Swarm control method based on swarm intelligence control

Abstract

The application relates to the technical field of unmanned aerial vehicle control and provides a swarm control method based on swarm intelligence control, which comprises the following steps: obtaining a hierarchical cluster class according to soil dimension data, and then obtaining the deviation degree of the hierarchical cluster class according to the difference of data in the hierarchical cluster class; obtaining a hierarchical region; obtaining the influence degree of each kind of soil dimension data on an analysis point according to the overlapping relationship between the hierarchical regions and the deviation degree; obtaining soil quality similarity according to the influence degree of each kind of soil dimension data on a prediction point and a sensor installation point; obtaining a theoretical crop height according to the soil quality similarity and the crop height measured by an unmanned aerial vehicle; obtaining a height deviation degree and a suspected deviation point according to the difference between the crop height measured by the unmanned aerial vehicle and the theoretical crop height, and then controlling the unmanned aerial vehicle swarm. The application controls the unmanned aerial vehicle swarm through the suspected deviation point and the height deviation degree, and the accuracy of monitoring the growth height of crops by the unmanned aerial vehicle is improved.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) control technology, and more specifically to a swarm control method based on swarm intelligence control. Background Technology

[0002] The continued growth of the population and the need to ensure food safety place stringent demands on both the scale and quality of agricultural product supply. Traditional agricultural models rely heavily on natural conditions and human experience, resulting in low resource utilization, unstable production efficiency, and rising labor costs. Overuse of chemical fertilizers and pesticides leads to soil degradation and environmental pollution. In contrast, smart agriculture, through IoT sensors and drone inspection technology, enables precise monitoring and intelligent decision-making regarding farmland environment and crop growth. Data-driven precision irrigation, intelligent fertilization, and pest and disease early warning significantly improve resource utilization efficiency, while automated equipment reduces reliance on human labor, forming a sustainable agricultural development model that is resource-saving, environmentally friendly, and highly productive.

[0003] In large-scale smart agriculture management, if drones fly too fast, the scanning accuracy of lidar will be affected by reduced point cloud density, accumulated motion errors, and insufficient scanning overlap. During high-speed flight, the mismatch between the lidar's pulse emission frequency and flight speed can lead to increased spacing between adjacent scan points, resulting in the loss of some plant height details. Simultaneously, the attitude angle measurement errors of the drone's inertial navigation system during high-dynamic flight can amplify point cloud position deviations, especially in undulating terrain or densely cropped areas. This can cause distorted height data or missed scans, leading to inaccurate acquisition of crop height in some farmland locations and consequently, misjudgments of crop growth status. Summary of the Invention

[0004] This invention provides a swarm control method based on swarm intelligence control to solve the problem of inaccurate crop height acquisition in some farmland locations obtained by drone scanning due to the excessive flight speed of drones. The specific technical solution adopted is as follows:

[0005] This invention proposes a swarm control method based on swarm intelligence control, which includes the following steps:

[0006] Several analysis points were set up in the farmland, including several sensor installation points and several prediction points. Several soil dimension data were obtained for each analysis point. The height of crops at each analysis point was obtained by drone scanning.

[0007] Based on the soil dimension data of the analysis points, several level clusters of each soil dimension data are obtained; based on the level clusters of the soil dimension data, the deviation degree of each level cluster of each soil dimension data is obtained; based on the level clusters, several level regions of each soil dimension data are obtained; based on the overlap relationship between the level regions of the soil dimension data and the deviation degree, the degree of influence of each analysis point on each soil dimension data is obtained.

[0008] Based on the similarity between the predicted points and sensor installation points in terms of the influence of each soil dimension data, the soil similarity between each predicted point and each sensor installation point is obtained; based on the soil similarity between the predicted points and sensor installation points, combined with the crop height measured by the UAV at each sensor installation point, the theoretical crop height at each predicted point is obtained; based on the difference between the crop height measured by the UAV at each predicted point and the theoretical crop height, the degree of height deviation at each predicted point is obtained.

[0009] Suspected deviation points are identified based on the degree of altitude deviation; the drone swarm is then controlled based on the suspected deviation points and the degree of altitude deviation.

[0010] Preferably, the specific method for obtaining several levels of clusters for each soil dimension data based on the soil dimension data of the analysis points includes:

[0011] For any type of soil dimension data, the data set consisting of the data values ​​of all soil integrated sensors and the predicted values ​​of all prediction points is denoted as the single-dimensional monitoring set of that type of soil dimension data. The elbow method is used to obtain the optimal number of clusters in the single-dimensional monitoring set. The optimal number of clusters is used as the K value in the K-means clustering algorithm. The elements in the single-dimensional monitoring set are clustered into K clusters, which are denoted as the hierarchical clusters of that type of soil dimension data.

[0012] Preferably, the specific method for obtaining the deviation degree of each level cluster of each soil dimension data based on the level clusters of the soil dimension data includes:

[0013] The deviation of the d-th level cluster of the q-th soil dimension data is calculated as follows:

[0014]

[0015] In the formula, C q,d E represents the degree of deviation of the d-th level cluster of the q-th soil dimension data; q,d E' represents the number of elements in the d-th level cluster of the q-th soil dimension data. qD represents the number of elements in the highest-order cluster among all order clusters of the q-th soil dimension data; q,d Let D' be the mean of the elements in the d-th level cluster of the q-th soil dimension data; q is the mean of the elements in the highest-numbered cluster among all the clusters of the q-th soil dimension data; exp() is the exponential function with the natural constant as the base; norm() is the linear normalization function; || is the absolute value function.

[0016] Preferably, the specific method for obtaining several level regions for each soil dimension data according to the level cluster class includes:

[0017] For any level cluster of any type of soil dimension data, the area formed by the analysis points corresponding to the elements belonging to that level cluster in the farmland is denoted as a level region of that type of soil dimension data.

[0018] Preferably, the method for determining the degree of influence of each analysis point on each type of soil dimension data based on the overlap between the graded regions of the soil dimension data and the degree of deviation includes:

[0019] The correlation between soil dimension data of type e and soil dimension data of type g in the f-th level region of soil dimension data of type e is calculated as follows:

[0020]

[0021] In the formula, B e,f,g For the f-th level region of the e-th soil dimension data, the correlation between the e-th soil dimension data and the g-th soil dimension data; F e,f G represents the number of analysis points in the f-th level region of the e-th soil dimension data; e,f,g,h G' represents the total number of analysis points in the level region of the f-th level region of the e-th soil dimension data, where the h-th analysis point is located in the level region of the g-th soil dimension data; e,f,g,h C' is the number of analytical points in the f-th level region of the e-th soil dimension data that overlap with the h-th analytical point in the g-th soil dimension data. e,f,h C″ represents the degree of deviation of the h-th analysis point within the f-th level region of the e-th soil dimension data from the corresponding level cluster; e,f,g,h ε represents the degree of deviation of the h-th analysis point in the f-th level region of the e-th soil dimension data from the level cluster of the g-th soil dimension data; || is the absolute value function; ε is the hyperparameter;

[0022] Based on the correlation between the soil dimension data of each analysis point and all other soil dimension data in the grade region where each soil dimension data is located, the degree of influence of each analysis point on each soil dimension data is obtained.

[0023] Preferably, the method for determining the degree of influence of each analysis point on each type of soil dimension data based on the correlation between the soil dimension data of each analysis point in the grade region where each type of soil dimension data is located and all other types of soil dimension data includes:

[0024] For any analysis point and any type of soil dimension data, the average correlation between that type of soil dimension data and all other types of soil dimension data within the grade region where that analysis point is located is recorded as the degree of influence of that type of soil dimension data on that analysis point.

[0025] Preferably, the method for obtaining the soil similarity between each predicted point and each sensor installation point based on the similarity relationship between the degree of influence of each soil dimension data on the predicted point and the sensor installation point is as follows:

[0026] The soil similarity between the k-th predicted point and the l-th sensor installation point is calculated as follows:

[0027] H k,l =exp(-(MAX(|B′) k -B′ l |)))

[0028] In the formula, H k,l The soil similarity between the k-th predicted point and the l-th sensor installation point; MAX(|B' k -B′ l | represents the maximum difference between the degree of influence of the k-th predicted point and the l-th sensor installation point on the same type of soil dimension data across all types of soil dimension data; || is the absolute value function; exp() is an exponential function with the natural constant as the base.

[0029] Preferably, the method for obtaining the theoretical crop height for each predicted point based on the soil similarity between the predicted point and the sensor installation point, combined with the crop height measured by the UAV at each sensor installation point, includes the following specific methods:

[0030] The theoretical crop height at the k-th prediction point is calculated as follows:

[0031]

[0032] In the formula, L k Here, H represents the theoretical crop height at the k-th prediction point; N is the number of sensor installation points;k,l M represents the soil similarity between the k-th predicted point and the l-th sensor installation point. l The drone measures crop height at the l-th sensor installation point; softmax() is the weight normalization function.

[0033] Preferably, the method for obtaining the degree of height deviation at each predicted point based on the difference between the crop height measured by the UAV at each predicted point and the theoretical crop height includes:

[0034] For any predicted location, the linear normalization result of the absolute value of the difference between the theoretical crop height at the predicted location and the crop height measured by the UAV is denoted as the degree of height deviation at the predicted location.

[0035] Preferably, the specific method for filtering suspected deviation points based on the degree of height deviation includes:

[0036] Predicted points with a height deviation greater than the preset deviation threshold are recorded as suspected deviation points.

[0037] The beneficial effects of this invention are:

[0038] When using drones to monitor crop height in farmland, soil conditions in some areas can affect crop growth. However, drones, flying too fast over farmland, cannot accurately capture the height of these localized areas affected by soil conditions. Therefore, soil analysis at different locations is necessary. This invention addresses this by deploying integrated soil sensors to analyze regional differences in soil dimensional data within localized farmland soil conditions, determining the degree of influence of each soil dimensional data point on the soil. For determining soil similarity, this invention compares the predicted location with the sensor installation location based on the influence of each soil dimensional data point on the soil. The similarity relationship of soil dimension data influence is used to obtain the soil similarity between each predicted point and each sensor installation point. To address the problem of inaccurate crop height measurement due to excessive drone flight speed in farmland, this invention combines the soil similarity between predicted and sensor installation points with the crop height measured by the drone at each sensor installation point to obtain the theoretical crop height for each predicted point. Then, the drone-measured crop height is compared with the theoretical crop height to determine the degree of height deviation for each predicted point, judging the possibility that crop growth at each predicted point is affected by soil conditions in localized areas of farmland. Thus, this invention identifies suspected deviation points through height deviation screening, and controls the drone swarm based on these suspected deviation points and the aforementioned height deviation degree. This improves the accuracy of drone monitoring of crop height for localized farmland areas with soil differences. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a schematic diagram of a cluster control method based on swarm intelligence control provided in an embodiment of the present invention. Detailed Implementation

[0041] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0042] Please see Figure 1 The diagram illustrates a swarm control method based on swarm intelligence control according to an embodiment of the present invention. The method includes the following steps:

[0043] Step S001: Set up several analysis points in the farmland and obtain several soil dimension data for each analysis point; use a drone to scan and obtain the drone-measured crop height for each analysis point.

[0044] It should be noted that, in order to improve the scanning accuracy of the drones while minimizing the impact on the scanning rate, when controlling the drone swarm, it is necessary to control the drones to perform low-speed targeted scanning of local areas where the crop growth height deviates from that of other farmland locations. The direct factor affecting the crop growth height is the soil quality, and the difference in soil quality is mainly reflected in the different contents of various soil properties in the farmland. Therefore, a comprehensive soil sensor is deployed in the farmland to acquire various soil properties.

[0045] Specifically, the farmland requiring intelligent management is divided into A×A square grids, with the center point of each square grid serving as a sensor installation point. A comprehensive soil sensor is installed at each sensor installation point, which integrates a soil moisture sensor, a soil temperature sensor, a soil conductivity sensor, a pH sensor, a soil nitrogen, phosphorus, and potassium sensor, and a soil carbon dioxide sensor. Here, A is the preset grid width, and this embodiment uses A = 20 meters as an example.

[0046] After all the integrated soil sensors are installed, for any one integrated soil sensor, a multidimensional soil monitoring data is collected once before the drone flies. The multidimensional soil monitoring data includes soil moisture, soil temperature, soil electrical conductivity, soil pH, soil nitrogen, phosphorus and potassium content, and soil carbon dioxide concentration. Each data value is a type of soil dimension data.

[0047] Based on an A×A square grid, each square grid is further divided into A equal parts. 2 A subgrid, that is, dividing each square grid into A 2 Each subgrid has a side length of 1 meter, and the center point of each subgrid is used as a prediction point.

[0048] For any type of soil dimension data, based on all integrated soil sensors for that soil dimension data, the ordinary kriging method is used to interpolate all predicted points to obtain the predicted value of each predicted point for that soil dimension data; among them, the ordinary kriging method is a well-known technique, and the specific method will not be introduced here.

[0049] The sensor installation points and prediction points in farmland are collectively referred to as analysis points;

[0050] A lidar is deployed on a drone, which then scans the farmland at a constant speed of 12 meters per second to obtain three-dimensional point cloud data, thus obtaining the drone-measured crop height at each analysis point.

[0051] Step S002: Based on the soil dimension data of the analysis points, obtain several level clusters for each type of soil dimension data; based on the level clusters of the soil dimension data, obtain the deviation degree of each level cluster for each type of soil dimension data; based on the level clusters, obtain several level regions for each type of soil dimension data; based on the overlap relationship between the level regions of the soil dimension data and the deviation degree, obtain the degree of influence of each analysis point on each type of soil dimension data.

[0052] It should be noted that the direct factor affecting crop growth is soil quality. However, in some localized areas, soil quality can change due to variations in slope, human factors, and other reasons. These changes primarily manifest as varying degrees of alteration in each soil dimension, thus affecting the height of crop growth. Therefore, it is necessary to determine the extent to which each soil dimension affects crops in each localized area.

[0053] Specifically, for any type of soil dimension data, the data set consisting of the data values ​​from all soil integrated sensors and the predicted values ​​from all prediction points is denoted as the single-dimensional monitoring set for that type of soil dimension data. The elbow method is used to obtain the optimal number of clusters in the single-dimensional monitoring set, and the optimal number of clusters is used as the K value in the K-means clustering algorithm. The elements in the single-dimensional monitoring set are clustered into K clusters, which are denoted as the hierarchical clusters of that type of soil dimension data. The elbow method and K-means clustering are well-known techniques, and the specific methods are not described here.

[0054] It should be noted that due to differences in soil quality in certain areas of farmland, some soil dimensional data may differ from those of other farmland areas.

[0055] Specifically, the deviation of the d-th level cluster of the q-th soil dimension data is calculated as follows:

[0056]

[0057] In the formula, C q,d E represents the degree of deviation of the d-th level cluster of the q-th soil dimension data; q,d E' represents the number of elements in the d-th level cluster of the q-th soil dimension data. q D represents the number of elements in the highest-order cluster among all order clusters of the q-th soil dimension data; q,d Let D' be the mean of the elements in the d-th level cluster of the q-th soil dimension data; q Let be the mean of the elements in the rank cluster with the most elements among all rank clusters of the q-th soil dimension data; exp() is an exponential function with the natural constant as the base; norm() is a linear normalization function, which normalizes the (|D) of all rank clusters of the q-th soil dimension data. q,d -D' q |);|| is the absolute value function.

[0058] In the formula, The larger the value, the fewer elements in the d-th level cluster of the q-th soil dimension data are in the q-th soil dimension data, indicating that the d-th level cluster of the q-th soil dimension data differs from the general data of the same soil dimension data; norm(|D q,d -D' q The larger the |), the greater the difference between the d-th level cluster of the q-th soil dimension data and the general data, and the greater the possibility that the d-th level cluster of the q-th soil dimension data deviates from the normal value.

[0059] It should be noted that in farmland soil monitoring, for each soil dimension data, there is a characteristic of local data similarity in farmland. For example, the gradual change in soil quality between farmlands leads to similarities in soil quality within a local area, but there are differences in soil quality between regions. For example, when the slope of farmland changes, water accumulates in low-lying areas, while strong evaporation at the top of the slope leads to salinization. The unevenness of human management measures, such as fertilization, irrigation, and tillage, often results in differences in implementation across regions. Therefore, for each soil dimension data, there is a characteristic of local data similarity in farmland. That is, for each level cluster of soil dimension data, there is a characteristic of clustered distribution in the two-dimensional plane space where the farmland is located.

[0060] Specifically, for any level cluster of any type of soil dimension data, the area formed by the analysis points corresponding to the elements belonging to that level cluster in the farmland is denoted as a level region of that type of soil dimension data.

[0061] It should be noted that in farmland soil monitoring, each soil dimension data affects crop growth. However, the impact on crop growth is not a linear additive relationship between soil dimension data; there are correlations between them, and certain soil dimension data collectively influence crop growth. These correlations are mainly manifested in the fact that different local areas, due to varying soil types, exhibit differences in certain soil dimension data compared to other locations. These differences cause these local soil dimension data to collectively affect crop growth. For example, in low-lying, waterlogged areas, high humidity and salt accumulation due to evaporation and concentration often occur simultaneously, forming salinization stress zones. In fields where organic fertilizers have been applied for a long time, increases in organic matter content, microbial activity, and CO2 concentration are often accompanied by enhanced pH buffering capacity, exhibiting a synergistic improvement in patchy distribution. Therefore, by assessing the overlapping distribution of different soil dimension data levels, we can determine the impact of the composition of soil dimension data for different soil types on crops, and thus obtain the correlations between soil dimension data under different soil types.

[0062] Specifically, the correlation between the e-th soil dimension data and the g-th soil dimension data in the f-th level region of the e-th soil dimension data is calculated as follows:

[0063]

[0064] In the formula, B e,f,g For the f-th level region of the e-th soil dimension data, the correlation between the e-th soil dimension data and the g-th soil dimension data; F e,f G represents the number of analysis points in the f-th level region of the e-th soil dimension data; e,f,g,hG' represents the total number of analysis points in the level region of the f-th level region of the e-th soil dimension data, where the h-th analysis point is located in the level region of the g-th soil dimension data; e,f,g,h C' is the number of analytical points in the f-th level region of the e-th soil dimension data that overlap with the h-th analytical point in the g-th soil dimension data. e,f,h C″ represents the degree of deviation of the h-th analysis point within the f-th level region of the e-th soil dimension data from the corresponding level cluster; e,f,g,h ε represents the deviation of the h-th analysis point in the f-th level region of the e-th soil dimension data from the level cluster of the g-th soil dimension data; || is the absolute value function; ε is a hyperparameter to prevent the denominator from being zero.

[0065] In the formula, The larger the value, the greater the overlap between the f-th level region of the e-th soil dimension data and the h-th analysis point within it in the g-th soil dimension data. The larger the value, the greater the value of C'. e,f,h With C″ e,f,g,h The fact that they are both relatively large and close indicates that the f-th level region of the e-th soil dimension data and the h-th analysis point within it both exhibit deviations from typical soil types in the g-th soil dimension data; if As it gets bigger, The larger the value, the more likely it is that the data for the e-th and g-th soil dimensions at the analysis point have shifted due to soil conditions; the e-th and g-th soil dimension data will jointly affect crop growth.

[0066] It should be noted that for a local farmland area with soil quality deviations compared to general farmland, some soil dimension data of this local farmland area will also be biased. That is, it shows that there is a high correlation between some soil dimension data at a single analysis point. These highly correlated soil dimension data collectively affect crop growth. Therefore, it is necessary to determine the main soil dimension data that affect crop growth in this local farmland area.

[0067] It should be further noted that if a certain soil dimension data is one of the main factors affecting crop growth in a local farmland area, then this soil dimension data will have a high correlation with the soil dimension data of other main factors. If a certain soil dimension data is not a main factor affecting crop growth in a local farmland area, then the correlation between this soil dimension data and other soil dimension data will be low.

[0068] Specifically, for any analysis point and any type of soil dimension data, the average correlation between that type of soil dimension data and all other types of soil dimension data within the grade region where that analysis point is located is recorded as the degree of influence of that type of soil dimension data on that analysis point.

[0069] Step S003: Based on the similarity between the predicted points and the sensor installation points in terms of the influence of each soil dimension data, obtain the soil similarity between each predicted point and each sensor installation point; based on the soil similarity between the predicted points and the sensor installation points, and combined with the crop height measured by the UAV at each sensor installation point, obtain the theoretical crop height at each predicted point; based on the difference between the crop height measured by the UAV at each predicted point and the theoretical crop height, obtain the height deviation degree at each predicted point.

[0070] It should be noted that in large-scale farmland, natural factors (such as micro-topography, parent material differences, and groundwater fluctuations) and human activities (such as zoned fertilization, uneven irrigation, and mechanical compaction) can lead to local soil variations. The combination of soil dimension data with a high degree of influence is different in different soil types (such as saline soil, clay soil, and sandy loam), which can be used to distinguish soil types.

[0071] It should be further explained that the data obtained by the drone is a height model. The direct factor affecting the growth height of crops is soil quality. Since the various soil dimension data at each sensor installation point are actually measured by the integrated sensor, it can represent the overall situation of each soil dimension data in the farmland within a local area. Therefore, by combining the degree of influence of each integrated sensor installation point on each soil dimension data with the crop height measured by the drone, and comparing it with the predicted points of soil dimension data with the same high degree of influence, the theoretical crop height of the predicted points can be inferred.

[0072] Specifically, the soil similarity between the k-th prediction point and the l-th sensor installation point is calculated as follows:

[0073] H k,l =exp(-(MAX(|B′) k -B′ l |)))

[0074] In the formula, H k,l The soil similarity between the k-th predicted point and the l-th sensor installation point; MAX(|B' k -B' l|) represents the maximum difference between the degree of influence of the k-th predicted point and the l-th sensor installation point on the same type of soil dimension data among all types of soil dimension data; || is the absolute value function; exp() is the exponential function with the natural constant as the base.

[0075] Among them, H k,l The larger the value, the greater the MAX(|B') value. k -B' l The smaller the |) is, the smaller the difference in influence between the k-th prediction point and the l-th sensor installation point on the soil dimension data with the greatest difference in influence is still very small. Since the combination of soil dimension data with a high degree of influence is different in different soil types, the soil at the k-th prediction point and the l-th sensor location is more likely to be similar.

[0076] It should be noted that similar soil types often produce crops of similar height. Since the various soil dimension data at each sensor installation point are actually measured by the integrated sensor, the accuracy of the various soil dimension data at each sensor installation point is relatively high. Therefore, the predicted height of the crop at the prediction point can be inferred based on the similarity of the soil type between the prediction point and the sensor installation point.

[0077] Specifically, the theoretical crop height at the k-th prediction point is calculated as follows:

[0078]

[0079] In the formula, L k Here, H represents the theoretical crop height at the k-th prediction point; N is the number of sensor installation points; k,l M represents the soil similarity between the k-th predicted point and the l-th sensor installation point. l The drone measures crop height at the l-th sensor installation point; softmax() is a weight normalization function, which normalizes the soil similarity between the k-th prediction point and all sensor installation points.

[0080] It should be noted that if there is a discrepancy between the theoretical crop height at the predicted location and the crop height measured by the drone at the predicted location, it indicates that the soil quality at that predicted location may have undergone local changes.

[0081] Specifically, for any given prediction point, the linear normalization result of the absolute value of the difference between the theoretical crop height and the crop height measured by the UAV at that prediction point is denoted as the degree of height deviation at that prediction point; where the object of linear normalization is the absolute value of the difference between the theoretical crop height and the crop height measured by the UAV at all prediction points.

[0082] Step S004: Filter out suspected deviation points based on the degree of altitude deviation; control the drone swarm based on the suspected deviation points and the degree of altitude deviation.

[0083] It should be noted that the greater the height deviation of the predicted points, the more likely that the soil at the predicted points has undergone local changes, and that these changes have affected crop growth. Therefore, it is necessary to improve the accuracy of UAV monitoring of crop growth at these predicted points.

[0084] Specifically, predicted points with height deviations greater than a preset deviation threshold are recorded as suspected deviation points; the preset deviation threshold is 0.7, and this embodiment will be described using this as an example.

[0085] In the next scan of farmland using drones, a particle swarm optimization algorithm will be used to control the flight of the drone swarm. The learning factor c2 and the inertial weight w in the particle swarm optimization algorithm will be adaptively adjusted. The specific adjustment process is as follows:

[0086] For any drone, the coordinates of the suspected deviation point that is closest to the drone in Euclidean distance are set as the global optimal position of the particle swarm optimization algorithm during the drone's flight.

[0087] The dynamic learning factor is calculated as follows:

[0088] c″2=c2+α×P

[0089] In the formula, c2 is a preset learning factor, and this embodiment uses c2 = 2 as an example; α is a preset gain coefficient, and this embodiment uses α = 0.1 as an example; P is the degree of height deviation of the suspected deviation point closest to the UAV in Euclidean distance;

[0090] The dynamic inertia weight is calculated as follows:

[0091] w'=w×e -P

[0092] In the formula, w' is the dynamic inertia weight; w is the preset inertia weight, and this embodiment uses w=0.9 as an example for description; e is the natural constant; P is the degree of altitude deviation of the suspected deviation point closest to the UAV's Euclidean distance.

[0093] It should be noted that by increasing the learning factor c2, the attractiveness of high deviation areas to the UAV is enhanced, and by reducing the inertial weight w, the inertia of the UAV is reduced, allowing it to decelerate and perform refined scanning near suspected deviation points.

[0094] Furthermore, the adjusted dynamic learning factor c'2 and dynamic inertia weight w' are used for control using the particle swarm optimization algorithm, and UWB local communication is used to maintain the formation spacing to avoid collisions; among them, the particle swarm optimization algorithm and the UWB local communication algorithm are well-known technologies, and the specific methods are not described here.

[0095] After scanning farmland using drones, three-dimensional point cloud data of crops in local farmland areas with local differences in soil quality are obtained. Then, the accurate crop height of each analysis point is obtained, and corresponding irrigation and fertilization management measures are implemented for crops at different heights.

[0096] This embodiment uses the exp(-MX) model to represent the inverse proportional relationship and normalization processing. MX is the input of the model, and the implementer can set the inverse proportional function and normalization function according to the actual situation.

[0097] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A cluster control method based on swarm intelligence control, characterized by, The method comprises the following steps: A plurality of analysis points are arranged in the farmland, the analysis points comprise a plurality of sensor installation points and a plurality of prediction points, and soil dimension data of each analysis point is acquired; and a crop height of each analysis point is acquired by using a UAV to scan; According to the soil dimension data of the analysis points, a plurality of hierarchical cluster classes of each soil dimension data are obtained; according to the hierarchical cluster classes of the soil dimension data, a deviation degree of each hierarchical cluster class of each soil dimension data is obtained; according to the hierarchical cluster classes, a plurality of hierarchical regions of each soil dimension data are obtained; and according to an overlapping relationship between the hierarchical regions of the soil dimension data and the deviation degree, an influence degree of each analysis point on each soil dimension data is obtained; According to a similarity relationship of the influence degrees of each prediction point and each sensor installation point on each soil dimension data, a soil quality similarity between each prediction point and each sensor installation point is obtained; according to the soil quality similarity between the prediction points and the sensor installation points, in combination with the UAV-measured crop height of each sensor installation point, a theoretical crop height of each prediction point is obtained; and according to a difference relationship between the UAV-measured crop height and the theoretical crop height of each prediction point, a height deviation degree of each prediction point is obtained; According to the height deviation degree, a suspected deviation point is screened out; and a UAV cluster is controlled according to the suspected deviation point and the height deviation degree.

2. The swarm control method based on swarm intelligence control according to claim 1, characterized in that, The method of obtaining the plurality of hierarchical cluster classes of each soil dimension data according to the soil dimension data of the analysis points comprises the following specific method: For any one soil dimension data, a data set composed of data values of all soil comprehensive sensors and prediction values of all prediction points is regarded as a single-dimension monitoring set of the soil dimension data, an elbow method is used to obtain an optimal cluster class number of the single-dimension monitoring set, the optimal cluster class number is taken as a K value in a K-means clustering algorithm, and elements in the single-dimension monitoring set are clustered into K cluster classes, which are regarded as hierarchical cluster classes of the soil dimension data.

3. The swarm control method based on swarm intelligence control according to claim 1, characterized in that, The method of obtaining the deviation degree of each hierarchical cluster class of each soil dimension data according to the hierarchical cluster classes of the soil dimension data comprises the following specific method: A calculation method of the deviation degree of the dth hierarchical cluster class of the qth soil dimension data is as follows: wherein C q,d is the deviation degree of the dth hierarchical cluster class of the qth soil dimension data; E q,d is the number of elements of the dth hierarchical cluster class of the qth soil dimension data; E' q is the number of elements of the hierarchical cluster class with the largest number of elements among all hierarchical cluster classes of the qth soil dimension data; D q,d is the mean of elements of the dth hierarchical cluster class of the qth soil dimension data; D' q is the mean of elements of the hierarchical cluster class with the largest number of elements among all hierarchical cluster classes of the qth soil dimension data; exp() is an exponential function with a natural constant as a base; norm() is a linear normalization function; and || is an absolute value function.

4. The swarm control method based on swarm intelligence control according to claim 1, characterized in that, The method of obtaining the plurality of hierarchical regions of each soil dimension data according to the hierarchical cluster classes comprises the following specific method: For any one hierarchical cluster class of any one soil dimension data, a region composed of analysis points corresponding to elements belonging to the hierarchical cluster class in the farmland is regarded as a hierarchical region of the soil dimension data.

5. The swarm control method based on swarm intelligence control according to claim 1, characterized in that, The method of obtaining the influence degree of each analysis point on each soil dimension data according to the overlapping relationship between the hierarchical regions of the soil dimension data and the deviation degree comprises the following specific method: A calculation method of the correlation of the e th soil dimension data and the g th soil dimension data in the f th hierarchical region of the e th soil dimension data is as follows: In the formula, B e,f,g is the relevance of the e-th soil dimension data to the g-th soil dimension data in the f-th hierarchical region of the e-th soil dimension data; F e,f is the number of analysis points in the f-th hierarchical region of the e-th soil dimension data; G e,f,g,h is the total number of analysis points in the hierarchical region of the g-th soil dimension data in which the h-th analysis point of the f-th hierarchical region of the e-th soil dimension data is located; G' e,f,g,h is the number of analysis points in the f-th hierarchical region of the e-th soil dimension data in which the h-th analysis point is overlapped with the hierarchical region in which the h-th analysis point of the e-th soil dimension data is located in the g-th soil dimension data; C' e,f,h is the deviation degree of the data corresponding to the h-th analysis point in the f-th hierarchical region of the e-th soil dimension data to the hierarchical cluster class; C" e,f,g,h is the deviation degree of the h-th analysis point in the f-th hierarchical region of the e-th soil dimension data to the hierarchical cluster class in the g-th soil dimension data; || is an absolute value function; and ε is a hyperparameter. According to the correlation of the soil dimension data of the analysis point in the grade area of each soil dimension data with all other soil dimension data, the influence degree of each analysis point on each soil dimension data is obtained.

6. The swarm control method based on swarm intelligence control according to claim 5, characterized in that, The specific method comprises the following steps: For any analysis point and any soil dimension data, the average of the correlation of the analysis point in the grade area of the soil dimension data with all other soil dimension data is recorded as the influence degree of the analysis point on the soil dimension data.

7. The swarm control method based on swarm intelligence control according to claim 1, characterized in that, The specific method comprises the following steps: The calculation method of the soil quality similarity between the kth prediction point and the lth sensor installation point is as follows: H k,l = exp(-(MAX(|B′ k -B′ l |))) wherein H k,l is the soil similarity between the kth prediction point and the lth sensor installation point; MAX(|B k -B l |) is the maximum value of the difference between the influence degree of the same soil dimension data on the kth prediction point and the lth sensor installation point in all soil dimension data; || is the absolute value function; and exp() is the exponential function with the natural constant as the base.

8. The swarm control method based on swarm intelligence control according to claim 1, characterized in that, The specific method comprises the following steps: The calculation method of the theoretical crop height of the kth prediction point is as follows: In the formula, L k is the theoretical crop height of the kth prediction point; N is the number of sensor installation points; H k,l is the soil similarity of the kth prediction point and the lth sensor installation point, M l is the unmanned aerial vehicle measured crop height of the lth sensor installation point; The softmax() is a weight normalization function.

9. The swarm control method based on swarm intelligence control according to claim 1, characterized in that, The specific method comprises the following steps: For any prediction point, the linear normalization result of the absolute value of the difference between the theoretical crop height and the unmanned aerial vehicle measured crop height of the prediction point is recorded as the height deviation degree of the prediction point.

10. The swarm control method based on swarm intelligence control according to claim 1, wherein, The specific method comprises the following steps: The prediction point with a height deviation degree greater than a preset deviation threshold is recorded as a suspected deviation point.

Images