Water resource intelligent management and control device debugging method and system based on short-distance transmission

Abstract

This invention discloses a debugging method for intelligent water resource management and control equipment based on short-range transmission, belonging to the field of equipment debugging technology. The key technical points include: for each intelligent water resource management and control device, obtaining the average competitive advantage index based on corresponding environmental data; obtaining the device calibration threshold and device calibration window based on the average competitive advantage index; generating basic debugging parameters based on the device calibration threshold and device calibration window; determining the interference type based on the current sensor data and random forest model of each intelligent water resource management and control device, and obtaining the corresponding optimized debugging parameters based on the interference type; and completing the full-dimensional debugging of the intelligent water resource management and control device based on the basic debugging parameters and optimized debugging parameters. This invention achieves refined monitoring and management of irrigation water resources through the synergy of basic debugging parameters and optimized debugging parameters.

Description

Technical Field

[0001] This invention relates to the field of equipment commissioning technology, and more specifically to a method and system for commissioning intelligent water resource management and control equipment based on short-range transmission. Background Technology

[0002] With the transformation of modern agriculture towards precision and intelligence, the refined monitoring and management of farmland irrigation water resources has become one of the core means to improve water resource utilization. Electromagnetic flowmeters, as key terminal equipment for distributed irrigation flow monitoring, need to achieve data interaction between devices and between devices and control terminals through short-range wireless communication (such as Bluetooth) to support functions such as irrigation scheduling and flow calibration. However, the farmland environment is complex, with various dynamic scenarios such as water droplet adhesion and crop shading. Furthermore, the devices are dispersed and have large differences in flow characteristics. Traditional monitoring solutions are difficult to adapt to the dynamic interference and heterogeneity of farmland, leading to problems such as high data transmission packet loss rates, inaccurate interference identification, and unreasonable device grouping and parameter configuration. These issues seriously affect the reliability and accuracy of irrigation flow monitoring, thus existing technologies have shortcomings. Summary of the Invention

[0003] To address the shortcomings of existing technologies, the present invention aims to provide a debugging method for intelligent water resource management and control equipment based on short-range transmission. The method groups the equipment by competitive potential index and optimizes the calibration threshold and calibration window for each group of equipment, thus avoiding the problem of mismatch between equipment configuration parameters and actual scenarios. Furthermore, the method uses a random forest model to identify dynamic interference in farmland and generate optimized debugging parameters, thereby avoiding the problem of weak anti-interference capability of equipment transmission at close range.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] This invention provides a debugging method for intelligent water resource management and control equipment based on short-range transmission, comprising:

[0006] For each intelligent water resource management and control device, the average competitive advantage index is obtained based on the corresponding environmental data.

[0007] The device calibration threshold and device calibration window are obtained based on the mean value of the competitive advantage index.

[0008] Basic debugging parameters are generated based on the equipment calibration threshold and the equipment calibration window.

[0009] The interference type is determined based on the current sensor data and random forest model of each water resource intelligent management and control device, and the corresponding optimization and debugging parameters of the water resource intelligent management and control device are obtained based on the interference type. The interference type includes water droplet interference and crop shading interference.

[0010] Based on the basic debugging parameters and the optimized debugging parameters, complete the full-dimensional debugging of the intelligent water resource management and control equipment.

[0011] As a further improvement of the present invention, the step of obtaining the device calibration threshold and the device calibration window based on the mean value of the competitive potential index includes:

[0012] Based on the average competitive advantage index, each water resource intelligent management and control device is divided into multiple first groups;

[0013] Based on the first group, the boundary device is determined, and based on the boundary device, the first group is updated to obtain multiple second groups;

[0014] Based on the second grouping and the particle swarm optimization algorithm, the device calibration threshold corresponding to each second group is obtained;

[0015] Based on the flow feature vector and flow fluctuation variance corresponding to the boundary device, the device calibration window corresponding to the boundary device is obtained.

[0016] As a further improvement of the present invention, the step of dividing each water resource intelligent management and control device into multiple first groups based on the average competitive advantage index includes:

[0017] The first iteration operation includes: obtaining a fusion weight matrix based on the mean competitive advantage index of each intelligent water resource management device and the current weight; obtaining a standardized Laplace matrix based on the fusion weight matrix and the degree matrix; performing spectral decomposition and cluster analysis based on the standardized Laplace matrix to obtain the current first group; determining whether the current first group has reached a preset termination condition; if not, updating the current weight until the current first group reaches the preset termination condition; and outputting the multiple first groups.

[0018] As a further improvement of the present invention, the step of updating the first group according to the boundary device to obtain a plurality of second groups includes:

[0019] Based on the mean of the competitive potential index, the feature vector corresponding to each boundary device is obtained;

[0020] Obtain the feature mean value corresponding to each first group, and obtain the similarity based on the feature vector and the feature mean value;

[0021] The first group is updated based on the similarity, resulting in multiple second groups.

[0022] As a further improvement of the present invention, the step of determining the boundary device according to the first group includes:

[0023] The embedding matrix is ​​obtained from the normalized Laplacian matrix;

[0024] The posterior probability corresponding to each intelligent water resource management and control device is obtained based on the embedding matrix and activation function.

[0025] Boundary devices are determined in the first group based on posterior probabilities.

[0026] As a further improvement of the present invention, the step of obtaining the device calibration threshold corresponding to each second group based on the second grouping and the particle swarm optimization algorithm includes:

[0027] The second iteration operation includes obtaining the traffic error rate and packet loss rate corresponding to the current device calibration threshold for each second group, obtaining a fitness function based on the traffic error rate and packet loss rate, updating the current device calibration threshold based on the fitness function, until a preset termination condition is reached, and outputting the device calibration threshold corresponding to each second group.

[0028] As a further improvement of the present invention, the step of obtaining the device calibration window corresponding to the boundary device based on the flow feature vector and flow fluctuation variance corresponding to the boundary device includes:

[0029] The irrigation method of the boundary device is determined based on the flow characteristic vector, the flow fluctuation variance, and the preset flow characteristic mean.

[0030] The standard deviation of flow fluctuation corresponding to the boundary equipment is obtained based on the irrigation method.

[0031] The device calibration window corresponding to the boundary device is obtained based on the standard deviation of the flow fluctuation.

[0032] As a further improvement of the present invention, the step of determining the interference type based on the current sensor data of each water resource intelligent management and control device and the random forest model includes:

[0033] Select splitting features for each tree in the random forest model based on the current covariance matrix;

[0034] For each splitting feature, determine whether the current sensor data satisfies the corresponding judgment rule;

[0035] The type of interference is determined based on the judgment result.

[0036] As a further improvement of the present invention, the above-mentioned intelligent water resource management and control device method based on short-range transmission also includes:

[0037] Obtain the feature mean value corresponding to the interference data to obtain the feature deviation vector;

[0038] Calculate the outer product matrix of the feature deviation vectors;

[0039] The current covariance matrix is ​​updated based on the outer product matrix, the current covariance matrix, and the amount of interference data.

[0040] This invention provides a commissioning system for intelligent water resource management and control equipment based on short-range transmission, which should include:

[0041] The calculation module is used to obtain the average competitive advantage index for each water resource intelligent management and control device based on the corresponding environmental data.

[0042] The calibration module is used to obtain the device calibration threshold and the device calibration window based on the mean value of the competitive advantage index;

[0043] The interference identification module is used to determine the type of interference based on the current sensor data and random forest model of each water resource intelligent management and control device, and to obtain the corresponding optimization and debugging parameters of the water resource intelligent management and control device based on the type of interference. The types of interference include water droplet interference and crop shading interference.

[0044] The debugging module is used to generate basic debugging parameters based on the equipment calibration threshold and the equipment calibration window; and to complete the full-dimensional debugging of the intelligent water resource management and control equipment based on the basic debugging parameters and the optimized debugging parameters.

[0045] This invention constructs a fusion weight matrix using a competitive potential index and divides the system into multiple first groups. This ensures that devices belonging to the same group have consistent spatial locations, pipeline affiliations, and crop competition statuses, serving as the basis for subsequent adjustments to device thresholds and device windows. For boundary devices, their grouping is determined through similarity, resolving the issue of unclear affiliation caused by physical location and cross-domain scenarios, thus avoiding the ambiguity of traditional static division. Furthermore, the calibration threshold is adjusted based on iterative steps to adapt to the dynamic characteristics of farmland. The device calibration threshold and device window threshold are then used as basic debugging parameters to complete the basic debugging of the intelligent water resource management equipment, providing accurate monitoring data for refined water resource management. In addition, this invention uses a random forest model to identify interference types in real time and generate corresponding optimized debugging parameters, achieving real-time response to variable interferences and ensuring the continuity and stability of monitoring data. Attached Figure Description

[0046] Figure 1 This is a schematic diagram of the method steps of the present invention;

[0047] Figure 2 This is a schematic diagram of the steps for the first iteration operation;

[0048] Figure 3 This is a schematic diagram of the steps for the second iteration operation;

[0049] Figure 4A schematic diagram illustrating the steps for determining the device calibration window. Detailed Implementation

[0050] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof.

[0051] The term "and / or" in the following text is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Additionally, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0052] Example 1

[0053] like Figure 1 As shown, this embodiment provides a debugging method for intelligent water resource management and control equipment based on short-range transmission, including:

[0054] For each intelligent water resource management and control device, the average competitive advantage index is obtained based on the corresponding environmental data.

[0055] The equipment calibration threshold and equipment calibration window are obtained based on the mean of the competitive advantage index.

[0056] Basic debugging parameters are generated based on the equipment calibration threshold and the equipment calibration window.

[0057] The type of interference is determined based on the current sensor data and random forest model of each water resource intelligent management and control device, and the corresponding optimization and debugging parameters of the water resource intelligent management and control device are obtained based on the type of interference. The types of interference include water droplet interference and crop shading interference.

[0058] Based on the basic debugging parameters and optimized debugging parameters, complete the full-dimensional debugging of the intelligent water resource management and control equipment.

[0059] The method provided in this embodiment is executed by a terminal device. The terminal device receives data sent by the electromagnetic flowmeter and issues a calibration threshold and calibration window to enable the electromagnetic flowmeter to complete accurate measurement and stable communication.

[0060] Specifically, the intelligent water resource management device in this embodiment is an electromagnetic flow meter. The method provided in this embodiment is applicable to agricultural scenarios. For example, for irregular fields with multi-branched pipe networks, multiple electromagnetic flow meters are installed near the pipe networks.

[0061] The basic function of an electromagnetic flowmeter is to collect the water flow velocity and flow rate in a pipe network. The electromagnetic flowmeter connects to the terminal device via Bluetooth. Bluetooth connection between the terminal device and the electromagnetic flowmeter is suitable for distributed, low-power, and short-range transmission in agricultural settings. Since electromagnetic flowmeters are dispersed in agricultural areas, wired connections would face problems such as complex wiring, high maintenance costs, and susceptibility to damage from agricultural machinery. Cellular networks would increase power consumption and communication costs, and signal coverage blind spots may exist in remote agricultural areas. Bluetooth technology, however, features low power consumption, stable short-range transmission, and no additional communication fees. It meets the short-range data interaction needs between the electromagnetic flowmeter and the terminal device while extending the flowmeter's battery life, making it suitable for long-term unattended operation scenarios in agricultural fields.

[0062] Specifically, when the overall farmland area is large, multiple terminal devices can be set up, each terminal device corresponding to a sub-area. A sub-area includes multiple electromagnetic flowmeters. Each terminal device and its corresponding multiple electromagnetic flowmeters should be within the Bluetooth communication range. This embodiment will only use a sub-area as an example for the following description. In addition, the device calibration threshold and device calibration window in this embodiment need to be updated every preset time to meet the current actual situation. This embodiment does not limit the preset time, and this embodiment only introduces the calculation steps once. These steps can be repeated every time an update is performed.

[0063] Furthermore, in this embodiment, the debugging of the intelligent water resource management and control equipment refers to the process of calculating the equipment calibration threshold and calibration window based on environmental data, identifying interference types, and generating optimized debugging parameters to complete the parameter configuration and operational optimization of the electromagnetic flowmeter. This ultimately achieves accurate monitoring and reliable operation of the water resource management and control equipment in the complex and dynamic environment of farmland. The basic debugging parameters are a set of calibration thresholds and calibration windows for each device, while the optimized debugging parameters include Bluetooth transmission power adjustment values ​​and Mesh network multipath transmission enable / disable flags. Full-dimensional debugging integrates two aspects: firstly, it uses basic debugging parameters to debug the data acquisition aspect of the electromagnetic flowmeter; secondly, it uses optimized debugging parameters to debug the communication aspect of the electromagnetic flowmeter. This embodiment does not strictly limit the order of the two debugging aspects. For example, the step of generating optimized debugging parameters based on the random forest model can be performed once every preset time. This preset time can be the same as or different from the preset time corresponding to updating the equipment calibration threshold and calibration window, but the two are not completely independent. For instance, when the random forest model identifies interference, it can additionally trigger an update of the basic debugging parameters to ensure the validity of the collected data.

[0064] This embodiment achieves precise measurement of irrigation flow by grouping and customizing calibration thresholds and calibration windows. Furthermore, by optimizing the debugging parameters and responding to interference in real time, it ensures the continuity and stability of monitoring data, enabling refined monitoring to be continuously implemented in complex and dynamic farmland environments.

[0065] Furthermore, this embodiment provides a step for obtaining a device calibration threshold and a device calibration window based on the mean of the competitive advantage index, including:

[0066] Based on the average competitive advantage index, each intelligent water resource management and control device is divided into multiple first groups;

[0067] The boundary devices are determined based on the first group, and the first group is updated based on the boundary devices to obtain multiple second groups;

[0068] Based on the second grouping and the particle swarm optimization algorithm, the device calibration threshold corresponding to each second group is obtained;

[0069] Based on the flow characteristic vector and flow fluctuation variance corresponding to the boundary device, the device calibration window corresponding to the boundary device is obtained.

[0070] Specifically, the environmental data corresponding to each electromagnetic flowmeter includes plant height, leaf density, and soil moisture. Among them, plant height can be obtained by the sensor based on the laser ranging principle. For example, the sensor can emit laser pulses towards the top of the crop. By recording the time difference between the laser pulse being emitted and being reflected back from the top of the crop, and combining this with the sensor's own location, the plant height can be calculated.

[0071] Leaf density can be obtained using an infrared sensor. For example, the transmitter in the sensor emits infrared light at an angle towards the crop canopy. When the light encounters the leaves, it undergoes diffuse reflection, and some of the reflected light returns to the receiver. When the leaf density is high, more infrared light is blocked and reflected by the leaves, resulting in stronger light intensity returning to the receiver. When the leaves are sparse, most of the light passes through the gaps between the leaves, resulting in weaker light intensity returning to the receiver. Therefore, based on this principle, a fitting model between emitted light intensity, received light intensity, and leaf density can be obtained through experimental measurement, thus yielding the final leaf density.

[0072] To ensure the accuracy of the fitting model, factors such as crop type can also be considered. The fitting model is a technical means that can be implemented by those skilled in the art, and this embodiment will not elaborate on it. Furthermore, this embodiment does not limit the type of model.

[0073] Soil moisture is used to measure soil water content, which can be obtained using a TDR sensor. For example, when the sensor probe is inserted into the soil, it emits a high-frequency electromagnetic pulse. The propagation speed and attenuation of the pulse in the soil are directly related to the soil's dielectric constant, which is primarily determined by the water content. Furthermore, the high-frequency electromagnetic pulse is essentially an electromagnetic wave. While electromagnetic waves travel at the speed of light in a vacuum, when they propagate through the soil, they are hindered by soil molecules, resulting in a slower propagation speed than the speed of light in a vacuum. The degree of this hindering is reflected in the soil's dielectric constant, from which the relative dielectric constant of the soil can be obtained. ,in This represents the dielectric constant of the soil. This represents the vacuum permittivity, which is a fixed value. Furthermore, it is combined with the formula for the propagation speed of electromagnetic waves in a medium. and the speed of light formula ,in Indicates the magnetic permeability of the soil. Representing the vacuum permeability, for weakly magnetic media such as soil, the soil permeability can be approximated as the vacuum permeability, thus obtaining... Assuming the length of the probe is The propagation time was obtained. By combining the two formulas, we can obtain Therefore, after measuring the pulse propagation time, it can be calculated. The soil moisture content can then be obtained using the Topp formula.

[0074] For each electromagnetic flowmeter, a data sensor for collecting environmental parameters is installed on the corresponding electromagnetic flowmeter. If the number of sensors is usually large and all of them are directly connected to the terminal device, it will exceed the concurrent connection limit of the terminal Bluetooth module. Therefore, in order to reduce the connection pressure on the terminal device, the sensor can first send the collected data to the corresponding electromagnetic flowmeter for relay, and finally the electromagnetic flowmeter sends the data to the terminal device.

[0075] It should be noted that the above sensors can only collect raw data such as time difference, emitted light intensity, received light intensity, and pulse propagation time. The step of calculating environmental data is performed in the terminal device.

[0076] The terminal device then calculates the average competitive advantage index based on the received data.

[0077] Specifically, for each intelligent water resource management device, the corresponding plant height difference rate can be calculated based on its corresponding plant height and the current plant height benchmark. The current plant height benchmark is the average plant height of crops in the current growth cycle, which can be determined based on historical growth data.

[0078] Next, the adjacent control devices of each water resource intelligent control device are determined. Based on the blade density of each water resource intelligent control device and the blade density of its adjacent control devices, the blade density gradient of each water resource intelligent control device can be obtained. The adjacent control devices are water resource intelligent control devices that are less than a preset distance from each water resource intelligent control device. In this embodiment, the preset distance is not limited.

[0079] In one embodiment, the average blade density of adjacent control devices is first calculated, then the average blade density is subtracted from the device's own blade density, and the difference is divided by a preset distance to obtain the blade density gradient corresponding to the device.

[0080] Next, the soil moisture corresponding to each intelligent water resource management device and the soil moisture corresponding to its adjacent management devices are obtained. The variance of soil moisture corresponding to each intelligent water resource management device is calculated. Finally, the plant height difference rate, leaf density gradient and soil moisture variance are normalized and weighted to obtain the mean of the competitive advantage index corresponding to itself. In this embodiment, the weights are not restricted.

[0081] This embodiment constructs a fusion weight matrix by using a competitive advantage index and divides the system into multiple first groups. This ensures that devices belonging to the same group have consistent spatial locations, pipeline affiliations, and crop competition statuses, serving as the basis for subsequent adjustments to device thresholds and device windows.

[0082] Furthermore, such as Figure 2 As shown, this embodiment provides a step for dividing each intelligent water resource management device into multiple first groups based on the average competitive advantage index, including:

[0083] The first iteration operation includes:

[0084] Based on the average competitive advantage index of each intelligent water resource management and control device and its current weight, a fusion weight matrix is ​​obtained;

[0085] Based on the fusion weight matrix and degree matrix, the standardized Laplacian matrix is ​​obtained;

[0086] Spectral decomposition and cluster analysis are performed based on the standardized Laplacian matrix to obtain the current first group;

[0087] Determine whether the current first group has reached the preset termination condition. If not, update the current weight until the current first group reaches the preset termination condition, and output multiple first groups.

[0088] Specifically, within a sub-region, each element in the fusion weight matrix represents the association weight between its corresponding two intelligent water resource management devices. Taking the calculation method of one element in the matrix as an example, firstly, the spatial distance between its corresponding two intelligent water resource management devices is obtained, then the spatial bandwidth is obtained, which is used to control the influence of spatial distance on weight. Next, the pipeline consistency identifier between the two intelligent water resource management devices is obtained. The pipeline consistency identifier is determined according to the pipeline code corresponding to the two intelligent water resource management devices. If the pipeline codes corresponding to the two intelligent water resource management devices are the same, the pipeline consistency identifier is 1; otherwise, it is 0. Finally, the element values ​​corresponding to the two intelligent water resource management devices in the fusion weight matrix are obtained based on the spatial distance, pipeline consistency identifier, and average competitive advantage index.

[0089] For example, for the first The and the first The element values ​​corresponding to the intelligent water resource management and control device are:

[0090]

[0091] in, Indicates the first The and the first The spatial distance between intelligent water resource management and control devices This refers to the spatial bandwidth. This embodiment does not limit its specific value. For example, the average distance between each intelligent water resource management device and its adjacent management devices can be calculated, and then the average of the average distances corresponding to each intelligent water resource management device can be used as the spatial bandwidth. For the first The and the first Each intelligent water resource management and control device corresponds to a pipeline network consistency identifier. and The first The and the first The average competitive advantage index corresponding to each intelligent water resource management and control device. The scaling parameter, representing the similarity of crop types, is used to control the rate at which differences in crop types decay in weight. This embodiment does not impose any restrictions on its specific value; for example, the standard deviation of the mean of all competitive advantage indices can be calculated as this scaling parameter. , and The weight is not limited to a specific value in this embodiment. Since this is the first iteration, the current weight can be obtained by random generation.

[0092] Next, a degree matrix is ​​constructed based on the current fusion weight matrix.

[0093] Specifically, the degree matrix is ​​a diagonal matrix, and the diagonal elements are the sum of the corresponding rows of the fusion weight matrix, reflecting the total association strength between a single device and all other devices. For example, the degree matrix contains the first element... Line number Column elements , This indicates the total number of devices.

[0094] Next, based on the fusion weight matrix Sum-degree matrix The standardized Laplace matrix is ​​obtained. This is used to eliminate the influence of differences in the overall correlation strength of the devices, making the subsequent spectral decomposition results more stable.

[0095] Next, spectral decomposition is performed on the standardized Laplace matrix to find the first K smallest eigenvectors in the standardized Laplace matrix. The first K smallest eigenvectors are then normalized, where K represents the number of the first group. Finally, K first groups are obtained based on the result of the normalization process, and each first group includes at least one intelligent water resource management device.

[0096] Next, for each first group, the number of devices with the same network coding and the total number of devices included are determined. The ratio of the number of devices with the same network coding to the total number of devices is used as the purity of the first group. If the purity of each first group is less than the preset purity, it means that the preset termination condition has been met. Otherwise, the current weight is updated until the current first group reaches the preset termination condition. Multiple first groups are output. In this embodiment, the value of the preset purity is not limited. Those skilled in the art can determine it according to the complexity of the network, for example, the preset purity is set to 0.85.

[0097] This embodiment does not limit the specific method of updating the weights. For example, for each intelligent water resource management device, its corresponding neighboring management devices and their corresponding soil moisture are determined. Then, the element values ​​of each intelligent water resource management device and its corresponding neighboring management devices are determined in the fusion matrix to obtain the predicted soil moisture value corresponding to each intelligent water resource management device. For example, for the first... The intelligent water resource management and control device has a corresponding predicted soil moisture value. ,in Indicates the first The total number of adjacent control devices corresponding to each intelligent water resource management and control device. express The first of the intelligent water resource management and control devices The system calculates the soil moisture corresponding to each intelligent water resource management device. Then, it calculates the absolute error between the predicted soil moisture value and the actual soil moisture corresponding to that device. The actual soil moisture is the soil moisture measured by sensors mentioned earlier. This allows the system to obtain the absolute error for each intelligent water resource management device and calculate the mean of all absolute errors as the average error. Then, it adjusts the weight values ​​according to the ratio of the current average error to the average error calculated in the previous iteration. For example, if the ratio is less than 1, it indicates a smaller current error, and the weight contributing the most can be strengthened according to (1 - ratio). If the ratio is greater than 1, it indicates a larger current error, and the weights can be corrected in the opposite direction according to (ratio - 1). It is crucial to ensure that the sum of the updated weights remains 1. This embodiment adjusts the weights based on the ratio between errors, ensuring that the weights always converge towards the direction of minimizing the predicted error, thus meeting the dynamic needs of agricultural scenarios.

[0098] This embodiment constructs a fusion weight matrix by integrating spatial distance, pipeline coding, and competitive advantage index. Combined with Laplace matrix spectral decomposition, the nonlinear spatial distribution of equipment is mapped to a low-dimensional space, so that the grouping results can accurately match the actual shape of the field and avoid cross-scene mixed grouping problems caused by regular division. At the same time, cluster analysis ensures that the equipment in the same first group is highly matched in terms of spatial location, pipeline affiliation, and crop competitive advantage, providing a clear grouping benchmark for subsequent boundary equipment affiliation determination.

[0099] Furthermore, this embodiment provides a step of updating a first group based on a boundary device to obtain multiple second groups, including:

[0100] Based on the mean of the competitive potential index, the feature vector corresponding to each boundary device is obtained;

[0101] Obtain the feature mean value corresponding to each first group, and calculate the similarity based on the feature vector and the feature mean value;

[0102] The first group is updated based on similarity, resulting in multiple second groups.

[0103] Furthermore, this embodiment provides a step for determining a boundary device based on a first group, including:

[0104] The embedding matrix is ​​obtained from the normalized Laplacian matrix;

[0105] The posterior probability of each intelligent water resource management device is obtained based on the embedding matrix and activation function.

[0106] The boundary device is determined in the first group based on the posterior probability.

[0107] Specifically, after performing eigenvalue decomposition on the Laplacian matrix, the eigenvectors corresponding to the first K smallest eigenvalues ​​are obtained. These eigenvectors are then concatenated row by row to obtain the embedding matrix. Each row in the embedding matrix corresponds to a smart water resource management device, and each column corresponds to a first group. Based on the essence of spectral clustering, which transforms the association relationships between devices (fusion feature matrix) into a similarity graph, the low-dimensional embeddings (eigenvectors) of this graph are found through eigenvalue decomposition of the Laplacian matrix, thereby discovering the clustering structure (first group) in the graph. The eigenvalues ​​of the Laplacian matrix... and eigenvectors satisfy ,when When smaller, there are ,Right now And then disassembled to obtain , Indicates equipment The corresponding feature vector, Indicates equipment The corresponding element in the degree matrix, Indicates equipment The corresponding feature vectors are such that similar devices have similar values ​​in the feature vectors. When the feature vectors are concatenated into an embedding matrix, the matrix represents the low-dimensional coordinates of the devices in the grouping dimension. The values ​​in the matrix represent the tendency of the intelligent water resource management device for the first group corresponding to the value. However, the values ​​in the feature vectors can be any real numbers without range restrictions, so the magnitude of the values ​​has no probabilistic meaning. Therefore, in this embodiment, the values ​​in the feature vectors are converted into the posterior probability corresponding to each intelligent water resource management device according to the activation function (such as Softmax). If there is a posterior probability of multiple first groups for each intelligent water resource management device that is higher than the preset probability, it indicates that the device is a boundary device. In this embodiment, the specific value of the preset probability is not limited. Those skilled in the art can determine it according to the complexity of the farmland scene, such as selecting it between 0.6 and 0.9.

[0108] Then, for each boundary device, its corresponding mean competitive advantage index, plant height difference rate, leaf density gradient, and soil moisture variance are concatenated to obtain its corresponding feature vector. For each first group, the mean of the feature vectors corresponding to all the intelligent water resource management devices included in it is calculated as the mean feature. Then, for each boundary device, the similarity between its corresponding feature vector and each mean feature is calculated, and it is assigned to the first group with the highest similarity, resulting in the updated first group, which is denoted as the second group.

[0109] This embodiment breaks through the limitations of traditional methods that only use physical location to classify equipment affiliation by matching the feature vectors, feature mean, and similarity of the boundary devices. It resolves the fundamental contradiction between the physical boundary and the characteristic boundary between crops and devices, ensuring that devices belonging to the same second group maintain a high degree of consistency in crop competition status and field environment characteristics. This provides a unified benchmark unit for subsequent calibration threshold iterations, avoids threshold mismatch caused by cross-group characteristic mixing, and improves the system's adaptation accuracy to heterogeneous farmland environments.

[0110] Furthermore, such as Figure 3 As shown, this embodiment provides a step for obtaining the device calibration threshold corresponding to each second group based on the second group and the particle swarm optimization algorithm, including:

[0111] The second iteration operation includes obtaining the traffic error rate and packet loss rate corresponding to the current device calibration threshold for each second group, obtaining the fitness function based on the traffic error rate and packet loss rate, updating the current device calibration threshold based on the fitness function, until a preset termination condition is reached, and outputting the device calibration threshold corresponding to each second group.

[0112] Specifically, in the particle swarm optimization algorithm corresponding to this embodiment, a particle represents the parameter combination corresponding to each group, specifically including two types of parameters: Bluetooth transmission power and device calibration threshold. The fitness function is used as a standard to judge the quality of each parameter combination. The fitness function can be calculated using two indicators: traffic error rate and Bluetooth packet loss rate. The smaller the fitness value, the better the performance of the corresponding parameter combination. For example, the traffic error rate is obtained through the traffic measurement value and the actual traffic value. The actual traffic value can be a reference traffic value obtained by manually controlling the irrigation traffic in the early stage of debugging. The Bluetooth packet loss rate is determined based on the number of lost data packets and the total number of transmitted data packets. The fitness function is a weighted value of the traffic error rate and the Bluetooth packet loss rate. This embodiment does not limit the weight.

[0113] In the first iteration, the particle swarm is initialized, generating multiple initial parameter combinations as particles for each second group. Each particle calculates its corresponding fitness value based on the current traffic error and packet loss rate data. In subsequent iterations, each particle references its best-performing parameters (individual optimum) and the best-performing parameters in the entire particle swarm (global optimum), while also considering the parameter inertia of each iteration, gradually adjusting the values ​​of Bluetooth transmission power and device calibration threshold. When the change in fitness value calculated in two consecutive rounds is less than a preset value, it indicates that the preset termination condition has been met, and the device calibration threshold corresponding to each second group is output. The particle swarm algorithm is a technical means that can be implemented by those skilled in the art. This embodiment does not limit the preset value, such as setting it to 0.01.

[0114] This embodiment uses traffic error rate and Bluetooth packet loss rate as dual objectives. A weighted fitness function balances their priorities, thus anchoring the core accuracy requirements of traffic monitoring while also ensuring communication stability. Furthermore, the iterative logic of individual and global optima in the particle swarm optimization algorithm allows parameters to quickly converge to the optimal combination suitable for the current micro-cell scenario, avoiding the problem of getting stuck in local optima with single parameter adjustments. Combined with multiple iterations and stability termination conditions, the adaptability and long-term stability of the parameters are further ensured. The final output is a unique parameter combination for each group, achieving parameter customization for different groups. This allows for precise control of traffic error for each group and reduces the Bluetooth packet loss rate to a lower level. Simultaneously, the terminal automatically completes iteration and parameter distribution, saving the cost of manual zone-by-zone debugging.

[0115] Furthermore, such as Figure 4 As shown, this embodiment provides a step for obtaining the device calibration window corresponding to the boundary device based on the flow feature vector and flow fluctuation variance corresponding to the boundary device, including:

[0116] The irrigation method for boundary equipment is determined based on the flow characteristic vector, flow fluctuation variance, and preset flow characteristic mean.

[0117] The standard deviation of flow fluctuation corresponding to the boundary equipment is obtained based on the irrigation method;

[0118] The device calibration window corresponding to the boundary device is obtained based on the standard deviation of flow fluctuation.

[0119] Specifically, the flow feature vector is obtained based on its own flow characteristics (such as fluctuation frequency and peak percentage). For each boundary device, real-time flow data within a preset time period is collected to calculate its corresponding flow feature vector and flow fluctuation variance. For each irrigation method (such as drip irrigation or flood irrigation), the mean flow feature value can be obtained based on historical irrigation data. By comparing the similarity between its own flow feature and the mean flow feature value corresponding to each irrigation method, the irrigation method corresponding to each boundary device is obtained. Then, the corresponding irrigation method is executed for each boundary device to obtain the standard deviation of flow fluctuation corresponding to the boundary device within a preset time period. If the standard deviation is less than or equal to the preset standard deviation, the current calibration window is maintained; if it is greater than the preset standard deviation, the current calibration window is increased. This embodiment does not limit the preset standard deviation, such as setting it to 0.6 m³ / h. For example, since the boundary device belongs to the transition zone of multiple groups, its corresponding calibration window can be obtained from the device calibration windows of multiple associated first groups. First, the first group with a posterior probability higher than the preset probability is determined. The average weight corresponding to each first group is determined according to the fusion weight matrix, and each average weight is normalized. Then, the average weight of each first group is obtained. The current device calibration window corresponding to each first group is weighted according to the normalized average weight to obtain a weighted window. Then, historical device calibration windows are obtained, and the device calibration windows are smoothed by moving average to obtain a smoothed window. In this embodiment, the value of the smoothing coefficient is not limited. In addition, in order to avoid false triggering, in addition to the standard deviation corresponding to the boundary device being greater than the preset standard deviation, the standard deviation of the traffic fluctuation corresponding to at least one associated first group must also be greater than the preset standard deviation. Finally, the smoothed window is multiplied by (1 + dynamic adjustment coefficient) to obtain the device calibration window. The dynamic adjustment coefficient is determined according to the empirical adjustment coefficient and the number of fluctuation propagation groups. The empirical adjustment coefficient is used to quantify the additional increase that the device calibration window needs to increase for each additional affected first group. In this embodiment, its specific value is not limited, such as being set to 0.2. The number of fluctuation propagation groups is determined according to the number of first groups whose traffic fluctuation standard deviation is greater than the preset standard deviation.

[0120] In this embodiment, the device calibration window specifically refers to the duration of each electromagnetic flow data collection. Its core purpose is to ensure the accuracy of flow calculation by using sufficient sample data. When the flow fluctuates greatly, the data collected in a short period of time is prone to contain extreme fluctuation values. Extending the window is equivalent to increasing the data sample size. More sampled data can make the high and low values ​​of the fluctuation cancel each other out, and the final calculated average value is closer to the actual flow. When the flow fluctuates little, a small amount of data can accurately reflect the true value. Shortening the window can reduce the data collection time and improve the calculation efficiency without affecting the accuracy, and avoid unnecessary resource consumption.

[0121] Furthermore, this embodiment provides a step for determining the type of interference based on the current sensor data of each water resource intelligent management and control device and a random forest model, including:

[0122] Select splitting features for each tree in the random forest model based on the current covariance matrix;

[0123] For each splitting feature, determine whether the current sensor data satisfies the corresponding judgment rule;

[0124] The type of interference is determined based on the judgment result.

[0125] Specifically, in agricultural scenarios, crop shading and moisture interference in farmland can affect the stability of Bluetooth signals, leading to transmission delays, packet loss, or data errors. Therefore, this embodiment identifies the current interference type based on random forest and dynamically optimizes the model through a data-driven approach. This approach can handle interference from the sensor itself and adapt to the dynamic characteristics of Bluetooth transmission.

[0126] Since water droplet interference can cause sudden and large changes in capacitance, and crop shading can not only interfere with the sensor probe, but also block the Bluetooth antenna, resulting in signal attenuation, the trigger time of the irrigation signal can be precisely aligned with the timestamp of capacitance change to avoid misjudgment caused by the asynchrony between the irrigation signal and sensor data. Therefore, the current sensor data in this embodiment includes three features: capacitance, leaf density and irrigation signal.

[0127] Specifically, the random forest model first needs to be pre-trained. An initial covariance matrix is ​​configured to define the correlation strength between three features in the sensor data and the type of interference. For example, the correlation between capacitance change and water droplet interference is set to 0.2, the correlation between leaf density change and crop shading is set to 0.2, and the correlation between irrigation signal and irrigation status is set to 0.1. Then, the above correlation values ​​are normalized to obtain the probability of selecting each feature when the decision tree splits nodes: for example, the probability of selecting capacitance change is 40%, the probability of selecting leaf density change is 40%, and the probability of selecting irrigation signal is 20%. This allows the decision tree to prioritize features with stronger correlation to interference. The above values ​​are only examples, and this embodiment does not limit the initial values.

[0128] Then, historical sensor data can be sampled to generate an independent training set for each decision tree. The training data includes: capacitance change, leaf density change, irrigation signal and corresponding interference type label. Then, split features are randomly selected according to the above probability, and the splitting threshold is set with the maximum information gain as the standard to obtain the judgment rule corresponding to each split feature (such as whether the capacitance change is greater than 0.5μF). Each split feature corresponds to a node in the decision tree.

[0129] In practical applications, real-time sensor data is input into a random forest model. The random forest model comprises multiple decision trees, each of which independently makes its own judgments. The final result is then output through voting. Each tree selects features according to a preset feature selection probability, making judgments layer by layer. For example, the first tree selects capacitance change as its root node with a 40% probability, based on whether the capacitance change is greater than 0.5μF. The input 0.6μF satisfies this condition, so it moves to the right child node. The right child node then selects leaf density change with a 40% probability, based on whether the leaf density change is greater than 0.1. The input 0.12 satisfies this condition, and the final output is crop interference. Similarly, the fourth tree selects leaf density change as its root node with a 40% probability, based on whether the leaf density change is greater than 0.1. The input 0.12 satisfies this condition, so it moves to the right child node. The right child node then selects capacitance change with a 40% probability, based on whether the capacitance change is greater than 0.5μF. The input 0.6μF satisfies this condition, and the final output is crop interference. The 10th tree selects the irrigation signal as its root node with a 20% probability. The judgment rule is whether the irrigation signal equals 1. If the input is 1, it satisfies the condition and enters the right child node. Subsequent trees continue to use the corresponding features for judgment, and finally output water droplet interference. If there are 10 trees in total, after the 10 trees have completed their judgments, the voting result is that 8 trees choose crop interference, 2 trees choose water droplet interference, and no trees choose no interference. The random forest then outputs the final result as crop interference in one go.

[0130] After determining the type of interference, the power can be adjusted to ensure the reliability of data transmission. For example, if the interference is determined to be crop shading, the Bluetooth transmission power is automatically increased from the previously obtained Bluetooth transmission power based on the particle swarm optimization algorithm to the compliant maximum power to enhance signal penetration. At the same time, the Mesh network multipath transmission enable / disable flag is set to 1 (enabled) to trigger Mesh network multipath transmission, forwarding data through two or more adjacent nodes to avoid packet loss caused by single-path shading. If the interference is determined to be water droplet interference, the real-time capacitance change value is multiplied by the water droplet interference power adjustment coefficient to obtain the power increment. The power increment is then added to the Bluetooth transmission power obtained based on the particle swarm optimization algorithm to obtain the adjusted Bluetooth transmission power. The water droplet interference power adjustment coefficient is used to quantify the contribution of capacitance change to the transmission power adjustment. It can be obtained through experimental calibration, such as simulating water droplet interference of different intensities in a controlled environment, recording multiple sets of capacitance change values ​​and corresponding required power increments, and fitting the results through linear regression. The Mesh network multipath transmission enable / disable flag is set to 0 (disabled) to reduce communication overhead and enhance signal penetration by increasing transmission power only.

[0131] Furthermore, the water resource intelligent management and control device method based on short-range transmission provided in this embodiment also includes:

[0132] Obtain the feature mean corresponding to the interference data to obtain the feature deviation vector;

[0133] Calculate the outer product matrix of the eigenvalue deviation vectors;

[0134] Update the current covariance matrix based on the outer product matrix, the current covariance matrix, and the amount of interference data.

[0135] Specifically, when the random forest model identifies a certain interference, it adds the corresponding interference label to the current sensor data, transforming it into interference data. When the interference data accumulates to a sufficient number (e.g., 100), it triggers the update of the covariance matrix. The difference between the feature mean corresponding to the interference data and the initial feature mean is used as the feature bias vector. Then, the feature bias vector is multiplied by its own transpose to obtain the outer product matrix. After that, the initial covariance matrix is ​​obtained and added to the outer product matrix. Finally, the ratio of the sum to the current data volume is used as the updated covariance matrix.

[0136] This embodiment, while preserving the correlation patterns of interference features accumulated from historical training data, incorporates deviation information between a single new interference data point and the historical feature mean. A weighted average is used to achieve lightweight matrix correction, avoiding the resource consumption of discarding historical data and recalculating the entire matrix, while ensuring that the on-site interference features of new data are accurately incorporated into the matrix. Furthermore, the update steps provided in this embodiment can dynamically adapt to changes in on-site interference. Since the feature correlations of farmland interference fluctuate dynamically with the environment, the deviation information from the new data allows the matrix to reflect these changes in real time, making the feature selection probabilities of the subsequent decision tree more closely aligned with the current scenario.

[0137] Furthermore, once the equipment calibration threshold corresponding to the second group is iteratively output through the particle swarm optimization algorithm, and the equipment calibration window based on the standard deviation of flow fluctuations of boundary equipment is calculated, the automated testing and acceptance process is automatically initiated. Before initiation, the intelligent water resource management and control equipment automatically verifies the preconditions, ensuring that the boundary equipment of the second group has been clearly identified, dynamic interference identification and transmission optimization have been completed, and the covariance matrix of the random forest model has been updated. Simultaneously, all electromagnetic flowmeter hardware is operating stably, Bluetooth communication links are unobstructed, and the power supply meets the requirements for continuous operation. Once all the above conditions are met, the terminal equipment automatically switches from calibration mode to acceptance mode and sends an acceptance preparation command to all associated electromagnetic flowmeters via Bluetooth. Each electromagnetic flowmeter suspends non-acceptance-related data acquisition and focuses on monitoring acceptance indicators.

[0138] During the acceptance preparation stage, the terminal device automatically calls the pre-stored acceptance threshold library, which is set based on agricultural precise irrigation standards, Bluetooth transmission industrial specifications, and electromagnetic flowmeter measurement accuracy requirements, and is used to clearly distinguish the qualified ranges of core indicators and non-core indicators. At the same time, the terminal device establishes exclusive acceptance communication links with each electromagnetic flowmeter and portable standard flowmeter. Among them, the terminal device and the electromagnetic flowmeter transmit acceptance data through Bluetooth encryption, and the terminal device and the standard flowmeter synchronously collect standard flow data through Bluetooth or short-distance wired connection to provide a benchmark for the determination of measurement accuracy.

[0139] No manual intervention is required throughout the core indicator collection process, and the data is transmitted and stored in real time to the local database of the terminal device. In terms of calibration accuracy indicators, the terminal device issues acceptance measurement instructions to the electromagnetic flowmeters. Each electromagnetic flowmeter continuously collects multiple groups of real-time flow data based on the current calibration threshold and calibration window. The standard flowmeter synchronously collects the corresponding period data, calculates the absolute error of each group of data and takes the average value as the single-device flow error, and then calculates the average error corresponding to all electromagnetic flowmeters according to the second grouping to obtain the grouped flow error mean value; at the same time, the standard deviation of flow fluctuations during the acceptance measurement period is extracted. If it meets the calibration window adaptation standard, the window is judged to be qualified, and the percentage of qualified electromagnetic flowmeters is counted to obtain the calibration window qualification rate. The transmission stability indicator sends test data packets containing grouping identifiers and timestamps to each group of electromagnetic flowmeters, and counts the percentage of successfully received and returned data packets to obtain the Bluetooth packet loss rate. At the same time, the average value of the time difference between the data packet sending and return is calculated as the Bluetooth transmission delay. The grouping effectiveness indicator extracts the number of devices in the same pipe network and the total number of devices in the grouping for each second grouping from the grouping module log, calculates the purity of each group and takes the average value to obtain the second grouping average purity, and then calls the standard grouping annotation data preset based on the field pipe network layout and crop distribution, compares the actual attribution results of the boundary devices, and counts the percentage of accurate matches as the boundary device attribution accuracy rate. The system reliability indicator extracts the total number of faults and the number of self-healed faults during the entire debugging process from the emergency handling module log, and calculates the percentage of self-healed faults to obtain the fault self-healing success rate.

[0140] After the indicator collection is completed, the terminal device automatically compares the collected data with the acceptance threshold library through the built-in algorithm, and determines the acceptance result according to the preset rules. When all core indicators meet the qualified standards and the number of unqualified non-core indicators does not exceed the preset quantity, it is judged as acceptance qualified; when a small number of core indicators do not meet the standards or some non-core indicators do not meet the standards, it is judged as automatic rectification required; when a large number of core indicators do not meet the standards or a large number of non-core indicators do not meet the standards, it is judged as acceptance unqualified. The determination process is completed efficiently, and the result is stored locally and a determination log containing indicator data, threshold comparison results, and determination basis is generated.

[0141] Feedback is provided through a combination of local and remote methods. When acceptance is successful, the green indicator light stays on and is accompanied by a preset prompt sound. At the same time, an automated acceptance report is uploaded to the water resources management system, including group information, final calibration thresholds, calibration window parameters, measured data of all indicators, and judgment results. When automatic rectification is required, the red indicator light flashes at a preset frequency and is accompanied by a corresponding prompt sound. An acceptance problem list is uploaded, specifying the non-compliant indicators, corresponding standards, and related modules. When acceptance fails, the red indicator light flashes at a high frequency and is accompanied by a warning prompt sound. An acceptance anomaly alarm report is uploaded, marked "manual intervention required" in addition to the problem list, and an alarm is triggered in the management system to notify the operation and maintenance personnel.

[0142] For situations requiring automatic rectification, the terminal device automatically triggers secondary optimization of the corresponding module without manual intervention. If the average group flow error does not meet the standard, the calibration module is restarted, the threshold search range and iteration parameters of the particle swarm optimization algorithm are adjusted, and the customized calibration threshold for the group is recalculated. If the calibration window pass rate does not meet the standard, the window adjustment module is restarted, the flow feature vector and flow fluctuation variance of the boundary devices are re-collected, the irrigation method judgment logic is corrected, and the calibration window is adjusted a second time. If the Bluetooth packet loss rate or transmission delay does not meet the standard, the interference identification module is restarted, the sensor data such as capacitance, leaf density, and irrigation signal are re-collected, the random forest covariance matrix is ​​updated, and the Bluetooth transmission parameters are optimized a second time. If the average purity of the group or the accuracy of the boundary device attribution does not meet the standard, the grouping module is restarted, the weight coefficients are adjusted, the competitive potential index of all electromagnetic flowmeters is recalculated, the fusion weight matrix is ​​updated, and the spectral clustering iteration is executed a second time to optimize the second grouping result. After each rectification is completed, it automatically returns to the acceptance triggering stage and restarts the complete acceptance process until it is judged as acceptable. If it is still in the automatic rectification state after a preset number of rectifications, it is automatically upgraded to acceptance failure and a manual intervention alarm is triggered.

[0143] After successful acceptance, the terminal device sends an acceptance completion command to all electromagnetic flowmeters. Each device switches from acceptance mode to normal operation mode, performing flow measurement, data uploading, and other control functions based on the final calibration threshold and calibration window. Simultaneously, the terminal device encrypts and stores the automated acceptance report, rectification records (if applicable), and all original data for all indicators locally, meeting the preset storage duration requirements. This data is also synchronized to the historical database of the control system, supporting subsequent maintenance queries, problem tracing, and parameter optimization references. Furthermore, the terminal device automatically triggers a low-power strategy, dynamically adjusting Bluetooth transmission power and wake-up intervals according to the irrigation cycle. During non-irrigation periods, power consumption is reduced by optimizing transmission parameters, ensuring the battery life of the solar-powered equipment and preventing interruptions in water resource management.

[0144] Example 2

[0145] This embodiment provides a debugging system for intelligent water resource management and control equipment based on short-range transmission, including:

[0146] The calculation module is used to obtain the average competitive advantage index for each water resource intelligent management and control device based on the corresponding environmental data.

[0147] The calibration module is used to obtain the device calibration threshold and device calibration window based on the mean of the competitive advantage index;

[0148] The interference identification module is used to determine the type of interference based on the current sensor data and random forest model of each water resource intelligent management and control device, and to obtain the corresponding optimization and debugging parameters of the water resource intelligent management and control device based on the type of interference. The types of interference include water droplet interference and crop shading interference.

[0149] The debugging module is used to generate basic debugging parameters based on the equipment calibration threshold and the equipment calibration window; and to complete the full-dimensional debugging of the intelligent water resource management and control equipment based on the basic debugging and optimized debugging parameters.

[0150] This application provides a method and system for debugging intelligent water resource management and control equipment based on short-range transmission. It constructs a fusion weight matrix by using a competitive potential index and divides the system into multiple first groups. This ensures that devices belonging to the same group have consistent spatial locations, pipeline affiliations, and crop competition statuses, which serve as the basis for subsequent adjustments to device thresholds and device windows. For boundary devices, the grouping is determined by similarity, which solves the problem of unclear affiliation of boundary devices due to physical location and cross-domain scenarios. This avoids the ambiguity of traditional static division and further adjusts the calibration threshold based on iterative steps, making the calibration threshold adaptable to the dynamic characteristics of farmland.

[0151] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0152] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0153] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0154] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A debugging method for intelligent water resource management and control equipment based on short-range transmission, characterized in that, include: For each intelligent water resource management and control device, the average competitive advantage index is obtained based on the corresponding environmental data. The device calibration threshold and device calibration window are obtained based on the mean value of the competitive advantage index. Basic debugging parameters are generated based on the equipment calibration threshold and the equipment calibration window. The interference type is determined based on the current sensor data and random forest model of each water resource intelligent management and control device, and the corresponding optimization and debugging parameters of the water resource intelligent management and control device are obtained based on the interference type. The interference type includes water droplet interference and crop shading interference. Based on the basic debugging parameters and the optimized debugging parameters, complete the full-dimensional debugging of the intelligent water resource management and control equipment.

2. The debugging method for a water resource intelligent management and control device based on short-range transmission according to claim 1, characterized in that, The step of obtaining the device calibration threshold and device calibration window based on the average value of the competitive advantage index includes: Based on the average competitive advantage index, each water resource intelligent management and control device is divided into multiple first groups; Based on the first group, the boundary device is determined, and based on the boundary device, the first group is updated to obtain multiple second groups; Based on the second grouping and the particle swarm optimization algorithm, the device calibration threshold corresponding to each second group is obtained; Based on the flow feature vector and flow fluctuation variance corresponding to the boundary device, the device calibration window corresponding to the boundary device is obtained.

3. The debugging method for a water resource intelligent management and control device based on short-range transmission according to claim 2, characterized in that, The method of dividing each intelligent water resource management device into multiple first groups based on the average competitive advantage index includes: The first iteration operation includes: obtaining a fusion weight matrix based on the mean competitive advantage index of each intelligent water resource management device and the current weight; obtaining a standardized Laplace matrix based on the fusion weight matrix and the degree matrix; performing spectral decomposition and cluster analysis based on the standardized Laplace matrix to obtain the current first group; determining whether the current first group has reached a preset termination condition; if not, updating the current weight until the current first group reaches the preset termination condition; and outputting the multiple first groups.

4. The debugging method for a water resource intelligent management and control device based on short-range transmission according to claim 2, characterized in that, The step of updating the first group based on the boundary device to obtain multiple second groups includes: Based on the mean of the competitive potential index, the feature vector corresponding to each boundary device is obtained; Obtain the feature mean value corresponding to each first group, and obtain the similarity based on the feature vector and the feature mean value; The first group is updated based on the similarity, resulting in multiple second groups.

5. The debugging method for a water resource intelligent management and control device based on short-range transmission according to claim 3, characterized in that, The step of determining the boundary device based on the first group includes: The embedding matrix is ​​obtained from the normalized Laplacian matrix; The posterior probability corresponding to each intelligent water resource management and control device is obtained based on the embedding matrix and activation function. Boundary devices are determined in the first group based on posterior probabilities.

6. The debugging method for a water resource intelligent management and control device based on short-range transmission according to claim 2, characterized in that, The step of obtaining the device calibration threshold corresponding to each second group based on the second grouping and the particle swarm optimization algorithm includes: The second iteration operation includes obtaining the traffic error rate and packet loss rate corresponding to the current device calibration threshold for each second group, obtaining a fitness function based on the traffic error rate and packet loss rate, updating the current device calibration threshold based on the fitness function, until a preset termination condition is reached, and outputting the device calibration threshold corresponding to each second group.

7. The debugging method for a water resource intelligent management and control device based on short-range transmission according to claim 2, characterized in that, The step of obtaining the device calibration window corresponding to the boundary device based on the flow feature vector and flow fluctuation variance corresponding to the boundary device includes: The irrigation method of the boundary device is determined based on the flow characteristic vector, the flow fluctuation variance, and the preset flow characteristic mean. The standard deviation of flow fluctuation corresponding to the boundary equipment is obtained based on the irrigation method. The device calibration window corresponding to the boundary device is obtained based on the standard deviation of the flow fluctuation.

8. The debugging method for a water resource intelligent management and control device based on short-range transmission according to claim 1, characterized in that, The step of determining the interference type based on the current sensor data of each water resource intelligent management and control device and the random forest model includes: Select splitting features for each tree in the random forest model based on the current covariance matrix; For each splitting feature, determine whether the current sensor data satisfies the corresponding judgment rule; The type of interference is determined based on the judgment result.

9. The debugging method for a water resource intelligent management and control device based on short-range transmission according to claim 8, characterized in that, Also includes: Obtain the feature mean corresponding to the interference data to obtain the feature deviation vector; Calculate the outer product matrix of the feature deviation vectors; The current covariance matrix is ​​updated based on the outer product matrix, the current covariance matrix, and the amount of interference data.

10. A debugging system for intelligent water resource management and control equipment based on short-range transmission, applied to the debugging method for intelligent water resource management and control equipment based on short-range transmission as described in any one of claims 1-9, characterized in that, include: The calculation module is used to obtain the average competitive advantage index for each water resource intelligent management and control device based on the corresponding environmental data. The calibration module is used to obtain the device calibration threshold and the device calibration window based on the mean value of the competitive advantage index; The interference identification module is used to determine the type of interference based on the current sensor data and random forest model of each water resource intelligent management and control device, and to obtain the corresponding optimization and debugging parameters of the water resource intelligent management and control device based on the type of interference. The types of interference include water droplet interference and crop shading interference. The debugging module is used to generate basic debugging parameters based on the equipment calibration threshold and the equipment calibration window; and to complete the full-dimensional debugging of the intelligent water resource management and control equipment based on the basic debugging parameters and the optimized debugging parameters.

Images