Sand barrier aerial seeding collaborative layout optimization system and method based on unmanned aerial vehicle cluster game

Abstract

The invention relates to the technical field of desertification control, in particular to a sand barrier aerial seeding collaborative layout optimization system and method based on an unmanned aerial vehicle cluster game, which constructs a microenvironment digital model of terrain, wind field, humidity and temperature through a multi-mode desert microenvironment sensing module and combines with a sand barrier-vegetation coupling dynamic model. Space-time correlation between sand barrier configuration and vegetation growth is achieved, the self-adaptive sand barrier parameter optimization module and the precise seed aerial seeding control module work cooperatively, seed putting is optimized, the sand stabilization efficiency and the vegetation survival rate are remarkably improved to 60%-75%, the treatment period is shortened to 1.5-2.5 years, and the control cost is reduced. The closed-loop feedback and dynamic optimization module further ensures continuous improvement of the treatment effect, a traditional'sand blocking 'strategy is converted into'ecological microenvironment engineering', a complete microenvironment regulation and control mechanism is constructed, the treatment efficiency is improved by 35%-50%, and fundamental change of the treatment concept is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of desertification control, and specifically relates to a collaborative layout optimization system and method for sand barrier aerial seeding based on drone swarm game. Background Art

[0002] Sand barrier sand fixation and vegetation restoration are two key measures for desertification control. At present, for the monitoring of the sand fixation function of sand barriers, traditional tagging methods or sand accumulation meters are still commonly used, that is, a fixed-point measurement method of setting stakes and sand accumulation meters in the sand barrier establishment area, and the sand fixation effect of the sand barrier is estimated by the change of the sand surface activity at the positioning points.

[0003] For example, Chinese Patent Application CN114897776A discloses a method for monitoring the sand fixation function of sand barriers based on drones. It obtains images of the target sand barrier control area through drones, performs aerial triangulation operations on the images to obtain a digital orthophoto map (DOM) and a digital surface model map (DSM), and processes them to obtain a digital elevation model map (DEM), and then draws a topographic map, and finally calculates the wind-deposited volume and wind-eroded volume. Although this method has made breakthroughs in monitoring technology, it still has the following deficiencies:

[0004] 1. Only a single drone is used for operation, which is inefficient and has limited coverage in a large-area desert environment;

[0005] 2. Only the effect of existing sand barriers is evaluated, and the layout of sand barriers cannot be actively optimized;

[0006] 3. The synergistic relationship between sand barriers and vegetation restoration is not considered, and the comprehensive ecological restoration mechanism is ignored;

[0007] 4. Parameters such as the shape, height, and density of sand barriers cannot be flexibly adjusted according to microenvironmental conditions;

[0008] 5. It has poor adaptability under strong wind conditions and cannot ensure the seed retention and vegetation restoration efficiency;

[0009] 6. The evaluation system shows obvious lag and lacks prediction and active optimization capabilities.

[0010] The above technical solutions only provide passive monitoring of the sand fixation effect of sand barriers, cannot achieve the collaborative optimization of sand barrier layout and vegetation restoration, and are difficult to meet the requirements of comprehensive desertification control. Therefore, there is an urgent need to develop a system that can actively optimize the sand barrier layout and collaborate with aerial seeding to improve the efficiency and quality of desertification control. Summary of the Invention

[0011] The purpose of the present invention is to provide a collaborative layout optimization system and method for sand barrier aerial seeding based on drone swarm game, aiming to solve the technical problems in the prior art such as the separation of sand barrier layout and vegetation restoration, the inability to flexibly adjust sand barrier parameters according to microenvironmental conditions, and the lack of active optimization capabilities.

[0012] The present invention proposes a collaborative layout optimization system for sand barrier aerial seeding based on UAV swarm game, including:

[0013] A multi-modal desert microenvironment perception module, which is used to collect multi-dimensional environmental data of the sandy land and construct a microenvironment digital model including terrain, wind field, humidity, and temperature;

[0014] A sand barrier-vegetation coupling dynamic model module, which is connected to the multi-modal desert microenvironment perception module for data, and is used to receive the microenvironment digital model and construct a spatio-temporal correlation model of the wind field regulation by the sand barrier and vegetation growth;

[0015] An adaptive sand barrier parameter optimization module, which is connected to the sand barrier-vegetation coupling dynamic model module for data, and is used to generate an optimal sand barrier configuration plan based on the spatio-temporal correlation model. The sand barrier configuration plan includes parameters such as sand barrier height, density, material, and arrangement pattern;

[0016] A precise seed aerial seeding control module, which is connected to the adaptive sand barrier parameter optimization module for data, and is used to receive the sand barrier configuration plan, adjust the seed delivery parameters according to real-time wind field data, and control the seed delivery position, time, and density;

[0017] A closed-loop feedback and dynamic optimization module, which is connected to the precise seed aerial seeding control module and the multi-modal desert microenvironment perception module for data, and is used to collect data on the treatment effect, evaluate the sand fixation efficiency of the sand barrier and the vegetation growth status, and feedback optimization suggestions to the adaptive sand barrier parameter optimization module.

[0018] Preferably, the multi-modal desert microenvironment perception module includes:

[0019] A multi-spectral data acquisition unit, which is used to collect surface information in visible light, near-infrared, and short-wave infrared bands;

[0020] A wind field monitoring unit, which is used to collect three-dimensional wind field data with a frequency of 0.1-10 Hz;

[0021] A terrain mapping unit, which is used to obtain a digital elevation model with a resolution of 5 cm;

[0022] A data fusion processing unit, which is used to integrate multi-source data to generate a unified microenvironment digital model.

[0023] Preferably, the sand barrier-vegetation coupling dynamic model module includes:

[0024] A wind field regulation model unit, which is used to simulate the influence of different sand barrier forms on the wind field;

[0025] A vegetation growth condition model unit, which is used to determine the threshold conditions of the microenvironment required for vegetation growth;

[0026] A spatio-temporal correlation calculation unit, which is used to establish a correlation model between the sand barrier layout and vegetation growth through a convolutional neural network algorithm.

[0027] Preferably, the adaptive sand barrier parameter optimization module includes:

[0028] A sand barrier specification library unit, which contains sand barrier parameters of 6 basic forms and 3 material combinations;

[0029] A multi-objective optimization unit, which is used to balance the objectives of sand fixation efficiency, vegetation growth and cost control;

[0030] A reinforcement learning unit, which is used to adaptively adjust the sand barrier parameters to adapt to different micro-environment conditions.

[0031] Preferably, the sand barrier parameters in the sand barrier specification library unit include:

[0032] The sand barrier height parameter, with a continuously adjustable range of 20 - 50 cm;

[0033] The sand barrier density parameter, with a continuously adjustable ventilation rate range of 30% - 70%;

[0034] The sand barrier material parameter, including wheat straw, reed, and shrub branches;

[0035] The sand barrier arrangement mode parameter, including square, strip, checkerboard, honeycomb, herringbone, and cross.

[0036] Preferably, the precise seed aerial seeding control module includes:

[0037] An intelligent seed delivery unit, which can automatically adjust the delivery parameters according to real-time wind field data;

[0038] A seed wrapping treatment unit, which is used to mix seeds with water retaining agents, nutrient agents, and adhesives to improve wind resistance;

[0039] A delivery trajectory optimization unit, which is used to plan the optimal delivery path and flight parameters.

[0040] Preferably, the control parameters of the precise seed aerial seeding control module include:

[0041] The seed delivery density, which can be precisely controlled within the range of 500 - 2000 seeds per mu;

[0042] The delivery height, which can be automatically adjusted within the range of 5 - 20 m;

[0043] The ratio of the seed wrapping material, prepared as seed: water retaining agent: nutrient agent: adhesive = 1: 0.5 - 2: 0.1 - 0.5: 0.1 - 0.4.

[0044] Preferably, the closed-loop feedback and dynamic optimization module includes:

[0045] Regular monitoring unit, which collects data on the treatment effect at three levels of short-term, medium-term, and long-term frequencies;

[0046] Evaluation index calculation unit, which is used to calculate the sand fixation rate, vegetation coverage, biomass, seed germination rate, and cost-benefit ratio;

[0047] Self-learning optimization unit, which is used to continuously improve the sand barrier aerial seeding collaborative strategy based on time series data.

[0048] Preferably, the monitoring frequency of the regular monitoring unit is set as follows:

[0049] Short-term monitoring, once every 7 days;

[0050] Medium-term monitoring, once every 30 days;

[0051] Long-term monitoring, once every 90 days;

[0052] When the environmental parameters change by more than 20%, emergency monitoring is triggered.

[0053] The collaborative layout optimization method of sand barrier aerial seeding based on unmanned aerial vehicle swarm game includes the following steps:

[0054] Collect multi-dimensional environmental data of the sandy land and construct a micro-environment digital model;

[0055] Based on the micro-environment digital model, establish a spatio-temporal correlation model of the sand barrier's regulation of the wind field and vegetation growth;

[0056] According to the spatio-temporal correlation model, generate an optimal sand barrier configuration plan;

[0057] Based on the sand barrier configuration plan, adjust the seed delivery parameters, and control the seed delivery position, time, and density;

[0058] Collect data on the treatment effect and evaluate the sand fixation efficiency of the sand barrier and the vegetation growth status;

[0059] According to the treatment effect data, update the sand barrier configuration plan and the seed delivery parameters to form a closed-loop optimization.

[0060] The beneficial effects of the present invention include:

[0061] 1. Realize the collaborative optimization of the sand barrier layout and vegetation restoration, enabling the sand fixation efficiency and vegetation recovery to form a positive gain, with the treatment efficiency increased by 35% - 50%, the vegetation survival rate increased from the traditional 30% - 40% to 60% - 75%, and the treatment cycle shortened from the traditional 3 - 5 years to 1.5 - 2.5 years;

[0062] 2. Through the multi-modal desert microenvironment perception and sand barrier-vegetation coupling dynamic model, a complete microenvironment regulation mechanism was constructed, transforming from simple sand blocking to complex ecological microenvironment engineering, achieving a qualitative change in the governance concept;

[0063] 3. An adaptive sand barrier parameter optimization system was developed, which can automatically adjust parameters such as the height, density, material, and arrangement pattern of the sand barrier according to microenvironment conditions, greatly improving the adaptability and sand fixation efficiency of the sand barrier;

[0064] 4. The precise seed aerial seeding control technology was innovatively designed, and the seed retention rate can still be maintained above 90% under 8-level wind conditions, significantly improving the vegetation restoration efficiency under extreme conditions;

[0065] 5. A closed-loop feedback and dynamic optimization mechanism was established, and the system can continuously self-learn and adjust according to the actual effect, transforming from static design to dynamic evolution, and continuously improving the governance effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 It is a schematic diagram of the architecture of the sand barrier aerial seeding collaborative layout optimization system based on UAV swarm game of the present invention;

[0067] Figure 2 It is a schematic diagram of the structure of the multi-modal desert microenvironment perception module of the present invention;

[0068] Figure 3 It is a flowchart of the working process of the sand barrier-vegetation coupling dynamic model module of the present invention;

[0069] Figure 4 It is a schematic diagram of the structure of the adaptive sand barrier parameter optimization module of the present invention;

[0070] Figure 5 It is a schematic diagram of the composition of the precise seed aerial seeding control module of the present invention;

[0071] Figure 6 It is a flowchart of the working process of the closed-loop feedback and dynamic optimization module of the present invention;

[0072] Figure 7 It is a schematic diagram of the process of the sand barrier aerial seeding collaborative layout optimization method based on UAV swarm game of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0073] Please refer to the attached Figures 1-7 , and the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0074] Example 1

[0075] Refer to Figure 1, the collaborative layout optimization system for sand barrier aerial seeding based on UAV swarm game provided by the present invention includes a multi-modal desert microenvironment perception module 1, a sand barrier-vegetation coupling dynamic model module 2, an adaptive sand barrier parameter optimization module 3, a precise seed aerial seeding control module 4, and a closed-loop feedback and dynamic optimization module 5.

[0076] The multi-modal desert microenvironment perception module 1 is used to collect multi-dimensional environmental data of the sandy land and construct a microenvironment digital model including terrain, wind field, humidity, and temperature. This module uses a UAV swarm equipped with multi-spectral cameras, thermal infrared sensors, and millimeter-wave radars, and adopts a four-dimensional scanning strategy (three-dimensional space + one-dimensional time) to obtain detailed environmental data of the desert area, and constructs a complete microenvironment digital model through a multi-source data fusion algorithm.

[0077] The sand barrier-vegetation coupling dynamic model module 2 is connected to the multi-modal desert microenvironment perception module 1 for data, and is used to receive the microenvironment digital model and construct a spatio-temporal correlation model of the sand barrier's regulation of the wind field and vegetation growth. This module constructs a regulation model of the sand barrier on the wind field based on the principle of fluid mechanics, and at the same time develops a microenvironment condition model for seed germination and seedling survival, and establishes a spatio-temporal correlation dynamic model of the sand barrier layout and vegetation growth through a convolutional neural network algorithm.

[0078] The adaptive sand barrier parameter optimization module 3 is connected to the sand barrier-vegetation coupling dynamic model module 2 for data, and is used to generate an optimal sand barrier configuration plan based on the spatio-temporal correlation model. This configuration plan includes parameters such as sand barrier height, density, material, and arrangement pattern. This module designs a variable form sand barrier specification library, which includes 6 basic forms and 3 material combinations, and adaptively optimizes the sand barrier parameters through a reinforcement learning algorithm to achieve multi-objective trade-offs (sand fixation efficiency, vegetation growth, cost control).

[0079] The precise seed aerial seeding control module 4 is connected to the adaptive sand barrier parameter optimization module 3 for data, and is used to receive the sand barrier configuration plan, adjust the seed delivery parameters according to the real-time wind field data, and control the seed delivery position, time, and density. This module develops an intelligent seed delivery system and seed wrapping technology, which can maintain a high seed retention rate under strong wind conditions and significantly improve the vegetation restoration efficiency.

[0080] The closed-loop feedback and dynamic optimization module 5 is connected to the precise seed aerial seeding control module 4 and the multi-modal desert microenvironment perception module 1 for data, and is used to collect governance effect data, evaluate the sand fixation efficiency of the sand barrier and the vegetation growth status, and feedback optimization suggestions to the adaptive sand barrier parameter optimization module 3. This module establishes a regular monitoring plan and an evaluation index system, and continuously improves the collaborative strategy of sand barrier aerial seeding through a self-learning optimization model of time series data.

[0081] The system constructs a complete closed-loop architecture of perception - analysis - decision - execution - optimization. Information exchange and collaborative work are achieved among modules through standardized data interfaces. The system adopts a cluster game mechanism, where each UAV acts as a game participant for local decision-making while considering the global optimal goal, realizing the collaborative optimization of sand barrier layout and aerial seeding.

[0082] Embodiment 2

[0083] Referring to Figure 2 , the multi-modal desert microenvironment perception module 1 in this embodiment includes a multi-spectral data acquisition unit 11, a wind field monitoring unit 12, a terrain mapping unit 13, and a data fusion and processing unit 14.

[0084] The multi-spectral data acquisition unit 11 is used to collect surface information in the visible light (RGB), near-infrared (NIR 760 - 900nm), and short-wave infrared (SWIR 1500 - 1700nm) bands. This unit is equipped with a high-resolution multi-spectral camera that can capture the surface reflection characteristics in different bands for identifying key information such as soil type, moisture content, and vegetation condition. During the scanning process, the sampling density of the multi-spectral camera is set to 25 - 100 sampling points per square meter and is automatically adjusted according to the required resolution.

[0085] The wind field monitoring unit 12 is used to collect three-dimensional wind field data with a frequency of 0.1 - 10Hz. This unit is equipped with a small weather station and an aerial wind field sensor that can monitor parameters such as wind speed, wind direction, and turbulence intensity in real time. The wind field sampling frequency can be adjusted within the range of 0.1 - 10Hz and automatically adapts according to the wind speed change rate. Low-frequency sampling is adopted when the wind field is relatively stable, and it automatically switches to high-frequency sampling when the wind field fluctuates greatly, ensuring both data quality and energy efficiency.

[0086] The terrain mapping unit 13 is used to obtain a digital elevation model with a resolution of 5cm. This unit uses lidar (LiDAR) technology that can penetrate surface vegetation to obtain accurate terrain data. The scanning height can be adjusted within the range of 50 - 200m, and the optimal flight height is automatically set according to the required resolution. In areas of key terrain features (such as dune ridge lines and slope bottom lines), the system will automatically increase the sampling density to ensure accurate capture of terrain details.

[0087] The data fusion processing unit 14 is used to integrate multi-source data to generate a unified digital model of the microenvironment. This unit adopts a multi-layer data fusion algorithm to integrate data from different sources and at different scales into a unified environmental representation model. The specific fusion process includes four steps: data preprocessing, spatial alignment, time synchronization, and feature extraction. In the data preprocessing stage, the system filters noise and detects outliers in the original data; in the spatial alignment stage, the data obtained by different sensors are mapped into a unified spatial coordinate system; in the time synchronization stage, data with different sampling frequencies are processed to establish time consistency; in the feature extraction stage, key environmental features are extracted from the fused data.

[0088] The finally generated digital model of the microenvironment contains the following core information: three-dimensional representation of the terrain (digital elevation model), three-dimensional velocity vector field of the wind field, soil moisture distribution map, surface temperature distribution map, and soil texture feature map. This model has high spatio-temporal resolution and can accurately capture the micro-scale changes in the desert environment, providing basic data support for subsequent sand barrier-vegetation coupling analysis.

[0089] Preferably, the multi-modal desert microenvironment perception module 1 adopts a distributed data acquisition strategy, and multiple drones simultaneously collect data in different regions and at different heights, greatly improving the efficiency and coverage of data acquisition. At the same time, the system adopts edge computing technology to perform preliminary data processing on the drones, reducing the data transmission burden and improving the overall operation efficiency of the system.

[0090] Embodiment 3: Detailed structure of the sand barrier-vegetation coupling dynamic model module

[0091] Refer to Figure 3 , the sand barrier-vegetation coupling dynamic model module 2 in this embodiment includes a wind field regulation model unit 21, a vegetation growth condition model unit 22, and a spatio-temporal correlation calculation unit 23.

[0092] The wind field regulation model unit 21 is used to simulate the influence of different sand barrier forms on the wind field. Based on the principle of computational fluid dynamics (CFD), this unit constructs a sand barrier-wind field interaction model, which can accurately predict the influence of different forms of sand barriers (height, density, arrangement pattern) on the wind speed distribution and turbulence characteristics. In the model, the sand barrier is regarded as a porous medium, and its ventilation rate is related to the sand barrier density. When the wind passes through the sand barrier, wind speed reduction areas and turbulence enhancement areas will be generated, and the size and intensity of these areas depend on the specific parameters of the sand barrier.

[0093] The wind field regulation effect is usually measured by two indicators: the wind speed attenuation rate and the turbulence intensity. The wind speed attenuation rate is defined as the relative change in wind speed before and after the sand barrier, usually in the range of 20% - 80%; the turbulence intensity is expressed by a dimensionless coefficient, usually in the range of 0.05 - 0.5. These two indicators jointly determine the sand fixation effect of the sand barrier and also have an important impact on the subsequent vegetation growth.

[0094] The vegetation growth condition model unit 22 is used to determine the microenvironment threshold conditions required for vegetation growth. This unit constructs an environmental response model for desert vegetation growth, including two key links: the seed germination stage and the seedling growth stage. The model considers the effects of various environmental factors such as wind speed, temperature, humidity, and light on vegetation growth and determines the appropriate range and critical thresholds of each factor.

[0095] For typical desert vegetation, the appropriate wind speed range for seed germination is 0 - 3.5 m / s, and the critical wind speed for seedling survival is 5 m / s. When the wind speed exceeds the critical value, it will cause the seeds to be blown away or the seedlings to be blown down. In addition, there is an optimal distance relationship between the sand barrier and the vegetation, usually expressed as a multiple of the sand barrier height, and the optimal range is between 0.5 - 2.0 times the sand barrier height. Within this range, the vegetation can be effectively protected by the sand barrier and will not be overly affected by the sand barrier shadow.

[0096] The spatio-temporal correlation calculation unit 23 is used to establish a correlation model between the sand barrier layout and vegetation growth through a convolutional neural network algorithm. This unit uses a deep learning method to construct an end-to-end prediction model of sand barrier parameters - microenvironment conditions - vegetation response. The model takes the sand barrier layout parameters and basic environmental conditions as inputs and predicts the microenvironment characteristics and vegetation growth probability at different locations.

[0097] In the design of the convolutional neural network, a multi-layer convolutional structure is adopted to capture spatial features, and at the same time, a processing mechanism for the time dimension is introduced, which can predict the environmental changes and vegetation responses at different time points. The key layer structures of the network include: the input layer (receiving sand barrier parameters and environmental data), the multi-layer convolutional layer (extracting spatial features), the recurrent neural network layer (processing time series), and the output layer (generating microenvironment predictions and vegetation growth probabilities).

[0098] The core innovation of the sand barrier - vegetation coupling dynamic model lies in the organic combination of the sand fixation function of the sand barrier and the growth conditions of the vegetation, constructing a unified analysis framework. Through this model, the system can predict the microenvironment changes under different sand barrier layouts, and then evaluate its impact on vegetation growth, providing a scientific basis for the optimization of sand barrier parameters.

[0099] Preferably, during the model training process, the spatio-temporal correlation calculation unit 23 adopts transfer learning technology to transfer the empirical knowledge accumulated in other regions to the new region, greatly reducing the amount of data and time required for model training. At the same time, the model has the ability of adaptive learning and can continuously adjust and optimize the prediction algorithm according to the actual monitoring data, improving the accuracy and adaptability of the prediction.

[0100] Embodiment 4

[0101] Referring to Figure 4 , the adaptive sand barrier parameter optimization module 3 in this embodiment includes a sand barrier specification library unit 31, a multi-objective optimization unit 32, and a reinforcement learning unit 33.

[0102] The sand barrier specification library unit 31 contains sand barrier parameters of 6 basic forms and 3 material combinations. This unit stores a preset sand barrier specification parameter library as the basic options for the optimization algorithm. The sand barrier parameters mainly include four aspects: sand barrier height, sand barrier density (ventilation rate), sand barrier material, and sand barrier arrangement pattern.

[0103] The sand barrier height parameter is continuously adjustable within the range of 20 - 50 cm, which can adapt to different wind conditions and sand fixation requirements. Generally speaking, higher sand barriers are used in areas with stronger winds, medium-height sand barriers are used in areas with moderate winds, and lower sand barriers can be used in areas with weaker winds. The sand barrier density parameter, expressed by the ventilation rate, is continuously adjustable in the range of 30% - 70%. If the ventilation rate is too low, strong eddies will appear on the leeward side of the sand barrier, which is not conducive to vegetation growth; if the ventilation rate is too high, the windproof and sand-fixing effect of the sand barrier will be weakened. The sand barrier materials include three optional materials: wheat straw, reed, and shrub branches. Different materials have different ventilation characteristics, service lives, and ecological adaptabilities. The sand barrier arrangement patterns include six options: square grid, strip, checkerboard, honeycomb, herringbone, and cross-shaped. Different arrangement patterns are suitable for different terrain and wind direction conditions.

[0104] The multi-objective optimization unit 32 is used to balance the three objectives of sand fixation efficiency, vegetation growth, and cost control. This unit adopts a multi-objective optimization algorithm to find the best balance point among multiple objectives. During the optimization process, the system will comprehensively consider the following aspects: sand fixation efficiency (measured by wind deposition and wind erosion), vegetation support score (evaluated by microenvironment suitability), cost per hectare (calculated by material usage and construction difficulty), and expected service life.

[0105] The multi-objective optimization adopts the Pareto Front search strategy to find a set of non-dominated solutions through continuous iteration, that is, solutions that cannot improve other objectives without sacrificing one objective. In practical applications, the system will select the most suitable solution from the Pareto Front according to specific governance requirements and resource constraints.

[0106] The reinforcement learning unit 33 is used to adaptively adjust the sand barrier parameters to suit different micro-environment conditions. Based on a deep reinforcement learning framework, this unit models the sand barrier parameter optimization problem as a Markov decision process (MDP). In this framework, the environmental state is the micro-environment characteristics, the action is the adjustment of the sand barrier parameters, and the reward is the evaluation of the sand barrier effect. By continuously interacting with the environment and obtaining feedback, the reinforcement learning algorithm can gradually improve the decision-making strategy and find the optimal sand barrier parameter configuration.

[0107] The reinforcement learning framework adopts a hybrid architecture that combines a deep Q-network (DQN) and policy gradient, enabling it to handle both continuous action spaces (such as the continuous adjustment of sand barrier height and density) and discrete action spaces (such as material selection and arrangement pattern selection) simultaneously. The core of the algorithm is a state-action value function Q(s,a), which is used to estimate the long-term cumulative reward for executing action a in state s.

[0108] Preferably, the adaptive sand barrier parameter optimization module 3 adopts a hierarchical optimization strategy. First, it determines the sand barrier types and layouts in different regions at the macroscopic scale, and then finely adjusts the specific parameters at the microscopic scale. This hierarchical strategy not only ensures the rationality of the overall layout but also takes into account the optimization of local details, greatly improving the computational efficiency and optimization quality. At the same time, the system also has a fast response ability and can quickly adjust the sand barrier configuration plan according to real-time information such as extreme weather warnings, enhancing the system's emergency adaptability.

[0109] Example 5

[0110] This example details various sand barrier parameters in the sand barrier specification library unit 31, including sand barrier height parameters, sand barrier density parameters, sand barrier material parameters, and sand barrier arrangement pattern parameters.

[0111] The range of the sand barrier height parameter is continuously adjustable from 20 to 50 cm. The sand barrier height is one of the key factors affecting the sand fixation effect. The greater the height, the wider the protection range of a single sand barrier, but the material consumption and construction difficulty also increase accordingly. Based on a large number of experiments and practical experience, there is an optimal sand barrier height under different wind speed conditions: when the wind speed is less than 5 m / s, a low sand barrier with a height of 20 - 30 cm can meet the requirements; when the wind speed is in the range of 5 - 10 m / s, a medium-height sand barrier of 30 - 40 cm has the best effect; when the wind speed is greater than 10 m / s, a high sand barrier of 40 - 50 cm is required for effective protection. The system will automatically calculate the optimal sand barrier height according to the regional wind speed characteristics and budget constraints.

[0112] The sand barrier density parameter, i.e., the ventilation rate, is continuously adjustable within the range of 30% to 70%. The ventilation rate is an important indicator to measure the density of the sand barrier structure and directly affects the aerodynamic characteristics of the sand barrier. Research shows that when the ventilation rate is below 30%, strong eddies will form on the leeward side of the sand barrier, leading to increased wind erosion; when the ventilation rate is above 70%, the wind blocking effect of the sand barrier is significantly reduced and the sand fixation ability is insufficient. The optimal ventilation rate is closely related to the height of the sand barrier and the local wind speed. Generally, the greater the wind speed, the lower the appropriate ventilation rate. The system uses a computational fluid dynamics model to accurately calculate the wind field distribution at different ventilation rates and find the optimal ventilation rate setting.

[0113] The sand barrier material parameters include three types: wheat straw, reed, and shrub branches. Different materials have different performance characteristics and application conditions. The wheat straw sand barrier is easy to make and has a low cost, suitable for short-term and rapid treatment, but its service life is only 1 - 2 years; the reed sand barrier has a higher strength and a service life of 3 - 4 years, and performs well in areas with relatively rich water; the shrub branch sand barrier has the best durability and a service life of more than 5 years, but it is complex to make and has a high cost. In practical applications, the system will select the most suitable sand barrier material according to the treatment cycle, budget constraints, and the availability of local resources. In some cases, the system will also recommend a mixed material solution, such as using reed as the main body and shrub branches for reinforcement, to balance cost and durability.

[0114] The sand barrier arrangement pattern parameters include six types: square grid, strip, checkerboard, honeycomb, herringbone, and cross. Different arrangement patterns are suitable for different topographies and wind direction characteristics. The square grid arrangement is the most traditional way to set up sand barriers and is suitable for areas with variable wind directions; the strip arrangement is set perpendicular to the dominant wind direction, with high sand fixation efficiency but poor adaptability to wind direction changes; the checkerboard arrangement can save materials while maintaining a certain sand fixation efficiency; the honeycomb arrangement has an all-round protection ability and is suitable for complex wind field environments; the herringbone arrangement performs excellently in areas with a clear dominant wind direction; the cross arrangement is effective in monsoon areas with alternating wind direction changes. The system will automatically select the optimal arrangement pattern according to the regional wind field characteristics and topographic conditions, and calculate the best arrangement spacing and coverage density.

[0115] In addition to the basic parameters, the sand barrier specification library unit 31 also contains a rich combination of solutions, which can generate personalized sand barrier configurations according to regional characteristics. For example, in areas with severe wind erosion, the system will recommend a denser sand barrier layout; in key areas for vegetation restoration, the sand barrier parameters will be optimized to create a microenvironment more suitable for vegetation growth; in phased treatment areas, the timing matching of sand barrier renewal and vegetation growth will be considered to formulate a phased implementation plan.

[0116] Preferably, the sand barrier specification library unit 31 uses knowledge graph technology to combine expert experience with experimental data to construct a knowledge base for sand barrier parameter selection. The system can quickly retrieve the best practices of similar cases according to the input environmental conditions as the initial optimization plan, greatly improving the optimization efficiency and plan quality.

[0117] Example 6

[0118] Refer to Figure 5 , the precise seed aerial seeding control module 4 in this embodiment includes an intelligent seed delivery unit 41, a seed wrapping treatment unit 42, and a delivery trajectory optimization unit 43.

[0119] The intelligent seed delivery unit 41 can automatically adjust the delivery parameters according to the real-time wind field data. This unit is equipped with a precision control system that can sense the changes in wind speed and direction in real time and adjust the delivery height, delivery rate, and delivery direction accordingly. The system adopts a closed-loop control strategy to ensure that the seeds accurately fall to the target position through three links: pre-delivery wind measurement, in-delivery correction, and post-delivery evaluation.

[0120] During the delivery process, the system will calculate the impact of wind speed and direction on the seed falling trajectory in real time and perform predictive compensation. The delivery height can be automatically adjusted within the range of 5 - 20m, reducing the drift by lowering the height when the wind is strong and increasing the coverage by raising the height when the wind is weak. The delivery accuracy reaches ±0.5m in the horizontal direction and ±0.2m in the vertical direction, which is much better than the traditional aerial seeding technology.

[0121] The seed wrapping treatment unit 42 is used to mix seeds with water-retaining agents, nutrient agents, and adhesives to improve the wind resistance and germination rate. This unit adopts advanced wrapping technology to compound seeds with various auxiliary materials to form seed balls with specific functions. The ratio of the seed wrapping materials is: seeds: water-retaining agent: nutrient agent: adhesive = 1: 0.5 - 2: 0.1 - 0.5: 0.1 - 0.4, which is adjusted according to the seed type and the environmental conditions of the target area.

[0122] The water-retaining agent can absorb and slowly release water to provide a stable water supply for seeds in arid environments; the nutrient agent contains the basic nutrients required for plant growth to promote the early growth of seedlings; the adhesive, on the one hand, increases the weight and structural stability of the seeds to improve the wind resistance, and on the other hand, binds with soil particles to prevent the seeds from being carried away by subsequent winds. The seeds treated by wrapping can still maintain a retention rate of more than 90% under the condition of an 8 - level wind (17.2 - 20.7m / s), significantly improving the seeding success rate in harsh environments.

[0123] The delivery trajectory optimization unit 43 is used to plan the optimal delivery path and flight parameters. Based on the microenvironment model of the target area and the sand barrier layout plan, this unit calculates the optimal position, time, and density of seed delivery. The system adopts a grid division strategy, divides the target area into several sub-areas, and determines personalized delivery parameters according to the characteristics of each sub-area.

[0124] The seed delivery density can be precisely controlled within the range of 500 - 2000 seeds per mu, and is dynamically adjusted according to vegetation type, soil conditions, and expected survival rate. The system will preferentially increase the delivery density in the microenvironment suitable for vegetation growth on the leeward side of the sand barrier, and at the same time consider the spatial distribution pattern of the vegetation to avoid over-concentration or over-dispersion.

[0125] In terms of flight path planning, the system considers the collaborative operation of multiple drones. Through spatio-temporal scheduling algorithms, it optimizes the overall flight efficiency and coverage integrity. The path planning not only considers geometric coverage, but also combines wind field prediction and sunlight changes to select the most suitable delivery time window to maximize the germination probability of seeds.

[0126] Preferably, the precise seed aerial seeding control module 4 also has an adaptive feedback ability, which can dynamically adjust subsequent delivery strategies according to the delivery effect data obtained in real time. For example, if it is detected that the seed retention rate in a certain area is lower than expected, the system will automatically increase the delivery density in this area or adjust the delivery parameters; if it is found that the seed germination rate is particularly high under certain microenvironment conditions, the system will preferentially select areas with similar conditions for delivery to achieve the optimal allocation of resources.

[0127] Example 7

[0128] This example details the control parameters of the precise seed aerial seeding control module 4, including seed delivery density, delivery height, and the ratio of seed wrapping materials.

[0129] The seed delivery density can be precisely controlled within the range of 500 - 2000 seeds per mu. The delivery density is a key parameter affecting the vegetation restoration effect and needs to be determined comprehensively according to various factors. For herbaceous plants with fast growth rate and strong tillering ability (such as Astragalus adsurgens Pall. and Artemisia desertorum Spreng.), the suitable delivery density is 500 - 800 seeds per mu; for shrub plants with slow growth but strong stress resistance (such as Caragana korshinskii Kom. and Hippophae rhamnoides L.), the suitable delivery density is 800 - 1200 seeds per mu; for rare plants with low germination rate, the delivery density can be increased to 1500 - 2000 seeds per mu. In addition, the delivery density also needs to be adjusted according to soil conditions: in semi-fixed sandy land with better fertility, the density can be appropriately reduced; in barren mobile sand dunes, the density needs to be increased to ensure a sufficient number of surviving plants.

[0130] The system adopts a variable-density delivery strategy and automatically adjusts the delivery density at different locations according to the micro-environment suitability distribution. In micro-regions with better moisture conditions, such as the leeward side of sand barriers and low-lying terrains, the system will increase the delivery density to maximize the resource utilization efficiency; in extremely harsh micro-environments such as the top of sand dunes, the density will be appropriately reduced to avoid resource waste.

[0131] The delivery height can be automatically adjusted within the range of 5 - 20m. The delivery height directly affects the dispersion range and drift degree of seeds, and needs to be optimized according to wind conditions and target accuracy. Under the condition of low wind speed (≤ level 3, 3.4 - 5.4m / s), the system adopts a higher delivery height of 15 - 20m to obtain a larger coverage range; under the condition of medium wind speed (level 4 - 5, 5.5 - 10.7m / s), a medium delivery height of 10 - 15m is adopted to balance the coverage range and accuracy; under the condition of strong wind (≥ level 6, 10.8 - 17.1m / s), the height is reduced to 5 - 10m to reduce the influence of wind and improve the delivery accuracy.

[0132] The system also takes into account the influence of terrain undulation on the delivery height, appropriately increases the flight height at terrain protrusions such as the ridge of sand dunes, maintains the stability of the relative height to the ground, and ensures the uniformity of the delivery density. For particularly complex terrains, the system will adopt a terrain-following mode to adjust the flight height in real time to maintain the best delivery effect.

[0133] The proportion of the seed coating materials is prepared as seed: water-retaining agent: nutrient agent: binder = 1: 0.5 - 2: 0.1 - 0.5: 0.1 - 0.4. The characteristics of different plant seeds vary, and personalized coating formulas are required. For smaller herbaceous plant seeds (such as Agriophyllum squarrosum and Salsola collina), the proportion of the water-retaining agent is relatively high (1: 1.5 - 2) to increase the weight and water-holding capacity; for larger shrub seeds (such as Hedysarum scoparium and Elaeagnus angustifolia), the proportion of the water-retaining agent is moderate (1: 0.8 - 1.2); for seeds with a certain natural coating structure (such as Haloxylon ammodendron), the proportion of the water-retaining agent is relatively low (1: 0.5 - 0.8).

[0134] The nutrient agent mainly includes three major elements of nitrogen, phosphorus, and potassium and trace elements such as iron, zinc, and boron. The proportion is adjusted according to plant requirements and soil conditions. In extremely barren pure sandy land, the proportion of the nutrient agent is increased (1: 0.3 - 0.5); in semi-fixed sandy land with a certain amount of organic matter, the proportion is appropriately reduced (1: 0.1 - 0.2). The binder mainly uses degradable organic materials such as modified starch and cellulose derivatives, which should not only ensure sufficient bonding strength but also not affect seed germination and root growth.

[0135] Preferably, the system will continuously optimize the package formula based on historical data and real-time feedback to form a formula library suitable for different plant species and different environmental conditions. For special difficult areas, the system will also add special functional additives, such as salt-resistant components, microbial growth promoters, etc., to further improve the success rate of vegetation restoration.

[0136] Example 8

[0137] Refer to Figure 6 , the closed-loop feedback and dynamic optimization module 5 in this embodiment includes a regular monitoring unit 51, an evaluation index calculation unit 52, and a self-learning optimization unit 53.

[0138] The regular monitoring unit 51 collects data on the treatment effect at three levels of short-term, medium-term, and long-term frequencies. This unit is responsible for planning and executing monitoring tasks, and collecting data on environmental changes and treatment effects in the treatment area. The monitoring adopts a multi-sensor collaborative strategy, combining drone aerial photography, ground sensor network, and manual sampling to comprehensively capture the changes during the treatment process.

[0139] The monitoring data mainly includes four aspects: sand barrier status (integrity, deformation degree), wind-sand activity (wind deposition amount, wind erosion amount), vegetation status (coverage, biomass, species composition), and environmental parameters (wind speed, precipitation, temperature). The system will reasonably arrange the monitoring frequency according to the data type and change rate, maximizing the resource utilization efficiency while ensuring the data quality.

[0140] The evaluation index calculation unit 52 is used to calculate evaluation indexes such as the sand fixation rate, vegetation coverage, biomass, seed germination rate, and cost-benefit ratio. This unit converts the original monitoring data into standardized evaluation indexes for quantitative evaluation and comparison of the treatment effect.

[0141] The sand fixation rate is the core index to measure the sand fixation effect of the sand barrier, and the calculation formula is:

[0142]

[0143] Among them, FixationRate is the sand fixation rate, S erosion is the wind erosion area during the monitoring period, and S total is the total monitoring area.

[0144] Vegetation coverage and biomass are important indexes to evaluate the vegetation restoration effect. Vegetation coverage is calculated through drone multi-spectral images, and biomass is obtained through a method combining quadrat surveys and remote sensing estimation. The seed germination rate is a direct index to evaluate the aerial seeding effect, and the calculation formula is:

[0145]

[0146] Among them, GerminationRate is the seed germination rate, N seedling is the number of seedlings germinated, N seed is the number of seeds released.

[0147] The cost-benefit ratio is a comprehensive indicator for evaluating the economic efficiency of governance, and the calculation formula is:

[0148]

[0149] Among them, CostBenefitRatio is the cost-effectiveness ratio, B ecological For ecological benefits (such as carbon sequestration and soil and water conservation value), B economic is the economic benefit (such as the output value of biological products), C material is the material cost, C operation For running costs.

[0150] The self-learning optimization unit 53 is used to continuously improve the sand barrier aerial seeding coordination strategy based on time series data. The unit uses machine learning methods to mine patterns from historical governance data and continuously update and improve the decision model. The system automatically adjusts the model parameters by comparing the difference between the predicted results and the actual results to improve the prediction accuracy and decision optimization effect.

[0151] The self-learning process includes four steps: data collection, pattern recognition, model update, and strategy optimization. In the data collection stage, the system integrates multi-source monitoring data to form a standardized data set; in the pattern recognition stage, data mining algorithms are applied to identify the correlation patterns between environmental conditions, governance measures, and effects; in the model update stage, the prediction model parameters are adjusted according to the new data; in the strategy optimization stage, an improved governance strategy is generated based on the updated model.

[0152] Preferably, the closed-loop feedback and dynamic optimization module 5 adopts a distributed architecture to delegate some processing tasks to edge devices, reduce the burden on the central system, and improve response speed. At the same time, the system has established a multi-level security mechanism, including data encryption, access control, and anomaly detection, to ensure the safe and stable operation of the system. In addition, the system also has the ability to work offline, and can still perform basic monitoring and optimization functions in remote areas with unstable network connections.

[0153] Example 9

[0154] This embodiment describes in detail the monitoring frequency setting of the periodic monitoring unit 51, including short-term monitoring, medium-term monitoring, long-term monitoring and emergency monitoring.

[0155] The short-term monitoring frequency is once every 7 days. The short-term monitoring mainly focuses on the status of sand barriers and the initial vegetation conditions, which is the basis for the system's rapid response and adjustment. In the initial stage of treatment (the first 1 - 2 months), short-term monitoring is particularly important as it can promptly detect and solve problems. The short-term monitoring mainly adopts a combination of drone aerial photography and ground observations at key points, focusing on the integrity of sand barriers, the distribution of wind-blown sand, and the seed germination situation.

[0156] The monitoring data includes the intact rate of sand barriers, the change in sand surface height, the density and distribution of germination points, etc. During the data collection process, the system will automatically identify abnormal areas and conduct high-resolution key sampling to ensure early detection and handling of problems. The short-term monitoring results are mainly used to fine-tune the sand barrier parameters and supplement aerial seeding, which belongs to the optimization and adjustment at the tactical level.

[0157] The medium-term monitoring frequency is once every 30 days. The medium-term monitoring mainly focuses on the overall change trend of the treatment area and the vegetation growth status, which is the main basis for the system to evaluate the phased effectiveness. The medium-term monitoring adopts a comprehensive census strategy to conduct systematic data collection on the entire treatment area, including detailed topographic changes, vegetation parameters, and environmental indicators.

[0158] The monitoring content includes changes in dune morphology (volume, ridge line position), vegetation parameters (height, coverage, biomass), and environmental indicators (soil moisture, organic matter content), etc. The medium-term monitoring results are used to evaluate the collaborative effect of sand barrier and aerial seeding, and to guide the strategy adjustment in the next stage, which belongs to the periodic assessment at the strategic level.

[0159] The long-term monitoring frequency is once every 90 days. The long-term monitoring mainly focuses on the overall recovery of the ecosystem and the sustainability of the treatment, which is an important means for the system to evaluate the final effectiveness. In addition to the conventional environmental and vegetation parameters, the long-term monitoring also adds in-depth indicators such as biodiversity, ecosystem functions, and socio-economic impacts.

[0160] The monitoring content includes changes in species composition, the degree of soil improvement, ecosystem service functions (such as carbon sequestration and oxygen release, soil and water conservation), and the improvement of local residents' livelihoods, etc. The long-term monitoring results are used to evaluate the comprehensive benefits of the treatment project and guide future large-scale planning decisions.

[0161] Emergency monitoring is triggered when the environmental parameter changes exceed 20%. Emergency monitoring is the system's rapid response mechanism to emergencies and abnormal changes. When the system detects significant changes (exceeding the threshold of 20%) in key environmental parameters (such as wind speed, precipitation, temperature), it will automatically trigger the emergency monitoring process to quickly assess the potential impact of the changes on the treatment area.

[0162] Emergency monitoring adopts a high-frequency sampling strategy in key areas to quickly obtain key data. By comparing with historical data, it assesses the severity and possible consequences of abnormal changes. The system will adjust the sand barrier parameters in a timely manner or take temporary protective measures according to the evaluation results to minimize the adverse effects. Common triggering events include abnormal meteorological events such as sandstorms, heavy rainfall, and extreme high temperatures.

[0163] Preferably, the monitoring frequency is not fixed but dynamically adjusted according to the treatment stage and actual needs. In the initial stage of treatment, the system tends to conduct more frequent monitoring to promptly address various problems. As the treatment enters the stable stage, the monitoring frequency can be appropriately reduced to minimize resource consumption. In addition, the system will also automatically adjust the monitoring strategy according to seasonal changes, such as increasing the monitoring frequency before the windy season and focusing on monitoring the vegetation status during the plant growth season.

[0164] Example 10

[0165] Refer to Figure 7 , the optimization method for collaborative layout of sand barrier aerial seeding based on UAV swarm game provided in this embodiment includes the following steps:

[0166] Step S1: Collect multi-dimensional environmental data of the sandy land and construct a micro-environment digital model.

[0167] In this step, multiple UAVs equipped with multi-spectral cameras, thermal infrared sensors, and millimeter-wave radars simultaneously collect multi-dimensional data of the sandy land. Using a four-dimensional scanning strategy (three-dimensional space + one-dimensional time), detailed environmental data of the target area is obtained. The collected data includes terrain data (digital elevation model with a resolution of 5 cm), wind field data (three-dimensional wind field with a frequency of 0.1 - 10 Hz), soil property data (humidity, temperature, texture), etc. Subsequently, through a multi-source data fusion algorithm, data from different sources and different scales are integrated into a unified micro-environment digital model to provide basic data support for subsequent analysis.

[0168] Step S2: Based on the micro-environment digital model, establish a spatio-temporal correlation model between the sand barrier's regulation of the wind field and vegetation growth.

[0169] In this step, first, based on the principle of computational fluid dynamics (CFD), a sand barrier - wind field interaction model is constructed to predict the influence of different forms of sand barriers on wind speed distribution and turbulence characteristics. Then, an environmental response model for desert vegetation growth is constructed to determine the micro-environmental conditions required for seed germination and seedling growth. Finally, through a convolutional neural network algorithm, an end-to-end prediction model of sand barrier parameters - micro-environmental conditions - vegetation response is established to realize the spatio-temporal correlation analysis of sand barrier layout and vegetation growth.

[0170] Step S3: Generate an optimal sand barrier configuration plan according to the spatio-temporal correlation model.

[0171] Based on the analysis results of the spatio-temporal correlation model, this step generates an optimal sand barrier configuration plan through multi-objective optimization and reinforcement learning algorithms. The system balances three objectives: sand fixation efficiency, vegetation growth, and cost control, and seeks the best balance point. The configuration plan includes parameters such as sand barrier height (continuously adjustable from 20 to 50 cm), sand barrier density (ventilation rate continuously adjustable from 30% to 70%), sand barrier material (wheat straw, reed, shrub branches), and sand barrier arrangement pattern (square, strip, checkerboard, honeycomb, herringbone, cross). The optimization process adopts a hierarchical strategy, first determining the regional layout at the macroscopic scale and then finely adjusting specific parameters at the microscopic scale.

[0172] Step S4: Based on the sand barrier configuration plan, adjust the seed placement parameters to control the seed placement position, time, and density.

[0173] This step accurately controls the seed placement position, time, and density according to the optimal sand barrier configuration plan and real-time wind field data. The system adopts intelligent seed placement technology, which can sense the changes in wind speed and direction in real time and adjust the placement height (automatically adjustable from 5 to 20 m), placement rate, and placement direction accordingly. At the same time, the seed coating technology is adopted to mix seeds with water-retaining agents, nutrient agents, and adhesives (the ratio is seed: water-retaining agent: nutrient agent: adhesive = 1: 0.5 - 2: 0.1 - 0.5: 0.1 - 0.4) to improve wind resistance and germination rate. The placement density is accurately controlled within the range of 500 - 2000 seeds per mu, and is dynamically adjusted according to the micro-environmental conditions, increasing the placement density in the micro-region most suitable for vegetation growth.

[0174] Step S5: Collect the data on the treatment effect and evaluate the sand fixation efficiency of the sand barrier and the vegetation growth status.

[0175] This step systematically collects the environmental changes and treatment effect data in the treatment area according to the three-level monitoring frequencies of short-term (once every 7 days), medium-term (once every 30 days), and long-term (once every 90 days). The monitoring contents include the sand barrier status, sand and wind activities, vegetation status, and environmental parameters, etc. The system converts the original monitoring data into standardized evaluation indicators, including sand fixation rate, vegetation coverage, biomass, seed germination rate, and cost-benefit ratio, etc., which is convenient for the quantitative evaluation and comparison of the treatment effect. When the environmental parameter changes exceed 20%, emergency monitoring will also be triggered to quickly evaluate the impact of abnormal changes.

[0176] Step S6: According to the treatment effect data, update the sand barrier configuration plan and seed placement parameters to form a closed-loop optimization.

[0177] Based on the evaluation results of governance effects, this step uses machine learning methods to mine patterns from historical data, continuously updating and improving the decision-making model. The system compares the differences between the predicted results and the actual effects, automatically adjusts the model parameters, and improves the prediction accuracy and decision-making optimization effects. The optimization process includes four links: data collection, pattern recognition, model update, and strategy optimization. Finally, an improved sand barrier configuration plan and seed placement parameters are generated to achieve the continuous optimization of the coordinated strategy of sand barrier aerial seeding.

[0178] By constructing a complete perception - analysis - decision - execution - optimization closed-loop mechanism, this method realizes the coordinated optimization of sand barrier layout and seed aerial seeding, significantly improving the efficiency and quality of desertification control. The system can not only adapt to the complex and changeable desert environment but also has the ability of self-learning and continuous optimization, representing an important development direction of desertification control technology.

[0179] Example 11

[0180] This example introduces the actual application case of the system in the Mu Us Sandy Land and its effect evaluation.

[0181] The test area is located at the southern edge of the Mu Us Sandy Land in Wushen Banner, Ordos City, Inner Mongolia Autonomous Region, with a total area of about 300 hectares. The main landform types are alternating semi-fixed dunes and mobile dunes. The average annual precipitation in this area is about 330 mm, the average annual evaporation is about 2100 mm, and the average annual wind speed is 3.5 m / s. It belongs to a typical arid and semi-arid climate region. The main vegetation in the area is drought-tolerant plants such as Artemisia desertorum, Salix psammophila, and Astragalus adsurgens, and the vegetation coverage is less than 15%, and the ecological system is fragile.

[0182] The specific process of applying this system for desertification control is as follows:

[0183] First, deploy a cluster system composed of 3 multi-functional unmanned aerial vehicles (UAVs), which are respectively responsible for three main tasks: environmental perception, sand barrier layout, and seed aerial seeding. In the system initialization stage, the environmental perception UAV conducts a detailed scan of the entire test area, collects high-resolution terrain data, wind field data, and soil property data, and constructs an accurate micro-environment digital model.

[0184] Based on the analysis of the micro-environment model, the system identifies the hotspots of wind-sand activities and potential vegetation suitable areas in the region, and generates a preliminary sand barrier layout plan. According to the regional wind field characteristics (the dominant wind direction is northwest, and the secondary dominant wind direction is southeast) and terrain features (most of the dune ridge lines are in the northeast - southwest direction), the system recommends a mixed arrangement mode mainly composed of honeycomb and herringbone, the sand barrier height is set at 35 - 45 cm, the ventilation rate is 40% - 60%, and the local abundant reed is selected as the material.

[0185] After the sand barriers were set up, the system formulated a differentiated aerial seeding strategy based on the results of the microenvironment assessment. For the suitable areas of the leeward side microenvironment of the sand barriers, Artemisia desertorum and Salix psammophila seeds were selected, and the seeding density was 1,200 - 1,500 seeds per mu; for the low-lying areas of the sand dunes, Astragalus adsurgens and Caragana korshinskii seeds were selected, and the seeding density was 900 - 1,200 seeds per mu; for the harsh areas at the top of the sand dunes, Haloxylon ammodendron seeds were selected, and the seeding density was 600 - 800 seeds per mu. All seeds were treated with wrapping, and the ratio was seeds: water-retaining agent: nutrient agent: binder = 1:1.2:0.3:0.25.

[0186] After the system had been running for one year, a comprehensive assessment of the treatment effect was carried out through a three-level monitoring system. The results showed that:

[0187] 1. The sand fixation effect of the sand barriers was remarkable, with the sand fixation rate reaching 87.6%, the wind erosion amount reduced by 72.3%, and the sand dune movement speed decreased from the original 3.5 m / year to 0.8 m / year;

[0188] 2. The vegetation restoration effect was good, with the average seed germination rate reaching 68.4%, much higher than 35% - 40% of the traditional method; the vegetation coverage increased from the initial less than 15% to 42.3%; the biomass increased from 45 g / m 2 to 186 g / m 2 ;

[0189] 3. The biodiversity was significantly improved, with the number of plant species monitored increasing from 8 to 23, and the number of insect species increasing from 12 to 37;

[0190] 4. The soil quality was improved, with the surface organic matter content increasing from 0.12% to 0.38%, and the soil stability was significantly enhanced;

[0191] 5. The treatment cost decreased, with the treatment cost per hectare saving about 42.7% compared with the traditional method, and the return on investment increased by 65.3%.

[0192] Compared with the traditional single sand barrier treatment or conventional aerial seeding technology, this system showed significant advantages in terms of sand fixation efficiency, vegetation survival rate, and cost-effectiveness. Especially in terms of adaptability under extreme weather conditions, the system demonstrated excellent performance. During the test period, the area encountered a 6 - level strong wind (wind speed reached 13.5 m / s), and the seed loss rate in the traditional treatment area was as high as 85%, while the seed retention rate in the treatment area of this system remained at 92.3%, fully proving the wind resistance performance and collaborative optimization effect of the system.

[0193] In addition, through the analysis of the data of the system running for one year, it is found that the collaborative optimization of sand barriers and aerial seeding has produced obvious synergistic effects. The sand barriers not only directly fix the sand surface, but also create a microenvironment suitable for vegetation growth; the growth of vegetation in turn enhances the stability of the sand surface and extends the service life of the sand barriers, forming a positive feedback loop. This kind of synergy cannot be achieved by a single technical means, which proves the innovation value and application prospect of this system.

[0194] In conclusion, the application case of this system in the Mu Us Sandy Land fully verifies the actual effect of the collaborative layout optimization system of sand barriers and aerial seeding based on the game of unmanned aerial vehicle clusters, providing a new technical path for the comprehensive management of desertification.

[0195] The specific embodiments of the present invention have been described in detail above. However, it should be understood that those skilled in the art can make various changes, modifications, substitutions and variations to these embodiments without departing from the principle and concept of the present invention. The protection scope of the present invention shall be subject to the appended claims.

Claims

1. The collaborative layout optimization system for sand barrier aerial seeding based on UAV swarm game, characterized in that, Including: A multi-modal desert micro-environment perception module, which is used to collect multi-dimensional environmental data of the sandy land and construct a micro-environment digital model including terrain, wind field, humidity, and temperature; A sand barrier-vegetation coupling dynamic model module, which is data-connected to the multi-modal desert micro-environment perception module and is used to receive the micro-environment digital model and construct a spatio-temporal correlation model of the wind field regulation by the sand barrier and vegetation growth; An adaptive sand barrier parameter optimization module, which is data-connected to the sand barrier-vegetation coupling dynamic model module and is used to generate an optimal sand barrier configuration plan based on the spatio-temporal correlation model. The sand barrier configuration plan includes parameters such as sand barrier height, density, material, and arrangement pattern; A precise seed aerial seeding control module, which is data-connected to the adaptive sand barrier parameter optimization module and is used to receive the sand barrier configuration plan, adjust the seed delivery parameters according to real-time wind field data, and control the seed delivery position, time, and density; A closed-loop feedback and dynamic optimization module, which is data-connected to the precise seed aerial seeding control module and the multi-modal desert micro-environment perception module, is used to collect data on the treatment effect, evaluate the sand fixation efficiency of the sand barrier and the vegetation growth status, and feedback optimization suggestions to the adaptive sand barrier parameter optimization module.

2. The system according to claim 1, wherein The multi-modal desert micro-environment perception module includes: A multi-spectral data acquisition unit, which is used to collect surface information in visible light, near-infrared, and short-wave infrared bands; A wind field monitoring unit, which is used to collect three-dimensional wind field data with a frequency of 0.1-10 Hz; A terrain mapping unit, which is used to obtain a digital elevation model with a resolution of 5 cm; A data fusion processing unit, which is used to integrate multi-source data to generate a unified micro-environment digital model.

3. The system according to claim 1, wherein The sand barrier-vegetation coupling dynamic model module includes: A wind field regulation model unit, which is used to simulate the influence of different sand barrier forms on the wind field; A vegetation growth condition model unit, which is used to determine the micro-environment threshold conditions required for vegetation growth; A spatio-temporal correlation calculation unit, which is used to establish a correlation model between the sand barrier layout and vegetation growth through a convolutional neural network algorithm.

4. The system according to claim 1, wherein The adaptive sand barrier parameter optimization module includes: A sand barrier specification library unit, which contains sand barrier parameters of 6 basic forms and 3 material combinations; A multi-objective optimization unit, which is used to balance the goals of sand fixation efficiency, vegetation growth, and cost control; A reinforcement learning unit, which is used to adaptively adjust the sand barrier parameters to adapt to different micro-environment conditions.

5. The system according to claim 4, wherein The sand barrier parameters in the sand barrier specification library unit include: A sand barrier height parameter, with a continuously adjustable range of 20-50 cm; A sand barrier density parameter, with a continuously adjustable ventilation rate range of 30% - 70%; A sand barrier material parameter, including wheat straw, reed, and shrub branches; A sand barrier arrangement pattern parameter, including square, strip, checkerboard, honeycomb, herringbone, and cross-shaped.

6. The system according to claim 1, wherein The precise seed aerial seeding control module includes: An intelligent seed delivery unit, which can automatically adjust the delivery parameters according to real-time wind field data; A seed wrapping treatment unit, which is used to mix seeds with water-retaining agents, nutrient agents, and adhesives to improve wind resistance; A delivery trajectory optimization unit, which is used to plan the optimal delivery path and flight parameters.

7. The system according to claim 6, wherein The control parameters of the precise seed aerial seeding control module include: The seed delivery density, which can be precisely controlled within the range of 500-2000 seeds per mu; The dropping height can be automatically adjusted within the range of 5 - 20m; The proportion of the seed wrapping material is prepared as seed: water retaining agent: nutrient agent: binder = 1: 0.5 - 2: 0.1 - 0.5: 0.1 - 0.

4.

8. The system according to claim 1, wherein The closed-loop feedback and dynamic optimization module includes: A regular monitoring unit that collects data on the treatment effect at three levels of short-term, medium-term, and long-term frequencies; An evaluation index calculation unit for calculating the sand fixation rate, vegetation coverage, biomass, seed germination rate, and cost-benefit ratio; A self-learning optimization unit for continuously improving the synergistic strategy of sand barrier aerial seeding based on time series data.

9. The system according to claim 8, wherein The monitoring frequency of the regular monitoring unit is set as: Short-term monitoring, once every 7 days; Medium-term monitoring, once every 30 days; Long-term monitoring, once every 90 days; When the environmental parameters change by more than 20%, emergency monitoring is triggered.

10. The collaborative layout optimization method for sand barrier aerial seeding based on the game of UAV swarms adopts the system of any one of claims 1-9, and is characterized in that It includes the following steps: Collect multi-dimensional environmental data of the sandy land and construct a micro-environment digital model; Based on the micro-environment digital model, establish a spatio-temporal correlation model of the wind field regulation and vegetation growth by the sand barrier; Generate an optimal sand barrier configuration plan according to the spatio-temporal correlation model; Based on the sand barrier configuration plan, adjust the seed dropping parameters to control the seed dropping position, time, and density; Collect data on the treatment effect and evaluate the sand fixation efficiency of the sand barrier and the vegetation growth status; Update the sand barrier configuration plan and the seed dropping parameters according to the treatment effect data to form a closed-loop optimization.

Images