).

[0083] II. Transport Layer and Edge Computing

[0084] 1. Transport Layer

[0085] This invention uses 5G slicing network technology to build an efficient and low-latency communication channel to meet the data transmission needs of different business scenarios.

[0086] ① Dynamic bandwidth allocation mechanism:

[0087] Emergency channel (used for data with high real-time requirements such as precipitation warnings and ecological disasters): The transmission path is space-based - transmission layer - edge layer - cloud platform, with dedicated bandwidth resources to ensure end-to-end latency of less than 50ms, prioritizing emergency response.

[0088] For routine monitoring channels (such as daily monitoring data like soil moisture): the transmission path is ground-transmission layer-edge layer-cloud platform, sharing bandwidth, with latency controlled within 200ms, balancing stability and efficiency.

[0089] Historical data backhaul (such as logs and archived data): scheduled for transmission during off-peak network hours at night, bandwidth can be automatically adjusted according to the current load, without affecting other critical tasks.

[0090] ②QoS guarantee algorithm:

[0091] Traffic scheduling optimization is achieved based on Weighted Fair Queuing (WFQ).

[0092] Dynamically adjust the data priority weights of each channel to ensure that high-priority data (such as flood warnings) always receive the best transmission resources, while maintaining the overall system throughput and stability.

[0093] 2. Edge Node Design

[0094] ①Regional edge layer:

[0095] Hardware: Equipped with a high-performance server cluster, featuring NVIDIA A100 GPUs, supporting parallel computing and deep learning inference tasks;

[0096] Core Function: Runs the SWAT-HS hydrological model to simulate the watershed-scale water cycle (time step 1 hour) and generates the Ecological Security Index (ESI). Formula:

[0097]

[0098] in, For soil moisture content, For soil moisture critical threshold, Normalized Difference Vegetation Index (NDVI) The normalized vegetation index baseline value, For groundwater reserves, Average groundwater storage over many years, weighting coefficient The weights are dynamically adjusted based on principal component analysis (PCA).

[0099] ②Site Edge Gateway:

[0100] Hardware: The core device is a Raspberry Pi 4 (ARM Cortex-A72 quad-core processor, 4GB RAM) paired with an Intel Neural Compute Stick 2 accelerator, which has lightweight AI inference capabilities;

[0101] Core functionality: TensorFlow Lite model: Real-time calculation of pasture carrying capacity (C):

[0102]

[0103] Where A is the grassland area (ha) and BD is the biomass (kg / ha). The coefficient of recovery is (0-1). Daily feed intake (kg / sheep unit). The number of days for grazing.

[0104] Anomaly detection mechanism:

[0105] The Isolation Forest algorithm is used to identify abnormal patterns.

[0106] When a sudden drop in soil moisture exceeding 25% over 6 hours is detected, an alarm is triggered, and drones are activated to perform high-frequency scans to obtain more detailed surface information.

[0107] III. Cloud Platform

[0108] 1. Digital Twin Engine:

[0109] This invention constructs a digital twin of a watershed by fusing LiDAR point clouds (5cm accuracy), Sentinel-2 imagery, and geological data. Its function is to achieve a dynamic mapping between the physical watershed and the digital space through virtual modeling and real-time data-driven approaches, thereby supporting the simulation, prediction, and decision optimization of eco-hydrological processes. In dynamic simulations and scenario extrapolations, the digital twin engine can simulate eco-hydrological responses under different management strategies (grazing intensity, precipitation changes), specifically including: predicting grassland degradation trends, assessing flood inundation extent, quantifying changes in carbon sink potential, and calculating grazing carrying capacity. These simulation results provide managers with a scientific basis to help them formulate more effective ecological protection and resource utilization strategies. In multi-source data fusion, the digital twin engine unifies and integrates data from space-based (satellite imagery), air-based (LiDAR point clouds), and ground-based (sensor data) observations into a three-dimensional space, solving the problem of data silos. (Its main technical advantages are supporting spatiotemporal matching and data assimilation; for example, fusing UAV LiDAR terrain data with Sentinel-2 imagery to generate high-precision land use maps; combining soil moisture sensor data with the GRACE-FO groundwater model to analyze water cycle paths.) In terms of intelligent decision support, the digital twin engine provides real-time virtual visualization output, such as grazing boundaries and grazing routes, and supports dynamic parameter injection (such as precipitation scenario simulation, soil moisture and vegetation cover changes), enabling real-time rendering of eco-hydrological responses and enhancing the visualization and interactivity of decision-making.

[0110] 2. Model Library:

[0111] The model library integrates various eco-hydrological analysis models, providing algorithmic support for data-driven decision-making. These models are used for different analytical tasks, such as predicting hydrological changes, assessing grassland productivity, or optimizing grazing routes. The library serves as a centralized location for storing and managing these models, facilitating their retrieval and updates.

[0112] In subsequent steps, the model library may be used in the data processing phase, such as for in-depth analysis on cloud platforms or for real-time computation at edge nodes. For instance, when an edge gateway detects abnormal data, it may need to call upon lightweight models from the model library for rapid analysis. Furthermore, during the decision-making phase, models from the library can be used to generate optimization suggestions or early warning information.

[0113] The model library spans the entire agro-pastoral eco-hydrological network platform system, playing a crucial role, particularly in in-depth analysis of multi-source data, multi-model analysis of the water-grass-livestock coupled system, multi-scenario simulation and intelligent decision-making. Specific model types are as follows:

[0114] Hydrological models: SWAT-HS is used for watershed-scale water cycle simulation, and HEC-RAS is used for rapid simulation of flood inundation extent;

[0115] Ecological models: CENTURY is used for grassland carbon cycle simulation, and TensorFlow Lite is used for local grassland carrying capacity assessment.

[0116] AI models: Deep reinforcement learning (DRL) is used to optimize grazing routes, and LSTM is used to predict drought risk.

[0117] 3. Blockchain Integration Module

[0118] ① Data preservation:

[0119] [1] Chain structure: The consortium chain is built on Hyperledger Fabric, with nodes including government, research institutions, herder associations and other participants to ensure trusted data sharing and collaborative governance;

[0120] [2] Evidence preservation process:

[0121] Sensor data is uploaded to the blockchain after being hashed using SHA-256.

[0122] The data block header includes a timestamp, GPS coordinates, and device ID to ensure the data source is traceable.

[0123] Smart contracts automatically verify data integrity to prevent tampering and forgery.

[0124] ② Ecological compensation:

[0125] [1] The blockchain-based ecological compensation mechanism achieves automated execution through smart contracts:

[0126] When the ESI of a certain grassland is >0.8 for 3 consecutive months, compensation is triggered:

[0127]

[0128] in, For grassland ecological compensation funds. Let A be a coefficient, and A be the grassland area. The price is for carbon sinks (RMB / ton).

[0129] IV. Decision-making level

[0130] ① Multi-objective optimization:

[0131] To achieve a balance between ecological and economic benefits, this invention employs a multi-objective optimization method, constructing the following objective function:

[0132]

[0133] in, For the benefit of the livestock economy, For ecological water demand, For net ecosystem productivity, For ecological security index, The carrying capacity of grassland. The maximum carrying capacity of the grassland. Soil moisture content, This represents the critical threshold for soil moisture. , , , , The weights are denoted by , and ESI stands for Ecological Security Index.

[0134] The solution algorithm employs the NSGA-II multi-objective genetic algorithm, outputting a Pareto front solution set for decision-makers to select the optimal compromise. This method effectively addresses the difficulty of balancing multiple objectives in traditional empirical decision-making, enhancing the scientific rigor and adaptability of the decision-making process.

[0135] ② AR interactive terminal:

[0136] This invention introduces Microsoft HoloLens 2 augmented reality glasses as a human-computer interaction terminal to improve the level of intelligence in on-site operations:

[0137] Real-time overlay of virtual grazing boundaries: green indicates grazing areas and red indicates no-grazing areas, helping herders to intuitively identify compliant areas;

[0138] Voice prompts provide key indicators, such as "Soil moisture ahead is 12%, it is recommended to stay for less than 2 hours," to help herders make immediate judgments.

[0139] Gesture operation to access historical data comparison: Users can use swipe gestures to view the difference in NDVI between this month and last month, and grasp the trend of grassland health status changes.

[0140] The specific implementation steps of this invention are as follows:

[0141] 1. Network Architecture Construction

[0142] ①Perception layer:

[0143] Space-based: Access Sentinel-2, GRACE-FO and FY-4A data, and update vegetation index, groundwater storage and precipitation forecast data daily and upload them directly to the cloud platform;

[0144] Airborne: UAV aerial survey data (LiDAR + multispectral) is transmitted to the edge layer via a 5G slicing network at the transmission layer, preprocessed, and then uploaded to the cloud platform;

[0145] Foundation: IoT sensors (soil multi-parameter instrument, livestock smart collar) transmit data to the edge layer via LoRaWAN in the transmission layer, and are then pre-processed by the site edge gateway before being synchronized to the cloud platform;

[0146] ②Transport layer:

[0147] 5G slicing network: Allocates dedicated channels for emergency data (such as rainstorm warnings), with latency <50ms.

[0148] BeiDou short message service: Ensuring data transmission in areas without network coverage (such as monitoring points in desert areas).

[0149] ③ Edge layer:

[0150] Regional peripheral center: Run the SWAT-HS model to generate weekly water supply and demand reports;

[0151] Station edge gateway: Real-time calculation of grassland carrying capacity,

[0152] ④ Cloud platform:

[0153] Digital Twin Engine: Constructing a 3D model of the watershed to simulate ecological responses under different grazing / precipitation scenarios.

[0154] Decision-Making Cockpit: Visualizes the Ecological Security Index (ESI), Water Pressure Level (WPL), etc.

[0155] 2. Intelligent Application Scenarios

[0156] ① Grassland degradation early warning:

[0157] The edge gateway collects and analyzes NDVI (Normalized Difference Vegetation Index) and soil moisture data from the perception layer in real time. When an NDVI decrease of more than 15% and soil moisture content is below 10% in a certain area, an alarm mechanism is triggered, and the relevant data is stored as evidence using blockchain technology.

[0158] After receiving data uploaded by the edge gateway, the cloud platform uses historical data and a deep reinforcement learning (DRL) model to assess the risk of grassland degradation in the affected area over the next three months. If the prediction shows a degradation risk probability exceeding 80%, a detailed ecological security report is generated.

[0159] Based on the assessment results, the system automatically generates grazing bans to prevent further deterioration. These bans are pushed directly to local herders via augmented reality (AR) glasses, while simultaneously initiating ecological compensation smart contracts to ensure that affected areas receive appropriate economic compensation or restoration measures.

[0160] ② Flood risk prevention and control:

[0161] The sensing layer continuously monitors precipitation in the region using space-based satellites (such as the FY-4A meteorological satellite). When the cumulative precipitation exceeds 50 mm within 24 hours, an initial warning signal is immediately generated and the edge layer is notified to prepare to activate the emergency response procedure.

[0162] Upon receiving the early warning signal, the edge layer rapidly activates a lightweight HEC-RAS flood model. Combined with LiDAR (VelodynePuck) point cloud analysis, it acquires runoff paths, waterlogged areas, and gully development information for running a simplified version of the HEC-RAS model, enabling rapid simulation of the flood inundation extent. The inundation extent map is generated within 10 minutes, with an accuracy controlled within ±5 meters.

[0163] LiDAR terrain data and precipitation data are fused and preprocessed, compressed, and then uploaded to the cloud platform. The cloud platform integrates all relevant information, including simulation results and real-time meteorological data, and hashes the early warning event on the blockchain. At the same time, the full-featured SWAT-HS model is called to verify the reliability of the edge layer model simulation results, forming comprehensive flood early warning information and pushing the final instructions to the decision-making level.

[0164] The decision-makers broadcast early warning information to residents in the relevant areas through the 5G network, providing specific evacuation guidelines and instructing livestock to be moved to safe areas.

[0165] Decision-makers may also use AR devices to provide intuitive operational guidance to on-site rescue personnel, thereby improving response efficiency.

[0166] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0167] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A cloud-edge-device collaborative intelligent management system for agricultural and pastoral ecological hydrology, characterized in that, include: Eco-hydrological information management module and eco-hydrological source tracing and early warning application module; The eco-hydrological information management module includes a sensing layer, a transmission layer, and an edge layer; the eco-hydrological source tracing and early warning application module includes a cloud platform and a decision-making layer. The sensing layer collects multi-source data through space-based, air-based, and ground-based systems; The transport layer uses 5G slicing network technology to build an efficient and low-latency communication channel, and transmits multi-source data to the edge layer through data transmission and priority scheduling mechanisms; The edge layer enables local data analysis and rapid response based on multi-source data, and sends multi-source data to the cloud platform. The edge layer includes: At the edge of the region, a high-performance server cluster is set up to run the SWAT-HS hydrological model, simulate watershed-scale water cycle and generate ecological security index. The edge gateway of the site runs a TensorFlow Lite model to calculate the carrying capacity of grassland in real time. The cloud platform integrates a digital twin engine, model library, and blockchain fusion module, providing comprehensive data processing, simulation prediction, and intelligent decision support. Based on the in-depth analysis results from the cloud platform, the decision-making level employs a multi-objective decision optimization algorithm to optimize the balance between livestock income, ecological water demand, and net ecosystem productivity, and to trigger the execution of the ecological compensation mechanism, generating multi-dimensional decision instructions. The ecological compensation mechanism specifically includes: When the Ecological Security Index (ESI) of a certain grassland is greater than 0.8 for three consecutive months, compensation is triggered. in, For grassland ecological compensation funds. Let A be a coefficient, and A be the grassland area. For carbon sink prices; The multi-objective optimization function is: in, For the benefit of the livestock economy, For ecological water demand, For net ecosystem productivity, For ecological security index, The carrying capacity of grassland. The maximum carrying capacity of the grassland. Soil moisture content, This represents the critical threshold for soil moisture. , , , , As weight, ESI It is the ecological security index.

2. The intelligent management system for agricultural and pastoral ecological hydrology based on cloud-edge-device collaboration according to claim 1, characterized in that, The perception layer includes: Space-based: Integrates different types of satellites to acquire daily data on Normalized Difference Vegetation Index (NDVI), Enhanced Difference Vegetation Index (EVI), groundwater storage, and precipitation forecasts; Airborne: Configure a drone swarm equipped with multispectral sensors, LiDAR, and thermal infrared sensors. The multispectral sensors are used to invert and process biomass density, the LiDAR is used to analyze point clouds to obtain runoff paths, waterlogged areas, and gully development, and the thermal infrared sensors are used to obtain grazing activity index and grazing intensity index. Based on the data obtained above, a high-precision vegetation cover heat map is generated in real time to optimize grazing paths. Foundation: Deploy IoT terminals, including a soil multi-parameter instrument and a smart collar for livestock. The soil multi-parameter instrument is used to collect soil moisture content and monitor surface parameters, while the smart collar for livestock is used to monitor livestock feeding behavior and GPS movement trajectory.

3. The intelligent management system for agricultural and pastoral ecological hydrology based on cloud-edge-device collaboration according to claim 2, characterized in that, The cloud platform layer includes: The digital twin engine integrates multi-source data through 3D modeling, supports dynamic parameter injection and real-time rendering, and is used to simulate ecological responses under different scenarios. The model library contains hydrological models, ecological models, and AI models for use by the edge layer and decision-making layer. The blockchain integration module utilizes Hyperledger Fabric to build a consortium blockchain for blockchain notarization and supports an ecosystem compensation mechanism.

4. The intelligent management system for agricultural and pastoral ecological hydrology based on cloud-edge-device collaboration according to claim 3, characterized in that, Hydrological models include the SWAT-HS hydrological model and the HEC-RAS flood simulation model; Ecological models include the CENTURY grassland carbon cycle model and the TensorFlow Lite model; AI models are used to optimize grazing routes based on high-precision vegetation cover heat maps.

5. The intelligent management system for agricultural and pastoral ecological hydrology based on cloud-edge-device collaboration according to claim 4, characterized in that, Calculate the ecological security index using the SWAT-HS hydrological model. ESI : in, For soil moisture content, For soil moisture critical threshold, Normalized Difference Vegetation Index (NDVI) The normalized vegetation index baseline value, For groundwater reserves, Average groundwater storage over many years, weighting coefficient ; Run TensorFlow Lite models to calculate pasture carrying capacity in real time C : Where A represents grassland area and BD represents biomass density. The coefficient of recovery, Daily food intake The number of days for grazing.

6. The intelligent management system for agricultural and pastoral ecological hydrology based on cloud-edge-device collaboration according to claim 1, characterized in that, The decision-making level also includes AR interactive terminals.

Images