Agricultural environment monitoring method and system based on 5G Internet of Things

Abstract

The invention belongs to the technical field of intelligent agriculture, and provides an agricultural environment monitoring method and system based on the 5G Internet of Things, and the method comprises the steps: extracting growth feature data of crops in an accept region in a plurality of historical periods from historical crop growth records, building a growth stage recognition model based on a machine learning model, and carrying out the recognition of the growth stage; identifying growth stages of all crops in the acclimatization area, and judging whether all crops in the acclimatization area are in the same growth stage or not; compared with the standard parameter range of the crop growth stage, the method effectively avoids the resource waste of the traditional uniform environment regulation and control, remarkably improves the targeting and efficiency of the environment regulation and control, provides datamation decision support for the growth of the crops in the optimal environment, and assists the fine management and yield improvement of agriculture.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of intelligent agriculture, and in particular relates to an agricultural environment monitoring method and system based on 5G Internet of Things. BACKGROUND

[0002] In the process of agricultural modernization, precision agriculture and environmental monitoring have become key technologies for improving crop yield, optimizing resource allocation and achieving sustainable development. Traditional agricultural management mode mainly relies on manual inspection and experience judgment, which has problems such as low data collection efficiency, poor real-time performance, and insufficient spatial resolution, and cannot meet the needs of fine planting.

[0003] With the development of Internet of Things, 5G communication, artificial intelligence and remote sensing technology, agricultural environment monitoring has gradually evolved towards intelligence and automation, but the existing technology still has the following significant defects: the lag and one-sidedness of growth stage monitoring, the existing crop growth monitoring mainly relies on a single data source (such as spectral reflectance or morphological features), and the model training relies on limited historical data, resulting in low accuracy of growth stage identification, traditional methods lack dynamic judgment ability for multi-growth stage coexistence scenarios, and cannot distinguish crops in different growth stages in the same area in real time, resulting in lack of pertinence of management measures.

[0004] The lack of spatial continuity analysis, the existing agricultural zoning technology does not fully consider the continuity of the actual spatial distribution of crops, resulting in fragmentation of management units and affecting the efficiency of environmental regulation.

[0005] Passivity and singleness of environmental regulation, the existing environmental monitoring system mainly focuses on single parameter threshold warning (such as triggering irrigation when soil humidity is below a certain value), lacks analysis of multi-parameter coupling effect, and the coordinated change of soil humidity and temperature may cause crop stress, but traditional methods cannot quantify the interference relationship between parameters, resulting in lag or invalidity of regulation measures, and the priority judgment of abnormal parameters relies on artificial experience, lacking a dynamic sorting mechanism driven by data.

[0006] Therefore, the present application provides an agricultural environment monitoring method and system based on 5G Internet of Things. SUMMARY

[0007] In order to make up for the deficiencies of the existing technology and solve at least one technical problem proposed in the background art.

[0008] In a first aspect, the present application provides an agricultural environment monitoring method based on 5G Internet of Things, comprising: extracting growth feature data of crops in the adoption area in multiple historical periods from historical crop growth records, constructing a growth stage identification model based on a machine learning model, identifying the growth stage of all crops in the adoption area, and determining whether all crops in the adoption area are in the same growth stage; If crops are not at the same growth stage, perform cluster analysis on crops at the same growth stage, introduce spatial constraints during the clustering process to ensure that crops within the same cluster are spatially continuous, output spatially continuous regional clusters with crops at the same growth stage, and determine the geographical boundaries of the regional clusters. Based on any spatially continuous regional cluster with the same crop growth stage, environmental parameters of the current crop growth stage are extracted from the regional cluster. By performing deviation analysis on the environmental parameters of the current crop growth stage, the growth environment of the crop at the current growth stage within the geographical boundary of the regional cluster is evaluated to determine whether the growth environment is suitable. If the growth environment is unsuitable, interference analysis is performed on the growth environment of crops at their current growth stage within the geographical boundaries of the regional clusters. Based on the analysis results, a control sequence table for abnormal environmental parameters is determined to make the growth environment more suitable for the current growth stage of the crops.

[0009] Preferably, the specific process for determining whether all crops in the adopted area are at the same growth stage is as follows: Based on the growth stage identification model, the growth stages of all crops in the adopted area are identified, and the percentage of each growth stage of all crops in the adopted area is calculated. If the percentage of crops in each growth stage is greater than or equal to the threshold, it means that all crops in the adopted area are not in the same growth stage; otherwise, it means that all crops in the adopted area are in the same growth stage.

[0010] Preferably, the specific process for identifying all crop growth stages in the adopted area is as follows: A pre-set monitoring period is established. During the monitoring period, drones are deployed to collect RGB images using high-resolution multispectral cameras mounted on the drones, while simultaneously acquiring the spectral reflectance of all crops in the adopted area; high spatial resolution morphological features are provided through the RGB images. Set growth stage labels for all crops within the adopted area, and use the spectral reflectance, morphological characteristics, and corresponding growth stage labels of the crops as growth characteristics. Input the growth characteristics of all crops in the adopted area, which are monitored in real time by 5G IoT, into the trained growth stage recognition model, and output the growth stage of all crops in the adopted area.

[0011] Preferably, the process for determining the geographical boundaries of the regional clusters is as follows: The spatial coordinates of all crops are recorded using RTK-GPS positioning. A spatial distance threshold is set, and an adjacency matrix is ​​constructed. The crop growth stage labels and the corresponding spatial coordinates of the crops are integrated into the feature vector of the crops. The K-Means algorithm is used for clustering to determine the number of clusters K. For each crop, the basic distance from the crop to the cluster center is calculated using the Euclidean distance formula. An adjacency matrix is ​​introduced, and the spatial constraints from the crop to the cluster center are calculated using the formula. summing the crop-to-cluster center base distance and the crop-to-cluster center spatial constraint, obtaining the total distance of the crop to the cluster center; assigning each crop to the cluster with the minimum total distance, updating the cluster center as the feature mean of the crops in the cluster; continuously repeating the steps of assigning crops to the cluster and updating the cluster center until the cluster center no longer changes or the preset number of iterations is reached, outputting the region cluster with continuous space and the same growth stage of the crops, and taking the maximum spatial boundary of the region cluster as the geographical boundary of the region cluster.

[0012] Preferably, the construction process of the adjacency matrix is: extracting two crops from all crops in the fostered region, calculating the distance between the two crops by the Euclidean distance formula; if the distance between the two crops is less than or equal to the spatial distance threshold, it indicates that the two crops are adjacent; otherwise, it indicates that the two crops are not adjacent; According to the adjacency relationship between all crops, all crops in the fostered region are numbered, and an adjacency matrix is constructed. If the crops are adjacent, the adjacency matrix is recorded as 1, and if the crops are not adjacent, the adjacency matrix is recorded as 0. The row and column of the adjacency matrix represent the crop number, and the adjacency relationship of the same crop in the same row and column is recorded as 0 in the adjacency matrix.

[0013] Preferably, the specific process of evaluating whether the growth environment of the crops in the current growth stage within the geographical boundary of the region cluster is qualified is: dividing the monitoring period into several time intervals, extracting the environmental parameters collected at the monitoring time points, and determining the abnormal environmental parameters; calculating the proportion of the number of abnormal environmental parameters in all environmental parameters; if it is greater than or equal to the abnormal environmental parameter proportion threshold, it indicates that the growth environment of the crops in the current growth stage within the geographical boundary of the region cluster is unqualified; otherwise, it indicates that the growth environment of the crops in the current growth stage within the geographical boundary of the region cluster is qualified.

[0014] Preferably, the determination process of the abnormal environmental parameter is: If the environmental parameter is not within the environmental parameter range, the corresponding environmental parameter is recorded as an abnormal environmental parameter.

[0015] Preferably, the process of performing interference analysis on the growth environment of the crops in the current growth stage within the geographical boundary of the region cluster is: the geographical boundary of the region cluster with unqualified growth environment of the crops in the current growth stage is recorded as an environment unqualified region; The number of abnormal environmental parameters of the environmental unqualified area in the monitoring period is counted, if the number of abnormal environmental parameters of the environmental unqualified area in the monitoring period is greater than 1, an analysis signal is generated, otherwise, the corresponding abnormal environmental parameter is recorded as an interference parameter, and the interference parameter is directly regulated.

[0016] Preferably, the specific process of determining the regulation sequence table of the abnormal environmental parameter is as follows: Based on the analysis signal, the number of abnormal environmental parameters of the environmental unqualified area in the monitoring period is integrated into a sequence of abnormal environmental parameter numbers, and any abnormal environmental parameter category is extracted as an independent variable, and other abnormal environmental parameter categories in the sequence of abnormal environmental parameters are used as explanatory variables to establish a linear regression model. If the variance inflation factor between the current abnormal environmental parameter category and the rth abnormal environmental parameter category is greater than 1, the rth abnormal environmental parameter category is recorded as a collinear parameter category. All common parameter categories of the current abnormal environmental parameter category are removed from the sequence of abnormal environmental parameter numbers, and the removed sequence of abnormal environmental parameter numbers is recorded as a sequence of abnormal environmental parameter numbers to be analyzed. In the monitoring period, the abnormal environmental parameters corresponding to the number of abnormal environmental parameters in the sequence of abnormal environmental parameters to be analyzed are extracted, and the Spearman correlation coefficient formula is used to calculate the interference representation value between the abnormal environmental parameters corresponding to the number of abnormal environmental parameters in the sequence of abnormal environmental parameters to be analyzed and the growth characteristics corresponding to the growth stage of crops, and the interference representation value is sorted from large to small to determine the regulation sequence table of the abnormal environmental parameter.

[0017] In a second aspect, the present application also provides an agricultural environment monitoring system based on 5G Internet of Things, which comprises: A growth stage identification module extracts growth characteristic data of crops in the fostered area in multiple historical periods from historical crop growth records, and establishes a growth stage identification model based on a machine learning model to identify the growth stages of all crops in the fostered area and determine whether all crops in the fostered area are in the same growth stage. A cluster analysis module performs cluster analysis on crops in the same growth stage, introduces spatial constraints in the cluster process to ensure that crops in the same cluster are spatially continuous, outputs a region cluster with spatial continuity and the same growth stage of crops, and determines the geographical boundary of the region cluster. The environment qualification judgment module: based on any spatially continuous and crop growth stage same region clustering cluster, extracting the environment parameters of the current growth stage of crops in the spatially continuous and crop growth stage same region clustering cluster, through deviation analysis on the environment parameters of the current growth stage of crops, evaluating whether the growth environment of the current growth stage of crops within the geographical boundary of the region clustering cluster is qualified; The interference analysis module: if the growth environment is not qualified, through interference analysis on the growth environment of the current growth stage of crops within the geographical boundary of the region clustering cluster, determining the regulation sequence table of abnormal environment parameters according to the analysis result, so that the growth environment is more suitable for the current growth stage of crops.

[0018] The beneficial effects of the present application are as follows: The present application realizes accurate monitoring and regional division of the growth state of the adopted region crops by combining machine learning and spatial clustering technology: the growth stage recognition model is constructed and trained by using the spectral reflectivity, morphological characteristics and growth stage label data of crops in the historical period, the multispectral and RGB image data are collected by the unmanned aerial vehicle in real time, and the current growth stage of each crop is recognized by inputting the preprocessed data into the model, the number and proportion of the stages are counted to determine whether they are synchronized; if not, the spatial coordinates of RTK-GPS positioning and the growth stage label are combined, and the improved K-Means algorithm is introduced for clustering with spatial constraints, so that the crops within the same cluster are spatially continuous and have consistent growth stages, and finally the region clustering cluster with clear geographical boundary is output, providing basis for fine agricultural management.

[0019] The present application accurately identifies the unqualified environment region by comparing with the standard parameter range of the growth stage of crops; when the environment is abnormal, the method combining statistical analysis and machine learning is adopted, which not only eliminates the collinearity interference among multiple parameters, but also quantifies the interference intensity of each abnormal parameter on crop growth, finally generates a regulation priority table according to the influence degree, effectively avoids the resource waste of traditional uniform environment regulation, significantly improves the targeting and efficiency of environment regulation, provides data-based decision support for crop growth in the best environment, and helps fine agricultural management and yield improvement. BRIEF DESCRIPTION OF DRAWINGS

[0020] The present application will be further described below in combination with the drawings.

[0021] Figure 1 is a step flow chart of an agricultural environment monitoring method based on 5G Internet of Things according to an embodiment of the present application; Figure 2 is a system block diagram of an agricultural environment monitoring system based on 5G Internet of Things according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to make the technical means, creative features, purposes and effects of the present application easy to understand, the present application will be further described below in conjunction with specific embodiments. Embodiment 1

[0023] Please refer to Figure 1 The agricultural environment monitoring method based on 5G Internet of Things provided by the embodiment of the present application comprises the following steps: Step 1: Extract the growth characteristic data of the crops in the adoption area in multiple historical periods from the historical crop growth records, construct a growth stage identification model based on a machine learning model, identify the growth stages of all the crops in the adoption area, and determine whether all the crops in the adoption area are in the same growth stage; Among them, the growth characteristic data of the crops in the adoption area in multiple historical periods includes spectral reflectance, morphological characteristics and corresponding growth stage labels, and the morphological characteristics include plant contour and leaf texture. It should be noted that the historical period includes but is not limited to 3 months, 6 months and 1 year. Based on the machine learning model, the growth stage identification model is constructed, the growth characteristic data of the crops in the adoption area in multiple historical periods is divided into a training set, a validation set and a test set according to a ratio of 8:1:1, the training set is used for model parameter training to ensure that the model learns enough feature patterns, the validation set is used for adjusting hyperparameters and evaluating the performance of the model on unseen data, and the test set is finally used for evaluating the generalization ability of the model and simulating the prediction effect in a real scene. The K-fold cross-validation is used to evaluate the generalization ability of the model to avoid overfitting, and when the amount of data is limited (the annotation cost is high in the agricultural scene), the cross-validation can fully utilize the growth characteristic data of the crops in the adoption area in multiple historical periods to evaluate the model. A preset monitoring period is set, and in the monitoring period, a UAV is deployed, a high-resolution multispectral camera carried by the UAV is used to collect RGB images, and the spectral reflectance of all the crops in the adoption area is obtained at the same time. All the collected data is corrected by atmospheric correction and radiation calibration to eliminate the interference of light, and the geometric correction is used to ensure the spatial alignment of the images. The RGB images provide high spatial resolution morphological characteristics. The growth stage labels of all the crops in the adoption area are set, the spectral reflectance, morphological characteristics and corresponding growth stage labels of the crops are used as the growth characteristics of the crops, and the growth characteristics of all the crops in the adoption area monitored in real time through the 5G Internet of Things are input into the trained growth stage identification model to output the growth stages of all the crops in the adoption area. The number of growth stages of all the crops in the adoption area is counted, and the proportion of the number of growth stages of all the crops in the adoption area is calculated. In some embodiments, the percentage of species in each growth stage is compared with a threshold for the percentage of species in each growth stage: If the percentage of crops in each growth stage is greater than or equal to the threshold for the percentage of crops in each growth stage, it means that all crops in the adopted area are not in the same growth stage. If the percentage of crops in each growth stage is less than the threshold for the percentage of crops in each growth stage, it means that all crops in the adopted area are in the same growth stage. Step 2: If crops are not in the same growth stage, perform cluster analysis on crops in the same growth stage. Introduce spatial constraints during the clustering process to ensure that crops in the same cluster are spatially continuous. Output spatially continuous regional clusters with crops in the same growth stage and determine the geographical boundaries of the regional clusters. The spatial coordinates of all crops are recorded using RTK-GPS positioning. A spatial distance threshold is set. The spatial distance threshold is set by those skilled in the art based on crop characteristics and historical experience. Two crops are randomly selected from all crops in the adoption area, and the distance between the two crops is calculated using the Euclidean distance formula. Compare the distance between two crops with a spatial distance threshold: If the distance between two crops is less than or equal to the spatial distance threshold, then the two crops are considered to be adjacent. If the distance between two crops is greater than the spatial distance threshold, it means that the two crops are not adjacent. Based on the proximity relationships between all crops, all crops in the adoption area are numbered and an adjacency matrix is ​​constructed. If crops are adjacent, they are recorded as 1 in the adjacency matrix; if crops are not adjacent, they are recorded as 0 in the adjacency matrix. The rows and columns in the adjacency matrix represent crop numbers. The proximity relationship of the same crop with the same row and column is recorded as 0 in the adjacency matrix. The crop growth stage labels and the corresponding spatial coordinates of the crop are integrated into the crop feature vector, and the K-Means algorithm is used for clustering to determine the number of clusters K. Specifically, the number of clusters K can be determined using the elbow method, and the sum of squared errors SSE can be calculated for different K values; Define the candidate range for K. For each K, use the K-Means clustering algorithm to cluster the data and calculate the corresponding SSE. Plot the K-SSE curve with K as the x-axis and SSE as the y-axis, calculate the rate of change of the decrease in SSE corresponding to adjacent K values, and select the K value whose rate of change flattens out after a sharp drop as the optimal number of clusters. K crops were randomly selected as the initial cluster centers; For each crop, the basic distance from the crop to the cluster center is calculated using the Euclidean distance formula. An adjacency matrix is ​​then introduced, and the following correction formula is applied: Calculate the spatial constraints from crop i to cluster center j In the formula, This represents the feature vector of crop i. The feature vector representing the cluster center j, The spatial weighting coefficients are set by those skilled in the art based on historical experience. This represents the elements corresponding to crop i and crop m in the adjacency matrix. The value represents the absolute difference between the k-th feature of crop i and the k-th feature of cluster center j, where n represents the feature dimension and M represents the total number of crops in the adoption area. The total distance from crops to cluster centers is obtained by summing the basic distance from crops to cluster centers and the spatial constraints from crops to cluster centers. Each crop is assigned to the cluster with the smallest total distance, and the cluster center is updated to the feature mean of the crops in the cluster. Repeat the above steps of assigning crops to clusters and updating cluster centers until the cluster centers no longer change or the preset number of iterations is reached. Output spatially continuous regional clusters with the same crop growth stage, and use the maximum spatial boundary of the regional clusters as the geographical boundary of the regional clusters. Determining the geographical boundaries of regional clusters serves the following purposes: Function 1: In the adoption model, the geographical boundaries clearly define the area of ​​the plot adopted by the consumer, and the consumer can view the crop growth progress through the map, which enhances the sense of trust and participation; Function 2: Geographic boundaries serve as the basis for spatial analysis, integrating multi-source data such as remote sensing and the Internet of Things (IoT) to improve decision-making accuracy. By combining satellite imagery and soil sensor data, cluster boundaries can be dynamically adjusted to adapt to environmental changes, which is beneficial for implementing differentiated environmental control strategies for different regions. Thirdly, the introduction of spatial constraints in the process of determining geographical boundaries optimizes the traditional K-Means clustering algorithm, making it more suitable for the spatial continuity requirements of agricultural scenarios. By automatically dividing management units through geographical boundaries, the cost of manual inspection is reduced. The technical solution of this embodiment is as follows: Growth characteristic data of crops in the adopted area over multiple historical periods are extracted from historical crop growth records. Based on a machine learning model, a growth stage identification model is constructed to identify the growth stages of all crops in the adopted area. It is then determined whether all crops in the adopted area are in the same growth stage. If they are not in the same growth stage, crops in the same growth stage are clustered. Spatial constraints are introduced during the clustering process to ensure spatial continuity of crops within the same cluster. The resulting clusters are spatially continuous and contain crops in the same growth stage. The geographical boundaries of these clusters are then determined. This invention, by combining machine learning and spatial clustering techniques, achieves the identification of crop growth in the adopted area. Precise monitoring and regional division of crop status: By utilizing the spectral reflectance, morphological characteristics, and growth stage label data of crops within historical cycles, a growth stage identification model is constructed and trained. Multispectral and RGB image data are collected in real time by drones, and after preprocessing, the data is input into the model to identify the current growth stage of each crop. The number and proportion of different stages are counted to determine whether they are synchronized. If they are not synchronized, the spatial coordinates of RTK-GPS positioning and growth stage labels are combined, and an improved K-Means algorithm with spatial constraints is introduced to cluster crops, ensuring that crops within the same cluster are spatially continuous and have consistent growth stages. The final output is a regional cluster with clear geographical boundaries, providing a basis for precision agricultural management. Example 2

[0024] Please see Figure 1 As shown in the figure, an agricultural environment monitoring method based on 5G Internet of Things according to an embodiment of the present invention further includes the following steps: Step 3: Based on any spatially continuous regional cluster with the same crop growth stage, extract the environmental parameters of the current crop growth stage in the regional cluster. By performing deviation analysis on the environmental parameters of the current crop growth stage, evaluate whether the growth environment of the current crop growth stage within the geographical boundary of the regional cluster is qualified. Among them, environmental parameters include, but are not limited to, soil moisture and soil temperature. IoT sensors are embedded within the geographical boundaries of regional clusters. Environmental parameters are collected through IoT sensors and transmitted to the cloud platform via 5G network. The environmental parameters are time-aligned with the growth characteristic data corresponding to the crop growth stage to ensure that the timestamps of the environmental parameters and the growth characteristic data corresponding to the crop growth stage are consistent. It should be noted that soil moisture values ​​are collected using a humidity sensor and soil temperature values ​​are collected using a temperature sensor. The monitoring period is divided into several monitoring time points with equal time intervals. Environmental parameters collected at each monitoring time point are extracted, and the environmental parameters are compared with the environmental parameter range. If an environmental parameter is outside the range of environmental parameters, the corresponding environmental parameter will be recorded as an abnormal environmental parameter. If the environmental parameters are within the environmental parameter range, then the corresponding environmental parameters are recorded as normal environmental parameters. It is understandable that the range of environmental parameters is set by those skilled in the art based on the growth characteristics of crops at different stages of production and in combination with historical experience. The number of abnormal environmental parameters among all environmental parameters within the statistical monitoring period is counted, and the percentage of abnormal environmental parameters among all environmental parameters is calculated. Compare the percentage of abnormal environmental parameters among all environmental parameters within the monitoring period with the threshold for the percentage of abnormal environmental parameters: If the proportion of abnormal environmental parameters among all environmental parameters in the monitoring period is greater than or equal to the threshold for the proportion of abnormal environmental parameters, it indicates that the growth environment of crops at the current growth stage within the geographical boundary of the regional cluster is unqualified. If the proportion of abnormal environmental parameters among all environmental parameters during the monitoring period is less than the threshold for the proportion of abnormal environmental parameters, it indicates that the growth environment of crops at the current growth stage within the geographical boundary of the regional cluster is qualified. Step 4: If the growth environment is unsuitable, conduct interference analysis on the growth environment of the crop at the current growth stage within the geographical boundary of the regional cluster, and determine the control sequence table of abnormal environmental parameters based on the analysis results to make the growth environment more suitable for the current growth stage of the crop. The geographical boundaries of clusters representing areas where the crop's current growth stage has an unsuitable growing environment are denoted as unsuitable environmental areas. The number of abnormal environmental parameters in the non-compliant area during the monitoring period is counted. If the number of abnormal environmental parameters in the non-compliant area during the monitoring period is greater than 1, an analysis signal is generated. If the number of abnormal environmental parameters in an area that does not meet environmental standards during the monitoring period is less than or equal to 1, then the corresponding abnormal environmental parameter is recorded as an interference parameter and the interference parameter is directly adjusted. Based on the analysis signal, the number of abnormal environmental parameters in the non-compliant area during the monitoring period is integrated into an abnormal environmental parameter number sequence. Any abnormal environmental parameter category is extracted from the abnormal environmental parameter number sequence as an independent variable, and other abnormal environmental parameter categories in the abnormal environmental parameter sequence are used as explanatory variables. A linear regression model is established, and the determination coefficient between the current abnormal environmental parameter category and other abnormal environmental parameter categories is calculated using the determination coefficient formula. Using the variance inflation factor formula: The variance inflation factor between the current abnormal environmental parameter category and the r-th abnormal environmental parameter category is calculated. , This represents the determination coefficient between the category of abnormal environmental parameter and the r-th category of abnormal environmental parameter. Compare the variance inflation factor between the current abnormal environmental parameter category and the r-th abnormal environmental parameter category with 1: If the variance inflation factor between the current abnormal environmental parameter category and the r-th abnormal environmental parameter category is greater than 1, it indicates that there is collinearity between the current abnormal environmental parameter category and the r-th abnormal environmental parameter category. The r-th abnormal environmental parameter category is denoted as the collinearity parameter category. If the variance inflation factor between the current abnormal environmental parameter category and the r-th abnormal environmental parameter category is less than or equal to 1, it indicates that there is no collinearity between the current abnormal environmental parameter category and the r-th abnormal environmental parameter category, and the r-th abnormal environmental parameter category is recorded as the non-collinear parameter category. Remove all common parameter categories of the current abnormal environment parameter category from the abnormal environment parameter species sequence, and so on. Record the abnormal environment parameter species sequence after removal as the abnormal environment parameter species sequence to be analyzed. Within the monitoring period, the abnormal environmental parameters corresponding to the number of abnormal environmental parameters in the sequence of abnormal environmental parameters to be analyzed are extracted. Using the Spearman correlation coefficient formula, the interference characterization value between the abnormal environmental parameters corresponding to the number of abnormal environmental parameters in the sequence of abnormal environmental parameters to be analyzed and the growth characteristics corresponding to the crop growth stage is calculated. The interference characterization values ​​between the abnormal environmental parameters corresponding to the number of abnormal environmental parameters in the sequence of abnormal environmental parameters to be analyzed and the growth characteristics corresponding to the crop growth stage are sorted from largest to smallest to determine the control order table of abnormal environmental parameters. Based on the control order table of environmental parameters, the crop growth environment is controlled to make the growth environment more suitable for the current growth stage of the crop. The purpose of determining the environmental parameter control sequence table is: Function 1: Calculate the interference characterization values ​​of abnormal environmental parameters and crop growth characteristics using Spearman correlation coefficient, and sort them from largest to smallest to clarify the degree of influence of each parameter on crop growth; Function 2: By prioritizing the processing of high-interference parameters, resources are avoided from being wasted on low-impact parameters, thus achieving efficient resource allocation; Function 3: By combining collinearity analysis and Spearman's coefficient, interrelated parameters are eliminated, ensuring that regulatory measures target independent and influential factors, thereby improving the accuracy of regulation; Function 4: Provides instruction sequence for automated environmental control systems, enabling on-demand control and adapting to dynamic environmental changes; Fifthly, by promptly processing highly disruptive parameters, the negative impact of environmental anomalies on crop growth can be reduced, thereby lowering production risks. The technical solution of this embodiment is as follows: Based on any spatially continuous regional cluster with the same crop growth stage, environmental parameters of the current crop growth stage are extracted from the regional cluster. By performing deviation analysis on the environmental parameters of the current crop growth stage, the growth environment of the crop within the geographical boundary of the regional cluster is evaluated as suitable. If the growth environment is unsuitable, interference analysis is performed on the growth environment of the crop within the geographical boundary of the regional cluster. Based on the analysis results, a control sequence table of abnormal environmental parameters is determined to make the growth environment more suitable for the current crop growth stage. This invention accurately identifies unsuitable environmental areas by comparing with the standard parameter range of the crop growth stage. When the environment is abnormal, a method combining statistical analysis and machine learning is used to eliminate collinear interference between multiple parameters and quantify the interference intensity of each abnormal parameter on crop growth. Finally, a control priority table is generated according to the degree of influence. This effectively avoids the resource waste of traditional uniform environmental control, significantly improves the targeting and efficiency of environmental control, provides data-driven decision support for crops to grow in the best environment, and helps to improve agricultural precision management and yield. Example 3

[0025] Please see Figure 2 As shown in the figure, an agricultural environment monitoring system based on 5G Internet of Things according to an embodiment of the present invention includes the following modules: Growth stage identification module: Extract growth characteristic data of crops in the adopted area from historical crop growth records over multiple historical periods, construct a growth stage identification model based on machine learning model, identify the growth stage of all crops in the adopted area, and determine whether all crops in the adopted area are in the same growth stage; Clustering analysis module: If crops are not in the same growth stage, cluster analysis will be performed on crops in the same growth stage. Spatial constraints will be introduced during the clustering process to ensure that crops in the same cluster are spatially continuous. The output will be a spatially continuous cluster of regions with crops in the same growth stage, and the geographical boundaries of the regional clusters will be determined. Environmental compliance assessment module: Based on any spatially continuous regional cluster with the same crop growth stage, extract the environmental parameters of the current crop growth stage in the regional cluster. By performing deviation analysis on the environmental parameters of the current crop growth stage, assess whether the growth environment of the current crop growth stage within the geographical boundary of the regional cluster is qualified. Interference Analysis Module: If the growth environment is unsuitable, interference analysis is performed on the growth environment of the crop at its current growth stage within the geographical boundaries of the regional cluster. Based on the analysis results, a control sequence table of abnormal environmental parameters is determined to make the growth environment more suitable for the current growth stage of the crop.

[0026] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. An agricultural environmental monitoring method based on 5G Internet of Things, characterized in that: include: Growth characteristic data of crops in the adopted area over multiple historical periods are extracted from historical crop growth records. Based on machine learning models, a growth stage identification model is constructed to identify the growth stages of all crops in the adopted area and determine whether all crops in the adopted area are in the same growth stage. If crops are not at the same growth stage, perform cluster analysis on crops at the same growth stage, introduce spatial constraints during the clustering process to ensure that crops within the same cluster are spatially continuous, output spatially continuous regional clusters with crops at the same growth stage, and determine the geographical boundaries of the regional clusters. Based on any spatially continuous regional cluster with the same crop growth stage, environmental parameters of the current crop growth stage are extracted from the regional cluster. By performing deviation analysis on the environmental parameters of the current crop growth stage, the growth environment of the crop at the current growth stage within the geographical boundary of the regional cluster is evaluated to determine whether the growth environment is suitable. If the growth environment is unsuitable, interference analysis is performed on the growth environment of crops at their current growth stage within the geographical boundaries of the regional clusters. Based on the analysis results, a control sequence table for abnormal environmental parameters is determined to make the growth environment more suitable for the current growth stage of the crops.

2. The agricultural environment monitoring method based on 5G Internet of Things according to claim 1, characterized in that: The specific process for determining whether all crops in the adopted area are at the same growth stage is as follows: Based on the growth stage identification model, the growth stages of all crops in the adopted area are identified, and the percentage of each growth stage of all crops in the adopted area is calculated. If the percentage of crops in each growth stage is greater than or equal to the threshold, it means that all crops in the adopted area are not in the same growth stage; otherwise, it means that all crops in the adopted area are in the same growth stage.

3. The agricultural environment monitoring method based on 5G Internet of Things according to claim 2, characterized in that: The specific process for identifying all crop growth stages in the adopted area is as follows: A pre-set monitoring period is established. During the monitoring period, drones are deployed to collect RGB images using high-resolution multispectral cameras mounted on the drones, while simultaneously acquiring the spectral reflectance of all crops in the adopted area; high spatial resolution morphological features are provided through the RGB images. Set growth stage labels for all crops within the adopted area, and use the spectral reflectance, morphological characteristics, and corresponding growth stage labels of the crops as growth characteristics. Input the growth characteristics of all crops in the adopted area, which are monitored in real time by 5G IoT, into the trained growth stage recognition model, and output the growth stage of all crops in the adopted area.

4. A method for agricultural environmental monitoring based on 5G Internet of Things according to claim 3, characterized in that: The process for determining the geographical boundaries of the regional clusters is as follows: The spatial coordinates of all crops are recorded using RTK-GPS positioning. A spatial distance threshold is set, and a proximity matrix is ​​constructed. The crop growth stage labels and the corresponding spatial coordinates of the crops are integrated into the feature vector of the crops. The K-Means algorithm is used for clustering to determine the number of clusters K. For each crop, the basic distance from the crop to the cluster center is calculated using the Euclidean distance formula. An adjacency matrix is ​​introduced, and the spatial constraints from the crop to the cluster center are calculated using a modified formula. The total distance from crops to cluster centers is obtained by summing the basic distance from crops to cluster centers and the spatial constraints from crops to cluster centers. Each crop is assigned to the cluster with the smallest total distance, and the cluster center is updated to the feature mean of the crops in the cluster. Repeat the above steps of assigning crops to clusters and updating cluster centers until the cluster centers no longer change or the preset number of iterations is reached. Output spatially continuous regional clusters with the same crop growth stage, and use the maximum spatial boundary of the regional clusters as the geographical boundary of the regional clusters.

5. A method for agricultural environmental monitoring based on 5G Internet of Things according to claim 4, characterized in that: The process of constructing the proximity matrix is ​​as follows: Randomly select two crops from all crops in the adopted area and calculate the distance between them using the Euclidean distance formula. If the distance between the two crops is less than or equal to the spatial distance threshold, the two crops are considered to be adjacent; otherwise, the two crops are considered to be not adjacent. Based on the proximity relationships among all crops, all crops in the adoption area are numbered, and an adjacency matrix is ​​constructed. If crops are adjacent, they are recorded as 1 in the adjacency matrix; if crops are not adjacent, they are recorded as 0 in the adjacency matrix. The rows and columns in the adjacency matrix represent crop numbers, and the proximity relationship of the same crop with the same row and column is recorded as 0 in the adjacency matrix.

6. A method for agricultural environmental monitoring based on 5G Internet of Things according to claim 3, characterized in that: The specific process for assessing whether the growth environment of crops at their current growth stage within the geographical boundaries of the evaluation area clusters is suitable is as follows: The monitoring period is divided into several monitoring time points with equal time intervals. Environmental parameters collected at each monitoring time point are extracted to determine abnormal environmental parameters. Calculate the percentage of abnormal environmental parameters among all environmental parameters; if it is greater than or equal to the threshold for the percentage of abnormal environmental parameters, it indicates that the growth environment of the crop at the current growth stage within the geographical boundary of the regional cluster is unqualified; otherwise, it indicates that the growth environment of the crop at the current growth stage within the geographical boundary of the regional cluster is qualified.

7. A method for agricultural environmental monitoring based on 5G Internet of Things according to claim 6, characterized in that: The process for determining the abnormal environmental parameters is as follows: If an environmental parameter is outside the range of environmental parameters, the corresponding environmental parameter will be recorded as an abnormal environmental parameter.

8. A method for agricultural environmental monitoring based on 5G Internet of Things according to claim 6, characterized in that: The process of performing interference analysis on the growth environment of crops at their current growth stage within the geographical boundaries of regional clusters is as follows: The geographical boundaries of clusters representing areas where the crop's current growth stage has an unsuitable growing environment are denoted as unsuitable environmental areas. The number of abnormal environmental parameters in the non-compliant area during the monitoring period is counted. If the number of abnormal environmental parameters in the non-compliant area during the monitoring period is greater than 1, an analysis signal is generated; otherwise, the corresponding abnormal environmental parameter is recorded as an interference parameter, and the interference parameter is directly adjusted.

9. A method for agricultural environmental monitoring based on 5G Internet of Things according to claim 8, characterized in that: The specific process for determining the control sequence table of abnormal environmental parameters is as follows: Based on the analyzed signals, the number of abnormal environmental parameters in the non-compliant areas during the monitoring period is integrated into an abnormal environmental parameter number sequence. Any abnormal environmental parameter category is extracted from it as an independent variable, and other abnormal environmental parameter categories in the abnormal environmental parameter sequence are used as explanatory variables. A linear regression model is established, and the determination coefficient between the current abnormal environmental parameter category and other abnormal environmental parameter categories is calculated using the coefficient of determination formula. The variance inflation factor between the current abnormal environmental parameter category and the r-th abnormal environmental parameter category is then calculated by substituting the coefficient of determination formula into the variance inflation factor formula. If the variance inflation factor between the current abnormal environmental parameter category and the r-th abnormal environmental parameter category is greater than 1, then the r-th abnormal environmental parameter category is denoted as the collinearity parameter category. Remove all common parameter categories of the current abnormal environment parameter category from the abnormal environment parameter species sequence, and so on. Record the abnormal environment parameter species sequence after removal as the abnormal environment parameter species sequence to be analyzed. Within the monitoring period, the abnormal environmental parameters corresponding to the number of abnormal environmental parameters in the sequence of abnormal environmental parameters to be analyzed are extracted. Using the Spearman correlation coefficient formula, the interference characterization values ​​between the abnormal environmental parameters corresponding to the number of abnormal environmental parameters in the sequence of abnormal environmental parameters to be analyzed and the growth characteristics corresponding to the crop growth stage are calculated. The abnormal environmental parameters are then sorted from largest to smallest to determine the regulation order table of the abnormal environmental parameters.

10. An agricultural environmental monitoring system based on 5G Internet of Things, characterized in that, The system is used to perform the method of any one of claims 1-9, and the system comprises: Growth stage identification module: Extract growth characteristic data of crops in the adopted area from historical crop growth records over multiple historical periods, construct a growth stage identification model based on machine learning model, identify the growth stage of all crops in the adopted area, and determine whether all crops in the adopted area are in the same growth stage; Clustering analysis module: If crops are not in the same growth stage, cluster analysis will be performed on crops in the same growth stage. Spatial constraints will be introduced during the clustering process to ensure that crops in the same cluster are spatially continuous. The output will be a spatially continuous cluster of regions with crops in the same growth stage, and the geographical boundaries of the regional clusters will be determined. Environmental compliance assessment module: Based on any spatially continuous regional cluster with the same crop growth stage, extract the environmental parameters of the current crop growth stage in the regional cluster. By performing deviation analysis on the environmental parameters of the current crop growth stage, assess whether the growth environment of the current crop growth stage within the geographical boundary of the regional cluster is qualified. Interference Analysis Module: If the growth environment is unsuitable, interference analysis is performed on the growth environment of the crop at its current growth stage within the geographical boundaries of the regional cluster. Based on the analysis results, a control sequence table of abnormal environmental parameters is determined to make the growth environment more suitable for the current growth stage of the crop.

Images