Ecological monitoring method and system for restoration of degraded grassland vegetation

Abstract

The present invention discloses an ecological monitoring method and system for restoration of vegetation in degraded grasslands, which relates to the field of ecological monitoring technology, including collecting environmental data of degraded grasslands, extracting environmental features and calculating the contribution of environmental features; establishing a plant species list according to the contribution of environmental features, and optimizing the selection of plant species and planting ratios from the plant species list to form a plant planting plan; planning planting paths based on the plant planting plan for precise planting; displaying the plant planting plan and planting paths and storing them in a database. The present invention greatly improves the correlation and adaptability between plant planting and the environment, improves the survival rate of plants, and simultaneously improves the efficiency of plant planting by matching the degraded grassland planting area to adapt to plants and planning plant planting paths.

Description

Technical Field

[0001] The invention relates to the technical field of ecological monitoring, and in particular to an ecological monitoring method and system for restoration of degraded grassland vegetation. Background Art

[0002] With the increasing global environmental problems, grassland degradation has become one of the main challenges facing the global ecosystem. As an important ecosystem, grassland not only undertakes functions such as soil and water conservation, carbon sink regulation and biodiversity maintenance, but is also an important foundation for the development of animal husbandry. However, due to factors such as overgrazing, unreasonable development and climate change, grassland degradation is becoming more and more serious, leading to the loss of ecological functions and a decline in productivity. Therefore, restoring degraded grassland vegetation and improving the ecological environment have become one of the hot topics in global research. At present, a variety of technologies and methods have been applied to grassland vegetation restoration, including traditional planting and artificial restoration technologies, such as seed sowing, fertilizer application and irrigation regulation; as well as intelligent and precise planting technologies developed in recent years, such as sowing and monitoring using drones. These technologies have improved the efficiency of grassland restoration to a certain extent. However, due to the complexity of the environment and the diversity of grassland degradation, the existing technologies still have certain limitations in dealing with grassland restoration effects under different environmental conditions. The existing technologies usually only select plants and formulate planting strategies based on shallow data, lack of in-depth analysis of the degraded grassland environment, resulting in insufficient matching between plant species selection and actual environmental conditions. Summary of the invention

[0003] In view of the problems existing in the above-mentioned existing ecological monitoring methods and systems for restoration of degraded grassland vegetation, the present invention is proposed.

[0004] Therefore, the problem to be solved by the present invention is that the existing technology usually only selects plants and formulates planting strategies based on shallow data, lacks in-depth analysis of the degraded grassland environment, resulting in insufficient matching between plant species selection and actual environmental conditions.

[0005] To solve the above technical problems, the present invention provides the following technical solutions: an ecological monitoring method for restoration of vegetation in degraded grasslands, which includes collecting environmental data of degraded grasslands, extracting environmental characteristics and calculating the contribution of environmental characteristics; establishing a plant species list according to the contribution of environmental characteristics, and optimizing the selection of plant species and planting ratios from the plant species list to form a plant planting plan; planning planting paths based on the plant planting plan for precise planting; and displaying the plant planting plan and planting paths and storing them in a database.

[0006] As a preferred solution of the ecological monitoring method for degraded grassland vegetation restoration described in the present invention, wherein: the collecting of degraded grassland environmental data, extracting environmental features and calculating environmental feature contribution refers to deploying a sensor network in the degraded grassland area to collect environmental data, cleaning the environmental data collected by the sensor network, and smoothing the data using a sliding average method and then standardizing the data, calculating the covariance matrix of the standardized environmental data, solving the eigenvalues ​​and eigenvectors of the covariance matrix, setting a threshold, and taking the eigenvector whose cumulative explained variance is greater than the threshold as the environmental eigenvector;

[0007] Construct the initial decision tree, set the number of decision trees to 100, and the maximum tree depth to 10. Obtain the historical environmental feature vector as the training set to train the initial decision tree to build a random forest model. During the training process, record the decision tree splitting process in real time, calculate the variance of the split child decision tree corresponding to the environmental feature vector, and calculate the contribution value of the variance reduction:

[0008] ,

[0009] in is the contribution of the environmental feature vector f to the reduction of model variance when splitting the decision tree node, is the variance of the decision tree before splitting, is the number of environmental feature vectors in the decision tree before splitting, and are the number of environmental feature vectors in the left and right child decision trees of the decision tree split, and is the variance of the left and right child decision trees;

[0010] When the environmental feature vector f is used for decision tree splitting, its contribution to the reduction of model variance is recorded, and the total contribution of the environmental feature vector f to the reduction of model variance during training is obtained comprehensively. ;

[0011] Get the total contribution value of each environmental feature vector and normalize it to get the feature contribution of each environmental feature vector .

[0012] As a preferred solution of the ecological monitoring method for degraded grassland vegetation restoration of the present invention, wherein: the establishment of a plant species list according to environmental feature contribution refers to constructing a feature contribution matrix C after obtaining the feature contribution of each environmental feature vector;

[0013] Select the radial basis function kernel to build the SVM model and optimize it based on the feature contribution matrix C:

[0014] ,

[0015] in is the optimized radial basis function kernel function, is the kernel function parameter, and are the input environment feature vector and the model training environment feature vector respectively;

[0016] The kernel function parameters and multiplication parameters are adjusted through cross-validation, and the loss function is defined, and the optimization goal is set to minimize the loss function;

[0017] Obtain plant planting environment data and extract plant environment feature vectors as training data to input into the SVM model for iterative optimization of model parameters;

[0018] The environmental feature vector is input into the trained SVM model, and the SVM model outputs the suitability of each plant, sets a suitability threshold, and sorts them from high to low according to suitability, and forms a plant species list with plants above the suitability threshold.

[0019] As a preferred embodiment of the ecological monitoring method for restoration of vegetation in degraded grasslands of the present invention, the step of optimizing the selection of plant species and planting ratios from the plant species list to form a plant planting plan refers to obtaining the plant species list, generating initial populations according to different plant species and planting ratios according to the plant species list, and calculating the plant diversity of each initial population using the Shannon diversity index according to the planting ratio:

[0020] ,

[0021] in is the plant diversity of the initial population, is the number of plant species in the initial population, is the planting ratio of the i-th plant;

[0022] Calculate the plant ecological suitability of each initial population based on the plant planting ratio:

[0023] ,

[0024] in is the ecological suitability of the initial population, is the fitness of the i-th plant;

[0025] The fitness function F is constructed based on the plant diversity and plant ecological suitability of the initial population:

[0026] ,

[0027] The fitness of each initial population is calculated, and the roulette wheel selection method is used to select the individual with the highest fitness for crossover and mutation operations, and a new generation of population is generated. The new population is iteratively generated, and the fitness of each new population is calculated. When the fitness improvement is lower than the preset threshold, the iteration is stopped, and the individual with the highest fitness in the new population is taken as the optimal population. The plant species and planting ratios in the optimal population are extracted as the plant planting plan.

[0028] As a preferred scheme of the ecological monitoring method for restoration of vegetation in degraded grasslands described in the present invention, wherein: the planning of planting paths based on plant planting schemes for precise planting refers to using a drone equipped with a LiDAR sensor and a multispectral camera to scan the degraded grassland area to generate a digital elevation model of the degraded grassland, and calculating the humidity through the normalized difference vegetation index to form a humidity distribution map, extracting the light intensity through the multispectral image to form a light distribution map, converting the digital elevation model of the degraded grassland, the humidity distribution map and the light distribution map into raster map formats respectively and superimposing them to form a multidimensional environmental feature map;

[0029] The principal component analysis of the elevation, humidity and light contained in each grid in the multidimensional environmental characteristic map is performed to extract the principal component feature vector, and the principal component feature vector is input into the random forest model to obtain the principal component feature contribution and form a principal component feature contribution matrix. The SVM model is optimized based on the principal component feature contribution matrix, and the principal component feature vector of each grid is input into the optimized SVM model to determine the plant suitability of each grid;

[0030] The total number of grids for each plant is calculated based on the plant planting ratio and grid number in the plant planting plan, and the optimization goal is defined as maximizing the comprehensive suitability. :

[0031] ,

[0032] in is the total number of plant species in the plant planting plan, n is the number of grids, is the suitability of planting the kth plant in the jth grid, is the planting ratio of the kth plant;

[0033] Iteratively adjust the grid plant species and calculate the comprehensive suitability after each adjustment Until the comprehensive suitability Stop after reaching the maximum value, and output the comprehensive suitability The grid plants are planted and labeled;

[0034] The planting grids of each plant are extracted to form a plant grid map. Each grid in the plant grid map is used as a node, and the shortest path optimization algorithm is used to plan the planting path of each plant.

[0035] As a preferred scheme of the ecological monitoring method for degraded grassland vegetation restoration described in the present invention, the display of plant planting plans and planting paths refers to associating the plant planting plans and planting paths and visually displaying them, and inputting the plant planting plans and planting paths into a convolutional neural network to obtain a completed plan view of plant planting in the degraded grassland area for synchronous display.

[0036] As a preferred scheme of the ecological monitoring method for restoration of degraded grassland vegetation described in the present invention, the storage in the database refers to obtaining the plant planting plan and the planting path of each plant and storing them in the database, and classifying and storing them according to timestamps. The database regularly performs integrity checks on the stored data and synchronously uploads the stored data to the cloud.

[0037] Another object of the present invention is to provide an ecological monitoring system for restoration of degraded grassland vegetation, which comprises:

[0038] A data acquisition module is used to collect environmental data of degraded grassland areas through a sensor network and extract environmental data features to calculate feature contributions;

[0039] A scheme optimization module is used to form a plant species list according to the contribution of environmental characteristics, and calculate the fitness of the plant species list to optimize the selection of plants and planting ratios to form a plant planting scheme;

[0040] Path planning module, used to match plant species in degraded grassland areas and plan plant planting paths;

[0041] The display and storage module is used to display and store plant planting plans and plant planting paths, and obtain the completed plant planting plan through neural algorithms for simultaneous display.

[0042] A computer device comprises: a memory and a processor; the memory stores a computer program, and the processor implements the steps of the above-mentioned ecological monitoring method for restoration of degraded grassland vegetation when executing the computer program.

[0043] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned ecological monitoring method for restoration of degraded grassland vegetation.

[0044] The beneficial effects of the present invention are as follows: the present invention collects degraded grassland environmental data, extracts environmental data features, calculates contributions, and forms a contribution matrix; optimizes the SVM model according to the contribution matrix to form a plant species list; and calculates plant fitness according to the plant species list to form a plant planting plan, thereby greatly improving the correlation and adaptability between plant planting and the environment, improving plant survival rate, and simultaneously improving plant planting efficiency by matching degraded grassland planting areas to adapt to plants and planning plant planting paths. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0046] Figure 1 Schematic diagram of the process of ecological monitoring methods for vegetation restoration in degraded grasslands.

[0047] Figure 2 Schematic diagram of the structure of the ecological monitoring system for restoration of degraded grassland vegetation. DETAILED DESCRIPTION

[0048] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0049] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0050] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments. Example 1

[0051] Reference Figure 1 , which is the first embodiment of the present invention, and provides an ecological monitoring method for restoration of degraded grassland vegetation. The ecological monitoring method for restoration of degraded grassland vegetation includes:

[0052] S1. Collect degraded grassland environmental data, extract environmental characteristics and calculate the contribution of environmental characteristics;

[0053] Specifically, collecting degraded grassland environmental data, extracting environmental features and calculating environmental feature contributions refers to deploying a sensor network in the degraded grassland area to collect environmental data, wherein the environmental data includes soil moisture, temperature, light intensity, soil pH value and nutrient elements, cleaning the environmental data collected by the sensor network, smoothing the data using a sliding average method and then standardizing the data, calculating the covariance matrix of the standardized environmental data, solving the eigenvalues ​​and eigenvectors of the covariance matrix, setting a threshold, and taking the eigenvectors with a cumulative explained variance greater than the threshold as environmental eigenvectors;

[0054] Construct the initial decision tree, set the number of decision trees to 100, and the maximum tree depth to 10. Obtain the historical environmental feature vector as the training set to train the initial decision tree to build a random forest model. During the training process, record the decision tree splitting process in real time, calculate the variance of the split child decision tree corresponding to the environmental feature vector, and calculate the contribution value of the variance reduction:

[0055] ,

[0056] in is the contribution of the environmental feature vector f to the reduction of model variance when splitting the decision tree node, is the variance of the decision tree before splitting, is the number of environmental feature vectors in the decision tree before splitting, and are the number of environmental feature vectors in the left and right child decision trees of the decision tree split, and is the variance of the left and right child decision trees;

[0057] When the environmental feature vector f is used for decision tree splitting, its contribution to the reduction of model variance is recorded, and the total contribution of the environmental feature vector f to the reduction of model variance during training is obtained comprehensively. ;

[0058] Get the total contribution value of each environmental feature vector and normalize it to get the feature contribution of each environmental feature vector .

[0059] The deployment of sensor networks in degraded grassland areas has enabled comprehensive and real-time collection of a variety of environmental data, including soil moisture, temperature, light intensity, soil pH, and nutrients. The diversity and real-time nature of this data provide a scientific basis for decision-making in grassland restoration. Smoothing and standardizing the data using the sliding average method can effectively reduce noise in the data and ensure data accuracy in subsequent calculations. This data preprocessing step significantly improves the accuracy of environmental feature extraction and ensures the reliability of environmental feature contribution calculations. The calculation of the covariance matrix helps identify the relationships and dependencies between environmental features, thereby determining the most important features for grassland restoration. By solving eigenvalues ​​and eigenvectors and setting a standard where the cumulative explained variance is greater than a threshold, it is possible to screen out The eigenvector with the greatest explanatory power for environmental changes reduces the data dimension while ensuring the representativeness of the data. This method greatly improves the efficiency of model calculation while ensuring the scientificity and efficiency of the grassland restoration plan. Through the decision tree splitting process in the random forest model, the impact of environmental characteristics on the model variance is recorded in real time, allowing us to quantitatively evaluate the contribution of each environmental eigenvector to grassland restoration. The model can not only identify which environmental characteristics have an important impact on grassland restoration, but also quantify its contribution to the optimization of the overall environmental restoration plan. Finally, the total contribution value of each eigenvector is obtained and normalized to form a feature contribution matrix. This processing method ensures the quantitative analysis of each environmental characteristic and effectively improves the matching degree between plant species selection and environmental characteristics.

[0060] S2. Establish a plant species list according to the contribution of environmental characteristics, and optimize the selection of plant species and planting ratios from the plant species list to form a plant planting plan;

[0061] Specifically, establishing a plant species list according to environmental feature contributions means obtaining the feature contribution of each environmental feature vector and then constructing a feature contribution matrix C;

[0062] Select the radial basis function kernel to build the SVM model and optimize it based on the feature contribution matrix C:

[0063] ,

[0064] in is the optimized radial basis function kernel function, is the kernel function parameter, and are the input environment feature vector and the model training environment feature vector respectively;

[0065] The kernel function parameters and multiplication parameters are adjusted through cross-validation, and the loss function is defined, and the optimization goal is set to minimize the loss function;

[0066] Obtain plant planting environment data and extract plant environment feature vectors as training data to input into the SVM model for iterative optimization of model parameters;

[0067] The environmental feature vector is input into the trained SVM model, and the SVM model outputs the suitability of each plant, sets a suitability threshold, and sorts them from high to low according to suitability, and forms a plant species list with plants above the suitability threshold.

[0068] After obtaining the contribution of each environmental feature vector, the present invention organizes these data into a feature contribution matrix C, which comprehensively records the degree of influence of environmental features on plant growth, and provides a scientific basis for the subsequent optimization of the SVM model. This structured data processing method effectively improves the utilization rate of feature information and ensures high accuracy and high credibility in plant selection and adaptability analysis. Through the optimization of the feature contribution matrix, the SVM model can more accurately identify the complex relationship between environmental features and plant species. In this process, the radial basis function kernel (RBF) is selected as the kernel function. The main reason is that the RBF kernel function has a strong nonlinear mapping ability, which can map the input environmental feature vector to a high-dimensional space, realize the accurate classification of various plant environmental features, obtain plant planting environment data, and extract the environmental feature vector of each plant as training data, ensuring that the model can be iteratively optimized based on historical data. This process ensures the accuracy and generalization ability of the model, so that it can have stable prediction performance under different environmental conditions. By inputting the environmental feature vector, the SVM model can iteratively update its parameters to minimize the loss function and gradually improve the model. The method improves the prediction ability of the type. This method ensures that the model can still accurately evaluate the suitability of plants in a new environment, thereby forming a more scientific list of plant species. The trained SVM model can calculate the suitability value of each plant according to the input environmental feature vector. This suitability reflects the growth potential of plants under specific environmental conditions. By setting a suitability threshold, only plants with a suitability higher than the threshold are selected. This screening mechanism ensures that each plant in the plant species list is highly matched with the environmental characteristics, thereby improving the survival rate of plants. In addition, sorting from high to low according to suitability to form a final list of plant species helps decision makers to prioritize the most suitable plants for planting and optimize the ecological restoration process. The final list of plant species is based on the optimized SVM model. This method not only ensures a high degree of match between plant species and the environment, but also ensures the survival rate and growth quality of plants after planting. Compared with traditional methods, the plant selection process in the present invention comprehensively considers the influence of environmental characteristics and plant adaptability, thereby achieving accurate plant species screening. The plant planting plan formulated in this way can maximize the ecological effect of plants and promote the rapid recovery of degraded grasslands.

[0069] Furthermore, optimizing the selection of plant species and planting ratios from the plant species list to form a plant planting plan means that after obtaining the plant species list, initial populations are generated according to different plant species and planting ratios according to the plant species list, and the plant diversity is calculated using the Shannon diversity index according to the planting ratio in each initial population:

[0070] ,

[0071] in is the plant diversity of the initial population, is the number of plant species in the initial population, is the planting ratio of the i-th plant;

[0072] Calculate the plant ecological suitability of each initial population based on the plant planting ratio:

[0073] ,

[0074] in is the ecological suitability of the initial population, is the fitness of the i-th plant;

[0075] The fitness function F is constructed based on the plant diversity and plant ecological suitability of the initial population:

[0076] ,

[0077] The fitness of each initial population is calculated, and the roulette wheel selection method is used to select the individual with the highest fitness for crossover and mutation operations, and a new generation of population is generated. The new population is iteratively generated, and the fitness of each new population is calculated. When the fitness improvement is lower than the preset threshold, the iteration is stopped, and the individual with the highest fitness in the new population is taken as the optimal population. The plant species and planting ratios in the optimal population are extracted as the plant planting plan.

[0078] The initial population is generated according to the plant species list, and arranged and combined according to different plant species and planting ratios. This way of generating the initial population ensures the diversity of the population, so that the algorithm has enough solution space to find the optimal solution in subsequent iterations. By calculating the Shannon diversity index, the diversity level of each initial population can be accurately evaluated to ensure that the plant planting plan has better stability and resistance in the ecosystem. A higher diversity index indicates that the plant species within the population are evenly distributed, which helps to improve the long-term effect and adaptability of grassland restoration. Through the fitness function, not only can the ecological value of each population be comprehensively evaluated, but also the population can be ensured in multiple A balance is struck between diversity and suitability, and the best individuals are selected according to the fitness value. This method can avoid the selection of a single standard and ensure that the generated plant planting scheme has high ecological value and adaptability under different environmental conditions. The roulette selection method is used to select individuals with higher fitness from the population in a probabilistic manner, and crossover and mutation operations are performed. This selection mechanism ensures that individuals with higher fitness have a higher probability of reproduction, thereby accelerating the convergence of the optimal solution. At the same time, through crossover and mutation operations, the newly generated population retains high-quality genes while maintaining a certain degree of variability, thereby ensuring the diversity of the population and the exploration ability of the algorithm.

[0079] During the iterative process, the iteration is stopped when the fitness improvement is lower than the preset threshold. This setting effectively avoids over-calculation and ensures that the algorithm converges to the optimal solution within a reasonable time. Finally, the individuals with the highest fitness are extracted as the optimal population, ensuring that the plant planting plan has the best diversity and ecological suitability in the current environment. After multiple rounds of iterative optimization, the plant species and planting ratios in the optimal population are extracted to form the final plant planting plan. This plan ensures that the selected plant species are not only suitable for growth in the current environment, but also have the optimal planting ratio, further improving the success rate of plant planting and grassland restoration effects. Compared with traditional planting plans, the optimization of plant selection and planting ratios in the present invention significantly improves the accuracy and sustainability of grassland restoration, ensuring the diversity and stability of the ecosystem.

[0080] S3. Plan the planting path based on the plant planting plan for precise planting;

[0081] Specifically, planning planting paths for precise planting based on plant planting plans means using drones equipped with LiDAR sensors and multispectral cameras to scan degraded grassland areas to generate digital elevation models of degraded grasslands, and calculating humidity through normalized difference vegetation index to form a humidity distribution map, extracting light intensity through multispectral images to form a light distribution map, and converting the digital elevation model of degraded grasslands, humidity distribution maps, and light distribution maps into raster map formats respectively and superimposing them to form a multidimensional environmental feature map;

[0082] The principal component analysis of the elevation, humidity and light contained in each grid in the multidimensional environmental characteristic map is performed to extract the principal component feature vector, and the principal component feature vector is input into the random forest model to obtain the principal component feature contribution and form a principal component feature contribution matrix. The SVM model is optimized based on the principal component feature contribution matrix, and the principal component feature vector of each grid is input into the optimized SVM model to determine the plant suitability of each grid;

[0083] The total number of grids for each plant is calculated based on the plant planting ratio and grid number in the plant planting plan, and the optimization goal is defined as maximizing the comprehensive suitability. :

[0084] ,

[0085] in is the total number of plant species in the plant planting plan, n is the number of grids, is the suitability of planting the kth plant in the jth grid, is the planting ratio of the kth plant;

[0086] Iteratively adjust the grid plant species and calculate the comprehensive suitability after each adjustment Until the comprehensive suitability Stop after reaching the maximum value, and output the comprehensive suitability The grid plants are planted and labeled;

[0087] The planting grids of each plant are extracted to form a plant grid map. Each grid in the plant grid map is used as a node, and the shortest path optimization algorithm is used to plan the planting path of each plant.

[0088] By using drones equipped with LiDAR sensors and multispectral cameras, the degraded grasslands were comprehensively scanned to generate digital elevation models (DEMs), humidity distribution maps, and light intensity distribution maps. This data collection method can quickly and accurately obtain the terrain and environmental characteristics of the grassland, and ensure the timeliness of the data. These distribution maps are converted into raster map formats and superimposed to form multidimensional environmental characteristic maps, so that each raster contains elevation, humidity, and light information, providing a complete and accurate data basis for subsequent environmental characteristic analysis and planting path optimization. This method greatly improves the efficiency and accuracy of grassland environmental feature extraction and is very useful in multiple After the 3D environmental feature map is constructed, principal component analysis (PCA) can effectively reduce the data dimension, thereby extracting the most important environmental feature vectors for plant growth. This not only reduces the complexity of subsequent model calculations, but also retains the most informative features, ensuring the scientific nature of the plant planting plan. The principal component feature vectors extracted by PCA can focus on those environmental factors that have the greatest impact on the suitability of grassland plants, making the model more efficient and accurate in subsequent calculations and optimizations. The principal component feature vectors are input into the random forest model to evaluate the contribution of each principal component feature to grassland plant growth. This process quantifies the contribution of each principal component feature to grassland plant growth. The influence of each feature is used to generate the principal component feature contribution matrix. This matrix can accurately show the influence of each environmental feature on plant suitability, thereby providing a scientific basis for the subsequent optimization of the SVM model. This feature contribution quantification method not only improves the accuracy of the model, but also ensures that the plant planting plan is highly matched with the grassland environment. The principal component feature contribution matrix is ​​used to optimize the SVM model, so that the model can more accurately predict the plant suitability of each grid. This optimization process ensures that the model can accurately calculate the plant suitability value of each grid when the principal component feature vector is input. This method improves the accuracy of plant species. The adaptability and accuracy of the planting scheme in the environment provide a scientific basis for the subsequent calculation of plant planting ratio and grid distribution. After obtaining the optimized optimal plant grid map, each grid is processed as a node, and the shortest path optimization algorithm is used to plan the planting path of each plant. This method can ensure that in the actual planting process, drones and seeding equipment can complete the seeding task with the shortest path and lowest energy consumption, thereby improving seeding efficiency. This path optimization algorithm enables each plant to be planted in the most suitable grid, which not only achieves precise planting, but also reduces unnecessary resource consumption, providing an efficient solution for grassland restoration.

[0089] S4, displaying the plant planting plan and planting path and storing them in a database;

[0090] Specifically, displaying the plant planting plan and the planting path means associating the plant planting plan and the planting path and displaying them visually, and inputting the plant planting plan and the planting path into the convolutional neural network to obtain a completed plan view of plant planting in the degraded grassland area for synchronous display.

[0091] Before displaying the plant planting plan and planting path, the two core elements are first associated. The plant planting plan defines the planting area and proportion of a specific plant species, while the planting path determines the actual operation path of the drone or seeding equipment on the grassland. Associating the two can not only ensure that the plants are planted in the most suitable area, but also accurately match the plant species and proportions in each area. This process ensures the accuracy of planting, reduces resource waste, and improves the efficiency of grassland restoration. Through this association, it is also possible to predict and monitor future plant growth and optimize subsequent management and maintenance strategies. The plant planting plan and planting path are visualized so that users can intuitively see the distribution of plant planting areas and the planning of planting paths. This visual display not only improves the transparency of information in the grassland restoration process, but also facilitates managers to make real-time adjustments and optimizations based on the display results. For example, if the planting path is found to be unreasonable in a certain area, the manager can adjust the path planning in time to ensure the coverage and uniformity of plant planting. In addition, through the visual display, a detailed planting report can be generated to record the completion of each planting task for subsequent The data support for grassland monitoring and maintenance is provided. By inputting the plant planting plan and planting path data into the convolutional neural network (CNN), the plant planting completion plan of the degraded grassland can be efficiently generated. This process uses the powerful image recognition and generation capabilities of CNN to convert the original data into a high-precision two-dimensional plan. Specifically, CNN extracts the spatial features of plant planting and path data through the convolution layer, and then reduces the computational complexity through the pooling layer, and finally generates a plan reflecting the actual planting status of the grassland. This method ensures the clarity and accuracy of the displayed image, which helps users to understand the specific situation of grassland restoration more intuitively. The plant planting completion plan generated by CNN can be displayed synchronously, which means that during the planting process, users can monitor the planting progress of each stage in real time. This synchronous display mechanism ensures that users can obtain the latest status of planting in a timely manner and make adjustments and optimizations according to actual conditions. For example, if it is found that the planting in some areas is inconsistent with the plan, users can immediately re-plan the path or adjust the plant planting ratio. This real-time feedback mechanism effectively improves the accuracy of the planting task and the overall restoration effect, and ensures the stability and sustainability of the grassland ecosystem during the restoration process.

[0092] Furthermore, storing in a database means storing the plant planting plan and the planting path of each plant in the database after obtaining them, and storing them in categories according to timestamps. The database regularly performs integrity checks on the stored data and synchronously uploads the stored data to the cloud. Example 2

[0093] Reference Figure 2 , which is the second embodiment of the present invention, is different from the previous embodiment and provides an ecological monitoring system for restoration of degraded grassland vegetation, which includes:

[0094] A data acquisition module is used to collect environmental data of degraded grassland areas through a sensor network and extract environmental data features to calculate feature contributions;

[0095] A scheme optimization module is used to form a plant species list according to the contribution of environmental characteristics, and calculate the fitness of the plant species list to optimize the selection of plants and planting ratios to form a plant planting scheme;

[0096] Path planning module, used to match plant species in degraded grassland areas and plan plant planting paths;

[0097] The display and storage module is used to display and store plant planting plans and plant planting paths, and obtain the completed plant planting plan through neural algorithms for simultaneous display.

[0098] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0099] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0100] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0101] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit with a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit with a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0102] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An ecological monitoring method for restoration of degraded grassland vegetation, characterized in that: include, Collect environmental data of degraded grassland, extract environmental characteristics and calculate the contribution of environmental characteristics; Establish a plant species list based on the contribution of environmental characteristics, and optimize the selection of plant species and planting ratios from the plant species list to form a plant planting plan; Plan planting paths based on plant planting plans for precise planting; Display plant planting plans and planting paths and store them in the database; The method of optimizing and selecting plant species and planting ratios from the plant species list to form a plant planting plan refers to generating initial populations according to different plant species and planting ratios according to the plant species list after obtaining the plant species list, and calculating the plant diversity of each initial population using the Shannon diversity index according to the planting ratio in the initial population: Where M is the plant diversity of the initial population, m is the number of plant species in the initial population, and p i is the planting ratio of the i-th plant; Calculate the plant ecological suitability of each initial population based on the plant planting ratio: Where S is the ecological suitability of the initial population, s i is the fitness of the i-th plant; The fitness function F is constructed based on the plant diversity and plant ecological suitability of the initial population: F = ω1*M+ω2*S; Calculate the fitness of each initial population, and use the roulette wheel selection method to select the individual with the highest fitness for crossover and mutation operations, and generate a new generation of populations. Iterate to generate new populations and calculate the fitness of each new population. When the fitness improvement is lower than the preset threshold, stop the iteration, take the individual with the highest fitness in the new population as the optimal population, and extract the plant species and planting ratio in the optimal population as the plant planting plan; The method of planning the planting path based on the plant planting plan for precise planting refers to using a drone equipped with a LiDAR sensor and a multispectral camera to scan the degraded grassland area to generate a digital elevation model of the degraded grassland, and calculating the humidity through the normalized difference vegetation index to form a humidity distribution map, extracting the light intensity through the multispectral image to form a light distribution map, and converting the digital elevation model of the degraded grassland, the humidity distribution map and the light distribution map into raster map formats respectively and superimposing them to form a multidimensional environmental feature map; The principal component analysis of the elevation, humidity and light contained in each grid in the multidimensional environmental characteristic map is performed to extract the principal component feature vector, and the principal component feature vector is input into the random forest model to obtain the principal component feature contribution and form a principal component feature contribution matrix. The SVM model is optimized based on the principal component feature contribution matrix, and the principal component feature vector of each grid is input into the optimized SVM model to determine the plant suitability of each grid; The total number of grids for each plant is calculated based on the plant planting ratio and grid number in the plant planting plan, and the optimization goal is defined as maximizing the comprehensive suitability S': Where K is the total number of plant species in the plant planting scheme, n is the number of grids, and s j，k is the suitability of planting the kth plant in the jth grid, p k is the planting ratio of the kth plant; Iteratively adjust the grid plant species, and calculate the comprehensive suitability S' after each adjustment until the comprehensive suitability S' reaches the maximum, then stop, output the grid plant species that makes the comprehensive suitability S' and mark it; The planting grids of each plant are extracted to form a plant grid map. Each grid in the plant grid map is used as a node, and the shortest path optimization algorithm is used to plan the planting path of each plant.

2. The ecological monitoring method for restoration of degraded grassland vegetation according to claim 1, characterized in that: The collecting of degraded grassland environmental data, extracting environmental features and calculating environmental feature contribution refers to deploying a sensor network in the degraded grassland area to collect environmental data, cleaning the environmental data collected by the sensor network, smoothing the data using a sliding average method and then standardizing the data, calculating the covariance matrix of the standardized environmental data, solving the eigenvalues ​​and eigenvectors of the covariance matrix, setting a threshold, and taking the eigenvector whose cumulative explained variance is greater than the threshold as the environmental eigenvector; Construct the initial decision tree, set the number of decision trees to 100, and the maximum tree depth to 10. Obtain the historical environmental feature vector as the training set to train the initial decision tree to build a random forest model. During the training process, record the decision tree splitting process in real time, calculate the variance of the split child decision tree corresponding to the environmental feature vector, and calculate the contribution value of the variance reduction: Where R f is the contribution of the environmental feature vector f to the reduction of model variance when the decision tree node is split, Var(D) is the variance of the decision tree before splitting, |D| is the number of environmental feature vectors in the decision tree before splitting, |D l | and |D r | are the number of environmental feature vectors in the left and right child decision trees of the decision tree split, Var(D l ) and Var(D r ) is the variance of the left and right child decision trees; When the environmental feature vector f is used for decision tree splitting, its contribution to the reduction of model variance is recorded, and the total contribution value R of the environmental feature vector f to the reduction of model variance during training is obtained comprehensively. zf ; Get the total contribution value of each environmental feature vector and perform normalization to get the feature contribution C of each environmental feature vector f .

3. The ecological monitoring method for restoration of degraded grassland vegetation according to claim 2, characterized in that: The establishing of the plant species list according to the environmental feature contribution means obtaining the feature contribution of each environmental feature vector and then constructing a feature contribution matrix C; Select the radial basis function kernel to build the SVM model and optimize it based on the feature contribution matrix C: K(x,x’)=exp(-γ‖C(x-x’)‖ 2 ); Where K(x, x') is the optimized radial basis function kernel function, γ is the kernel function parameter, x and x' are the input environment feature vector and the model training environment feature vector respectively; The kernel function parameters and multiplication parameters are adjusted through cross-validation, and the loss function is defined, and the optimization goal is set to minimize the loss function; Obtain plant planting environment data and extract plant environment feature vectors as training data to input into the SVM model for iterative optimization of model parameters; The environmental feature vector is input into the trained SVM model, and the SVM model outputs the suitability of each plant, sets a suitability threshold, and sorts them from high to low according to suitability, and forms a plant species list with plants above the suitability threshold.

4. The ecological monitoring method for restoration of degraded grassland vegetation as claimed in claim 3, characterized in that: The display of the plant planting plan and the planting path refers to associating the plant planting plan and the planting path and visually displaying them, and inputting the plant planting plan and the planting path into the convolutional neural network to obtain the completed plan view of the plant planting in the degraded grassland area for synchronous display.

5. The ecological monitoring method for restoration of degraded grassland vegetation according to claim 4, characterized in that: The storing in the database means storing the plant planting plan and the planting path of each plant in the database after obtaining them, and storing them in categories according to timestamps. The database regularly performs integrity checks on the stored data and synchronously uploads the stored data to the cloud.

6. An ecological monitoring system for degraded grassland vegetation restoration according to the ecological monitoring method for degraded grassland vegetation restoration as claimed in any one of claims 1 to 5, characterized in that: include, A data acquisition module is used to collect environmental data of degraded grassland areas through a sensor network and extract environmental data features to calculate feature contributions; A scheme optimization module is used to form a plant species list according to the contribution of environmental characteristics, and calculate the fitness of the plant species list to optimize the selection of plants and planting ratios to form a plant planting scheme; Path planning module, used to match plant species in degraded grassland areas and plan plant planting paths; The display and storage module is used to display and store the plant planting plan and plant planting path, and obtain the plant planting completion plan through the neural algorithm for synchronous display; The method of optimizing and selecting plant species and planting ratios from the plant species list to form a plant planting plan refers to generating initial populations according to different plant species and planting ratios according to the plant species list after obtaining the plant species list, and calculating the plant diversity of each initial population using the Shannon diversity index according to the planting ratio in the initial population: Where M is the plant diversity of the initial population, m is the number of plant species in the initial population, and p i is the planting ratio of the i-th plant; Calculate the plant ecological suitability of each initial population based on the plant planting ratio: Where S is the ecological suitability of the initial population, s i is the fitness of the i-th plant; The fitness function F is constructed based on the plant diversity and plant ecological suitability of the initial population: F = ω1*M+ω2*S; Calculate the fitness of each initial population, and use the roulette wheel selection method to select the individual with the highest fitness for crossover and mutation operations, and generate a new generation of populations. Iterate to generate new populations and calculate the fitness of each new population. When the fitness improvement is lower than the preset threshold, stop the iteration, take the individual with the highest fitness in the new population as the optimal population, and extract the plant species and planting ratio in the optimal population as the plant planting plan; The method of planning the planting path based on the plant planting plan for precise planting refers to using a drone equipped with a LiDAR sensor and a multispectral camera to scan the degraded grassland area to generate a digital elevation model of the degraded grassland, and calculating the humidity through the normalized difference vegetation index to form a humidity distribution map, extracting the light intensity through the multispectral image to form a light distribution map, and converting the digital elevation model of the degraded grassland, the humidity distribution map and the light distribution map into raster map formats respectively and superimposing them to form a multidimensional environmental feature map; The principal component analysis of the elevation, humidity and light contained in each grid in the multidimensional environmental characteristic map is performed to extract the principal component feature vector, and the principal component feature vector is input into the random forest model to obtain the principal component feature contribution and form a principal component feature contribution matrix. The SVM model is optimized based on the principal component feature contribution matrix, and the principal component feature vector of each grid is input into the optimized SVM model to determine the plant suitability of each grid; The total number of grids for each plant is calculated based on the plant planting ratio and grid number in the plant planting plan, and the optimization goal is defined as maximizing the comprehensive suitability S': Where K is the total number of plant species in the plant planting scheme, n is the number of grids, and s j，k is the suitability of planting the kth plant in the jth grid, p k is the planting ratio of the kth plant; Iteratively adjust the grid plant species, and calculate the comprehensive suitability S' after each adjustment until the comprehensive suitability S' reaches the maximum, then stop, output the grid plant species that makes the comprehensive suitability S' and mark it; The planting grids of each plant are extracted to form a plant grid map. Each grid in the plant grid map is used as a node, and the shortest path optimization algorithm is used to plan the planting path of each plant.

7. A computer device comprising: Memory and processor; The memory stores a computer program, characterized in that the processor implements the steps of the ecological monitoring method for restoration of degraded grassland vegetation according to any one of claims 1 to 5 when executing the computer program.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the ecological monitoring method for restoration of degraded grassland vegetation as described in any one of claims 1 to 5 are implemented.

Images