A smart agricultural irrigation regulation and control system based on multi-source data fusion

Abstract

This invention belongs to the field of smart agriculture technology. It discloses a smart agricultural irrigation control system based on multi-source data fusion. The system includes: uniformly dividing a target farmland into multiple unit farmlands; collecting multi-source crop phenotypic data from each unit farmland and performing fusion analysis to identify the crop type and growth stage of each unit farmland; acquiring the soil moisture status of each unit farmland and, combined with the crop type and growth stage, quantitatively assessing the water deficit degree of each unit farmland; acquiring the three-dimensional distribution of crop roots in each unit farmland and, combined with the water deficit degree, dynamically calculating the irrigation hydraulic parameters of each unit farmland; and, based on the irrigation hydraulic parameters, determining the irrigation operation parameters for each unit farmland and precisely controlling the irrigation equipment. This invention enables differentiated and precise irrigation at the unit farmland level, significantly improving water resource utilization efficiency and ensuring that crops receive the most suitable water supply at different growth stages.

Description

Technical Field

[0001] This invention relates to the field of smart agriculture technology, and more specifically, to a smart agricultural irrigation control system based on multi-source data fusion. Background Technology

[0002] With the increasing scarcity of global water resources and the acceleration of agricultural modernization, precision irrigation, as an important means to improve water resource utilization efficiency, is playing an increasingly prominent role in modern agriculture. Traditional agricultural irrigation management mainly relies on single sensor data and fixed irrigation systems for decision-making. Due to the complexity of farmland ecosystems and the spatiotemporal variability of crop water requirements, single data sources and static management models cannot accurately reflect the dynamic interaction between crops, soil, and the environment. Therefore, problems such as low water resource utilization efficiency, poor irrigation uniformity, and inability to adapt to dynamic changes in crop growth needs exist, making it difficult to achieve precise matching between water supply and crop needs. This results in serious water waste and difficulty in guaranteeing crop yield and quality.

[0003] Existing agricultural irrigation technologies suffer from the following shortcomings: Limited data acquisition and fusion: Irrigation systems generally use single-type sensors for monitoring, lacking comprehensive consideration of crop phenotypes and root distribution, making it difficult to achieve "on-demand irrigation"; Inefficient spatial management and water requirement assessment: Existing systems treat the entire farmland as a unified management unit, ignoring the spatial heterogeneity within the farmland, and using fixed crop coefficients to assess water requirements, lacking dynamic responses to crop growth stages and actual water conditions; One-sided irrigation control: Irrigation decisions primarily focus on controlling irrigation volume, lacking the ability to precisely regulate irrigation depth, and failing to achieve stratified irrigation that matches crop root distribution.

[0004] In view of this, the present invention proposes a smart agricultural irrigation control system based on multi-source data fusion to solve the above problems. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: a smart agricultural irrigation control system based on multi-source data fusion, comprising: The data acquisition module is used to evenly divide the target farmland into multiple unit farmlands and collect multi-source crop phenotypic data of each unit farmland in real time. The crop identification module is used to fuse and analyze multi-source crop phenotypic data to accurately identify the crop type and growth stage of each unit of farmland. The moisture monitoring module is used to acquire the soil moisture status of each unit of farmland and perform coupled calculations with crop type and growth stage to quantitatively assess the degree of water deficit in each unit of farmland. The hydraulic calculation module is used to obtain the three-dimensional distribution of crop roots in each unit of farmland, determine the root characteristic parameters of each unit of farmland based on the three-dimensional distribution of crop roots, and dynamically calculate the irrigation hydraulic parameters of each unit of farmland according to the root characteristic parameters and the degree of water deficit. The irrigation control module is used to determine the irrigation operation parameters of each unit of farmland based on the irrigation hydraulic parameters of each unit of farmland and in combination with the pre-built water transport model through numerical solution, and to precisely control the irrigation equipment according to the irrigation operation parameters.

[0006] Furthermore, multi-source crop phenotypic data includes crop image data and multispectral reflectance data; The content of accurately identifying the crop type and growth stage of farmland in each unit includes: The crop image data of each unit of farmland is input into the trained crop recognition model, and the type label of each unit of farmland is output. Based on the type label, the crop type corresponding to each unit of farmland is obtained. The crop image data and type labels of each unit of farmland are integrated to obtain the labeled data of each unit of farmland; the labeled data of each unit of farmland are input into the trained stage recognition model to output the predicted probability set corresponding to each unit of farmland; based on the real-time collected multispectral reflectance data, the stage probability set corresponding to each unit of farmland is calculated in sequence. The predicted probability set includes the predicted probabilities of different growth stages corresponding to crop types, and the stage probability set includes the stage probabilities of different growth stages corresponding to crop types. Different probability weights are set for the predicted probability and the stage probability. The predicted probability and the stage probability are weighted and summed for the same unit farmland and the same growth stage to obtain the comprehensive probability of each unit farmland for different growth stages. The comprehensive probabilities corresponding to the same unit farmland are compared, and the growth stage with the highest comprehensive probability is taken as the growth stage of the corresponding unit farmland.

[0007] Furthermore, the content of the stage probability set corresponding to each unit of farmland is calculated sequentially, including: The multispectral historical data corresponding to each unit of farmland is obtained, and combined with the real-time collected multispectral reflectance data, the spectral variation characteristics of each unit of farmland under different bands are calculated in turn. A set of typical features is pre-defined, which includes typical variation features of different growth stages corresponding to different crop types under different wavelengths; based on the crop type corresponding to each unit of farmland, the typical variation features of each unit of farmland are obtained; based on the spectral variation features and typical variation features of each unit of farmland, the feature similarity of different growth stages corresponding to each unit of farmland under different wavelengths is calculated in turn. Different bands are assigned corresponding band weights in sequence, and the feature similarity of each unit farmland corresponding to the same growth stage is calculated by weighted summation based on the band weights to obtain the overall similarity of each unit farmland corresponding to different growth stages; the overall similarity of each unit farmland corresponding to different growth stages is then processed by Softmax normalization to obtain the stage probability set of each unit farmland.

[0008] Furthermore, the quantitative assessment of the water deficit degree of each unit of farmland includes: Each unit of farmland is divided into multiple soil layers, each corresponding to a depth range. The field water holding capacity and permanent wilting point of each unit of farmland corresponding to different soil layers are obtained. Combined with the soil moisture status of each unit of farmland, the effective water content of each unit of farmland corresponding to different soil layers is calculated sequentially. Based on the crop type and growth stage corresponding to each unit of farmland, the crop coefficient and crop water sensitivity coefficient corresponding to each unit of farmland are obtained from the preset coefficient set; the reference evapotranspiration corresponding to the target farmland is calculated, and combined with the crop coefficient corresponding to each unit of farmland, the actual evapotranspiration of the crop corresponding to each unit of farmland is calculated; based on the effective water content of different soil layers corresponding to each unit of farmland, the water deficit ratio of different soil layers corresponding to each unit of farmland is calculated; the water deficit ratio of different soil layers corresponding to each unit of farmland, the actual evapotranspiration of the crop, and the crop water sensitivity coefficient are multiplied in sequence to obtain the degree of water deficit of different soil layers corresponding to each unit of farmland.

[0009] Furthermore, the root system characteristic parameters include the horizontal expansion index, the root depth index, and the root density index; The content for determining the horizontal expansion index per unit of farmland includes: From the three-dimensional distribution of crop roots, the three-dimensional coordinates and unit length corresponding to each root point are obtained sequentially, and the root point corresponding to the base of the crop stem is marked as the center point; the Euclidean distance between each root point and the center point is calculated sequentially and marked as the root distribution distance; all root distribution distances are sorted from smallest to largest to generate a distance sequence; the number of root points is counted to obtain the number of units; the product of the number of units and the preset effective root coefficient is calculated to obtain the effective number; the distance sequence is then ranked... The root distribution distance at each location is used as a horizontal expansion index; where... This is the effective quantity.

[0010] Furthermore, the determination of the root depth index per unit of farmland includes: Add up the unit lengths corresponding to all root points sequentially to obtain the total root length; sort the depth coordinates in the three-dimensional coordinates corresponding to all root points from smallest to largest to generate a depth sequence; Based on the depth sequence, calculate the cumulative root length corresponding to each root system point in sequence; calculate the ratio of each cumulative root length to the total root system length to obtain the root length percentage corresponding to each root system point; calculate the percentage difference corresponding to each root system point based on the root length percentage and the effective root system coefficient; take the depth coordinate corresponding to the root system point with the smallest percentage difference as the root system depth index. The determination of the root density index per unit area of ​​farmland includes: From the three-dimensional distribution of crop roots, the root points corresponding to each soil layer are obtained sequentially; based on the unit length of the root points corresponding to each soil layer, the regional root length corresponding to each soil layer is calculated; based on the depth range of each soil layer, the soil thickness corresponding to each soil layer is calculated sequentially; based on the soil thickness and unit area corresponding to each soil layer, the regional volume of each soil layer is calculated; based on the regional root length and regional volume corresponding to each soil layer, the root density index of each soil layer is calculated.

[0011] Furthermore, irrigation hydraulic parameters include total irrigation volume and irrigation depth; The dynamic calculation of irrigation hydraulic parameters for each unit of farmland includes: The water deficit degree, soil thickness, and unit area of ​​different soil layers in the same unit of farmland are multiplied sequentially to obtain the stratified irrigation amount for each soil layer in each unit of farmland. The root density index of different soil layers in the same unit of farmland is averaged to obtain the average root density of each unit of farmland. The ratio of the root density index to the average root density of each soil layer in each unit of farmland is calculated to obtain the root factor for each soil layer in each unit of farmland. Based on the root factor, the stratified irrigation amount of different soil layers in each unit of farmland is weighted and summed to obtain the total irrigation amount for each unit of farmland. The root depth index of each unit of farmland is used as the corresponding irrigation depth.

[0012] Furthermore, the steps for determining irrigation operation parameters per unit of farmland include: Step S1: Set according to the total irrigation amount Preliminary operating parameters for the group; Step S2: Based on the unit area and irrigation depth, from Selected from the initial operation parameters of the group Group candidate job parameters, and Group candidate job parameters are divided into Group of candidate parameter sets; Step S3: Select a set of candidate operation parameters from each set of candidate parameters as representative operation parameters, and filter a set of candidate parameters as the optimal parameter set based on the representative operation parameters; Step S4: Perform continuous optimization search on the candidate job parameters in the optimal parameter set to obtain a set containing... The optimal parameter range for the group of candidate job parameters; Step S5: Calculate the irrigation deviation for each group of candidate operation parameters in the optimal parameter range in turn, and select the candidate operation parameter with the smallest irrigation deviation as the irrigation operation parameter for the corresponding unit of farmland.

[0013] Furthermore, in step S1, setting The preliminary operation parameters for the group include: A preset flow rate range is established. A value is randomly selected from this range as the initial flow rate, and the ratio of the total irrigation volume to the initial flow rate is calculated to obtain the initial time. Based on the initial flow rate and initial time, a set of initial operating parameters is set. This process is repeated until a total of [number missing] parameters are set. Preliminary operating parameters for the group; In step S2, the following are selected: The parameters for the group of candidate jobs include: Obtain the saturated and unsaturated hydraulic conductivity, calculate the product of the saturated hydraulic conductivity and the unit area to obtain the upper limit of flow; calculate the lower limit of flow based on the irrigation depth, the unsaturated hydraulic conductivity, and the preset maximum irrigation time; construct a flow range based on the upper and lower limits of flow; compare the preliminary flow in each set of preliminary operating parameters with the flow range in turn; if the preliminary flow is within the flow range, mark the corresponding preliminary operating parameter as a candidate operating parameter.

[0014] Furthermore, in step S4, obtaining the optimal parameter range includes: The candidate job parameters in the optimal parameter set are randomly divided into... For each set of optimizations, the simplex method is used to perform continuous optimization search to obtain the simplex corresponding to each set of optimizations. Each simplex is iteratively optimized sequentially. After each iteration, the provisional optimal parameters for each simplex are obtained, and clustering is performed on the provisional optimal parameters for all simplexes. When a simplex converges, the iterative optimization process for that simplex ends. When all simplexes converge, the provisional optimal parameters for each simplex are obtained, and the irrigation deviation corresponding to each set of provisional optimal parameters is calculated. The simplex with the smallest irrigation deviation is selected. The optimal operating parameters are integrated to form the best parameter range; The clustering guidance for all temporary optimal operation parameters corresponding to the simplex includes: Cluster the temporary optimal operation parameters corresponding to each simplex to obtain multiple parameter clusters; count the number of temporary optimal operation parameters in each parameter cluster and mark them as the number of temporary optimal operation parameters; if the number of temporary optimal operation parameters in a parameter cluster is greater than 1, determine the corresponding nearest operation parameter, and mark the simplex corresponding to the temporary optimal operation parameters that are not the nearest operation parameters as repulsion simplexes, and perform repulsion operations on the centroids corresponding to each repulsion simplex in turn.

[0015] The technical effects and advantages of the intelligent agricultural irrigation control system based on multi-source data fusion proposed in this invention are as follows: By uniformly dividing the target farmland into multiple unit farmlands and collecting multi-source crop phenotypic data such as crop images and multispectral reflectance from each unit farmland in real time, deep learning technology is used to accurately identify the crop type and growth stage of each unit farmland, providing reliable basic data for subsequent precise water demand assessment and stratified irrigation. Based on the dynamic coupling analysis of crop growth needs and soil moisture status, the water deficit degree of each unit farmland is quantitatively assessed, and irrigation depth and total irrigation amount are determined by acquiring three-dimensional distribution data of crop roots, achieving precision irrigation that matches the actual water demand of crops. Numerical simulation is used to predict the spatiotemporal distribution of water in the soil, and a multi-stage screening and improved multi-starting-point simplex method is used for continuous optimization search. A clustering guidance mechanism avoids local optima and ensures that the globally optimal irrigation scheme is found, thereby minimizing irrigation deviation and improving the efficiency and uniformity of irrigation water use. Differentiated precision irrigation at the unit farmland level is achieved, significantly improving water resource utilization efficiency, reducing water waste, and ensuring that crops receive the most suitable water supply at different growth stages, thereby improving crop yield and quality and providing effective technical support for the sustainable development of smart agriculture. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of a smart agricultural irrigation control system based on multi-source data fusion, according to Embodiment 1 of the present invention. Detailed Implementation

[0017] 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.

[0018] Example 1

[0019] Please see Figure 1As shown in the figure, the smart agricultural irrigation control system based on multi-source data fusion described in this embodiment includes a data acquisition module, a crop identification module, a water monitoring module, a hydraulic calculation module, and an irrigation control module; the modules are connected by wired and / or wireless means to realize data transmission between the modules.

[0020] The data acquisition module is used to evenly divide the target farmland into multiple unit farmlands and collect multi-source crop phenotypic data of each unit farmland in real time.

[0021] Based on a preset unit area, the target farmland is uniformly divided using an equal area segmentation algorithm (such as the grid method, Voronoi diagram method, etc.) to obtain multiple unit farmlands with the same area; where the target farmland refers to the farmland area for irrigation regulation, and the unit area is preset by those skilled in the art based on factors such as the overall area of ​​the target farmland, soil heterogeneity, and the precision of irrigation equipment.

[0022] Multi-source crop phenotypic data includes, but is not limited to, crop image data, multispectral reflectance data, etc. Crop image data refers to the appearance images of crops within a unit of farmland, while multispectral reflectance data refers to the reflectance data of crops within a unit of farmland under different wavelengths (such as blue light, green light, red light, etc.). Multi-source crop phenotypic data is acquired through multi-source acquisition devices deployed in each unit of farmland, including RGB cameras and multispectral sensors.

[0023] The crop identification module is used to fuse and analyze multi-source crop phenotypic data to accurately identify the crop type and growth stage of each unit of farmland.

[0024] The content of accurately identifying the crop type and growth stage of farmland in each unit includes: The crop image data of each unit of farmland is input into the trained crop recognition model, which outputs the type label corresponding to each unit of farmland and obtains the crop type corresponding to each unit of farmland based on the type label. The type label is a numerical label corresponding to the crop type, and the numerical label corresponding to different crop types is different. Crop types include rice, wheat, cotton, peanuts, etc. The crop recognition model is a convolutional neural network (CNN) model, an extension of deep neural networks. It mainly includes an input layer, multiple convolutional layers, pooling layers (downsampling layers), fully connected layers, and an output layer. Each convolutional layer consists of multiple convolutional kernels (filters), each sliding with the input data to extract local features. Weight parameters are included in the convolution operation to learn the importance of features. Pooling layers are typically followed by convolutional layers to reduce the size of the feature maps, preserving key information while reducing computational complexity. Fully connected layers follow the convolutional and pooling layers to map high-dimensional features to the output space. In the convolutional and fully connected layers, each neuron typically applies an activation function, introducing non-linearity to enhance the model's ability to express complex image patterns and its generalization capabilities. The specific training process of the crop recognition model includes: Multiple crop image data are collected in advance, and each crop image data is labeled with the corresponding type label. The labeled crop image data is divided into a training set and a test set, with 70% of the crop image data used as the training set and 30% used as the test set. The crop recognition model is trained using the training set and tested using the test set. A preset error threshold is set. When the mean prediction error of all crop image data in the test set is less than the error threshold, the crop recognition model is output. The prediction error is calculated using the average cross-entropy loss, and the error threshold is preset according to the accuracy required by the crop recognition model.

[0025] The crop image data and type labels of each unit of farmland are integrated to obtain the labeled data of each unit of farmland. The labeled data of each unit of farmland is then input into the trained stage recognition model, which outputs the predicted probability set corresponding to each unit of farmland. The predicted probability set includes the predicted probability of different growth stages corresponding to crop types. For example, the growth stages are seedling stage, tillering stage, and heading stage for rice, and seedling stage, budding stage, and boll-forming stage for cotton. The stage recognition model is also a convolutional neural network model, and the specific training process is the same as that of the crop recognition model. The multispectral historical data corresponding to each unit of farmland is obtained, and combined with the real-time collected multispectral reflectance data, the spectral variation characteristics of each unit of farmland under different bands are calculated in turn. A set of typical features is pre-defined, which includes typical variation characteristics of different growth stages of different crop types under different wavelengths. Among them, typical variation characteristics refer to the standard reference pattern formed by the change of reflectance data over time under a specific wavelength for a specific crop type and growth stage, which serves as the benchmark for spectral variation characteristics. The set of typical features is pre-defined by those skilled in the art based on experimental analysis, literature experience, and other factors. Based on the crop type corresponding to each unit of farmland, typical variation characteristics of each unit of farmland are obtained. Based on the spectral variation characteristics and typical variation characteristics of each unit of farmland, the cosine similarity of different growth stages of each unit of farmland under different bands is calculated sequentially and marked as feature similarity. Corresponding band weights are set for different bands sequentially, and the sum of all band weights is 1. The band weights are all pre-set by those skilled in the art based on the actual distinguishing role of different bands for growth stages. Based on the band weights, the feature similarity of each unit of farmland corresponding to the same growth stage is calculated by weighted summation to obtain the overall similarity of each unit of farmland corresponding to different growth stages. The overall similarity of each unit of farmland corresponding to different growth stages is sequentially subjected to Softmax normalization to obtain the stage probability set of each unit of farmland. The stage probability set includes the stage probability of different growth stages. Softmax normalization is an existing technology, and the specific calculation process will not be elaborated here. Different probability weights are set for the predicted probability and the stage probability, and based on the predicted probability set and stage probability set corresponding to each unit of farmland, the predicted probability and stage probability of the same unit of farmland and the same growth stage are weighted and summed to obtain the comprehensive probability of each unit of farmland corresponding to different growth stages; the comprehensive probabilities corresponding to the same unit of farmland are compared, and the growth stage with the highest comprehensive probability is taken as the growth stage of the corresponding unit of farmland; wherein, the probability weights are preset by those skilled in the art according to the actual situation.

[0026] The multispectral historical data includes multispectral pre-data and multispectral preceding data. Multispectral pre-data is the multispectral reflectance data that is one position before the current time (i.e., the time when multispectral reflectance data is collected in real time) at the corresponding time. Multispectral preceding data is the multispectral reflectance data that is two positions before the current time at the corresponding time (i.e., the multispectral reflectance data collected two positions before the current time). Spectral feature variations include first-order features and second-order features; The calculation method for the first-order feature is as follows: the real-time acquired multispectral reflectance data is marked as multispectral real-time data, and the multispectral real-time data is paired with the reflectance data of the same band in the multispectral pre-data to form a reflectance set; for each reflectance set, the pre-data (i.e., the reflectance data corresponding to the multispectral real-time data) is subtracted from the real-time data (i.e., the reflectance data corresponding to the multispectral pre-data) to obtain the first-order feature corresponding to different bands. The calculation method for the second-order features is as follows: the reflectance data in the multispectral preceding data is sequentially added to the reflectance set of the same band to form an extended set; for each extended set, the real-time data is subtracted by twice the preceding data, and then the preceding data (i.e., the reflectance data corresponding to the multispectral preceding data) is added to obtain the second-order features corresponding to different bands.

[0027] The moisture monitoring module is used to obtain the soil moisture status of each unit of farmland and perform coupled calculations with crop type and growth stage to quantitatively assess the degree of water deficit in each unit of farmland.

[0028] The quantitative assessment of the water deficit degree of each unit of farmland includes: Based on the soil type and characteristics of each unit of farmland, those skilled in the art can divide each unit of farmland into multiple soil layers in sequence; each soil layer corresponds to a depth range, and the depth ranges corresponding to each soil layer may be different. The field holding capacity and permanent wilting point of different soil layers in each unit of farmland were obtained. Combined with the soil moisture status of each unit of farmland, the effective water content of different soil layers in each unit of farmland was calculated sequentially. Field holding capacity refers to the maximum water content that the soil can retain after free drainage. Permanent wilting point refers to the critical soil moisture content at which crops cannot absorb sufficient water from the soil and will permanently wilt. Both field holding capacity and permanent wilting point were obtained through field measurements by those skilled in the art. Soil moisture status includes the actual water content corresponding to different soil layers, which is obtained by averaging data collected from multiple capacitive soil moisture sensors deployed in the corresponding soil layers. The expression for effective water content is: ; In the formula, Indicates effective moisture content, Indicates the actual moisture content. Indicates the permanent withering point. Indicates field holding capacity; Based on the crop type and growth stage of each unit of farmland, the crop coefficient and crop water sensitivity coefficient of each unit of farmland are obtained from the preset coefficient set. The reference evapotranspiration of the target farmland is calculated. The reference evapotranspiration is calculated using the FAO Penman-Monteith method. The meteorological data (such as temperature, solar radiation, wind speed, etc.) used in the FAO Penman-Monteith method are obtained from the official website of the local meteorological department. The FAO Penman-Monteith method is an existing technology, and the specific calculation process will not be described in detail here. Calculate the product of the crop coefficient and the reference evapotranspiration for each unit of farmland to obtain the actual evapotranspiration of the crop for each unit of farmland; calculate the water deficit ratio for each unit of farmland corresponding to different soil layers based on the effective water content of each unit of farmland corresponding to different soil layers; multiply the water deficit ratio for each unit of farmland corresponding to different soil layers, the actual evapotranspiration of the crop, and the crop water sensitivity coefficient in sequence to obtain the degree of water deficit for each unit of farmland corresponding to different soil layers.

[0029] The coefficient set includes a crop coefficient set and a sensitivity coefficient set, which are pre-set by those skilled in the art based on crop physiological characteristics and practical experience. The crop coefficient set includes crop coefficients corresponding to different crop types and growth stages. Crop coefficients refer to the correction coefficients for the water requirement characteristics of a specific crop at a specific growth stage, which are used to convert reference evapotranspiration into actual crop evapotranspiration. The sensitivity coefficient set includes crop water sensitivity coefficients corresponding to different crop types and growth stages. Crop water sensitivity coefficients are used to characterize the sensitivity of a specific crop to water stress at a specific growth stage. The larger the value, the more severe the impact of water shortage on crop growth and yield at the corresponding growth stage. The water deficit ratio is equal to the difference between the effective water content and the water deficit ratio.

[0030] The hydraulic calculation module is used to obtain the three-dimensional distribution of crop roots in each unit of farmland, determine the root characteristic parameters of each unit of farmland based on the three-dimensional distribution of crop roots, and dynamically calculate the irrigation hydraulic parameters of each unit of farmland according to the root characteristic parameters and the degree of water deficit.

[0031] The method for obtaining the three-dimensional distribution of crop roots is as follows: multiple transparent observation tubes are pre-buried at an angle within a unit of farmland, and root images are periodically taken along the tube walls using a micro-root tube camera equipped with a special light source; each root image is processed sequentially using professional software (such as RootSnap, WinRHIZO Tron, etc.) to extract the root characteristic parameters (such as root length, diameter, etc.) corresponding to each root image, and combined with the spatial position and tilt angle of the transparent observation tubes, the three-dimensional distribution of crop roots is reconstructed from all root images through spatial coordinate transformation (such as geometric projection transformation, tilt angle correction algorithm, etc.) and statistical interpolation (such as Kriging interpolation, inverse distance weighted interpolation, etc.). Root system characteristic parameters include horizontal expansion index, root depth index, and root density index; the horizontal expansion index refers to the maximum horizontal distance that the crop roots extend from the center of the crop outward; the root depth index refers to the maximum vertical depth of the crop roots; and the root density index refers to the density of the crop roots per unit volume of soil. Irrigation hydraulic parameters include total irrigation volume and irrigation depth; total irrigation volume refers to the total volume of water that needs to be applied to the entire unit of farmland; irrigation depth refers to the maximum soil depth that irrigation water needs to penetrate.

[0032] The content for determining the root system characteristic parameters of a unit farmland includes: From the three-dimensional distribution of crop roots, the three-dimensional coordinates and unit length corresponding to each root point are obtained sequentially; where root point represents the sampling unit of root system in three-dimensional space, and unit length refers to the length of root segment between two adjacent root points of the same root. Mark the root points corresponding to the base of the crop stem (i.e., the part where the crop stem connects to the root system) as the center point; calculate the Euclidean distance between each root point and the center point, and mark it as the root distribution distance; sort all root distribution distances from smallest to largest to generate a distance sequence; count the number of root points to obtain the number of units; calculate the product of the number of units and the preset effective root coefficient to obtain the effective number; and select the root points ranked in the distance sequence... The root distribution distance at each location is used as a horizontal expansion index; in, The effective root coefficient is the number of roots that can be used; the effective root coefficient is preset by those skilled in the art based on actual conditions. In this embodiment, the preferred effective root coefficient is [value missing]. It should be noted that the root distribution distance refers to the planar distance on the horizontal plane. Therefore, when calculating the Euclidean distance, it is calculated based on the horizontal and vertical coordinates of each root point and the center point in the corresponding three-dimensional coordinates.

[0033] The total root length is obtained by summing the unit lengths corresponding to all root points sequentially. The depth coordinates of all root points in the three-dimensional coordinate system are then sorted from smallest to largest to generate a depth sequence. The depth coordinates at the surface are set to 0, and the depth coordinates increase as the roots grow downwards. Based on the depth sequence, the cumulative root length corresponding to each root point is calculated sequentially. The ratio of each cumulative root length to the total root length is calculated to obtain the root length percentage for each root point. The absolute value of the difference between each root length percentage and the effective root coefficient is then calculated to obtain the percentage difference for each root point. The depth coordinate corresponding to the root point with the smallest percentage difference is taken as the root depth index. The method for calculating the cumulative root length is as follows: mark the root point for calculating the cumulative root length as the current point, and add the unit length corresponding to the current point and the unit lengths of all root points in the depth sequence that are ahead of the current point in turn to obtain the cumulative root length corresponding to the current point.

[0034] From the three-dimensional distribution of crop roots, the root points corresponding to each soil layer are obtained sequentially; the unit lengths of the root points corresponding to each soil layer are added sequentially to obtain the regional root length of each soil layer; according to the depth range of each soil layer, the soil thickness corresponding to each soil layer is calculated sequentially; the product of the soil thickness and the unit area of ​​each soil layer is calculated sequentially to obtain the regional volume of each soil layer; the ratio between the regional root length and the regional volume of each soil layer is calculated to obtain the root density index of each soil layer. The method for calculating soil thickness is as follows: calculate the difference between the maximum value and the minimum value of the depth range to obtain the soil thickness.

[0035] The dynamic calculation of irrigation hydraulic parameters for each unit of farmland includes: The water deficit degree, soil thickness, and unit area of ​​different soil layers in the same unit of farmland are multiplied sequentially to obtain the stratified irrigation amount for each soil layer in each unit of farmland. The root density index of different soil layers in the same unit of farmland is averaged to obtain the average root density of each unit of farmland. The ratio of the root density index to the average root density of each soil layer in each unit of farmland is calculated to obtain the root factor for each soil layer in each unit of farmland. Based on the root factor, the stratified irrigation amount of different soil layers in each unit of farmland is weighted and summed to obtain the total irrigation amount for each unit of farmland. It should be noted that in calculating the irrigation amount, the degree of water deficit (e.g., ), soil thickness (e.g.) ) and unit area (e.g. To standardize the units used in irrigation calculations, we can ensure the accuracy of the calculations and obtain usable volume units, which will facilitate the control of subsequent irrigation equipment. The root depth index of each unit of farmland is used as the corresponding irrigation depth.

[0036] The irrigation control module is used to determine the irrigation operation parameters of each unit of farmland based on the irrigation hydraulic parameters of each unit of farmland and in combination with the pre-built water transport model through numerical solution, and to precisely control the irrigation equipment according to the irrigation operation parameters.

[0037] Water transport models are mathematical models that describe the infiltration, distribution and movement of irrigation water in soil profiles. They are used to predict the spatiotemporal distribution changes of water in different soil layers after irrigation. Water transport models take into account factors such as soil properties, soil moisture status, root water uptake, and evaporation, and can simulate the movement of water in both vertical and horizontal directions. The pre-construction process of the water transport model is as follows: Those skilled in the art collect soil column samples (i.e., columnar soil samples) from each unit of farmland and experimentally determine soil hydraulic parameters, such as saturated hydraulic conductivity (i.e., the water permeability of soil in a fully saturated state), unsaturated hydraulic conduction function (i.e., the law governing the change of water permeability of soil in an unsaturated state with water content or suction), and water characteristic curves (i.e., the relationship between soil moisture state and soil suction). Based on capacitive soil moisture sensors installed in each unit of farmland, the water transport process under different irrigation operation parameters is recorded. Using professional software such as HYDRUS, and based on the Richards equation, a mathematical model for each unit of farmland is established. The water transport process under different irrigation operation parameters is imported into the corresponding mathematical model for parameter optimization and model verification, ultimately resulting in a water transport model that can input irrigation hydraulic parameters such as total irrigation volume and irrigation depth, and output the spatiotemporal distribution law of water in the soil.

[0038] Irrigation operation parameters include irrigation flow rate and irrigation time; irrigation flow rate refers to the total volume of water applied to a unit of farmland per unit time; irrigation time refers to the duration of the irrigation operation. The steps for determining irrigation operation parameters for a unit of farmland include: Step S1: Set according to the total irrigation amount Preliminary operating parameters for the group; Step S2: Based on the unit area and irrigation depth, from Selected from the initial operation parameters of the group Group candidate job parameters, and Group candidate job parameters are divided into Group of candidate parameter sets; Step S3: Select a set of candidate operation parameters from each set of candidate parameters as representative operation parameters, and filter a set of candidate parameters as the optimal parameter set based on the representative operation parameters; Step S4: Perform continuous optimization search on the candidate job parameters in the optimal parameter set to obtain a set containing... The optimal parameter range for the group of candidate job parameters; Step S5: Calculate the irrigation deviation for each group of candidate operation parameters in the optimal parameter range in turn, and select the candidate operation parameter with the smallest irrigation deviation as the irrigation operation parameter for the corresponding unit of farmland.

[0039] In step S1 above, set The preliminary operation parameters for the group include: A preset flow range is defined, referring to the feasible range of irrigation flow rates, which is pre-set by those skilled in the art based on the performance of the irrigation equipment. A value is randomly selected from this range as the initial flow rate, and the ratio of the total irrigation volume to the initial flow rate is calculated to obtain the initial time. Based on the initial flow rate and initial time, a set of initial operating parameters is set; this process is repeated until all parameters are set. Preliminary operational parameters for the group. It is an integer greater than 1; In step S2 above, the following are selected: The parameters for the group of candidate jobs include: Saturated hydraulic conductivity and unsaturated hydraulic conductivity are obtained from the soil hydraulic parameters acquired during the pre-construction process of the water transport model; among them, unsaturated hydraulic conductivity is obtained from the unsaturated hydraulic conduction function according to the soil moisture state corresponding to the unit farmland. The upper limit of flow rate is obtained by multiplying the saturated hydraulic conductivity by the unit area; the lower limit of flow rate is calculated based on the irrigation depth, unsaturated hydraulic conductivity, and the preset maximum irrigation time; the expression for the lower limit of flow rate is: ; In the formula, Indicates the lower limit of traffic. Indicates irrigation depth. Indicates unsaturated hydraulic conductivity. This indicates the maximum irrigation time; the maximum irrigation time is preset by those skilled in the art based on the performance of the irrigation equipment and the efficiency of farmland operations. A flow range is constructed based on the upper and lower flow limits. The maximum value of the flow range is the upper flow limit, and the minimum value is the lower flow limit. The initial flow rate in each set of preliminary operation parameters is compared with the flow range in turn. If the initial flow rate is within the flow range (i.e., the initial flow rate is less than or equal to the upper flow limit and greater than or equal to the lower flow limit), the corresponding preliminary operation parameter is marked as a candidate operation parameter. If the initial flow rate is outside the flow range (i.e., the initial flow rate is greater than the upper flow limit or less than the lower flow limit), the corresponding preliminary operation parameter is not marked. in, , Furthermore, this embodiment is preferred. .

[0040] In step S3 above, the selection of representative parameters from the candidate parameter set includes: The average flow rate is calculated by averaging all the initial flow rates in the candidate parameter set; the average time is calculated by averaging the initial times in the candidate parameter set; the average operating parameters are constructed based on the average flow rate and average time; the Euclidean distance between each group of candidate operating parameters and the average operating parameters in the candidate parameter set is calculated sequentially and marked as the parameter deviation; the candidate operating parameter with the smallest parameter deviation is taken as the representative operating parameter. The process of selecting a set of candidate parameters as the optimal set includes: Each set of representative operation parameters is input into the corresponding water transport model to obtain the irrigation effect parameters corresponding to each set of representative operation parameters. Among them, the irrigation effect parameters include the predicted expansion index, the predicted depth, and the predicted irrigation amount corresponding to each soil layer. The predicted expansion index refers to the horizontal expansion index output by the water transport model, the predicted depth refers to the irrigation depth output by the water transport model, and the predicted irrigation amount refers to the layer irrigation amount output by the water transport model. The differences between the horizontal expansion index, irrigation depth, and the corresponding layer irrigation amount for each soil layer, and the corresponding parameters in each group of irrigation effect parameters, are calculated sequentially to obtain the parameter deviation for each group of representative operational parameters. The parameter deviation includes expansion deviation, depth deviation, and irrigation amount deviation for each soil layer. Based on the root factor, the irrigation amount deviations for each soil layer are weighted and summed to obtain the total deviation. A preset weight set is established, including the deviation weights corresponding to expansion deviation, depth deviation, and total deviation. The weight set is pre-set by those skilled in the art based on the degree of influence of expansion deviation, depth deviation, and total deviation on crop growth. The parameter deviations corresponding to each group of representative operation parameters are normalized sequentially to obtain the standard deviation. Based on the deviation weight, the standard deviations corresponding to each group of representative operation parameters are weighted and summed sequentially to obtain the irrigation deviation for each group of representative operation parameters. The set of candidate parameters corresponding to the representative operation parameter with the smallest irrigation deviation is taken as the optimal parameter set.

[0041] In step S4 above, the content of obtaining the optimal parameter range includes: The candidate job parameters in the optimal parameter set are randomly divided into... Group optimization set, , The number of candidate job parameter sets in the optimal parameter set; each optimization set includes three candidate job parameters; the simplex method is used to continuously optimize and search each optimization set to obtain the simplex corresponding to each optimization set; the simplex is a geometric shape formed by the three candidate job parameters in the corresponding optimization set; the simplex method is an existing technology, and the specific process will not be described in detail here. Each simplex is iteratively optimized sequentially. After each iteration, the provisional optimal operating parameters for each simplex are obtained, and clustering is performed on all provisional optimal operating parameters corresponding to the simplexes. The provisional optimal operating parameters are the candidate operating parameters with the smallest irrigation deviation among the candidate operating parameters corresponding to the simplexes. When a simplex converges, the iterative optimization process for that simplex ends. When all simplexes converge, the provisional optimal operating parameters for each simplex are obtained, and the irrigation deviation corresponding to each group of provisional optimal operating parameters is calculated. The simplex with the smallest irrigation deviation is then selected. The optimal operating parameters are integrated to form the best parameter range.

[0042] The clustering guidance for all temporary optimal operation parameters corresponding to the simplex includes: Clustering is performed on the temporarily optimal operation parameters corresponding to each simplex to obtain multiple parameter clusters. The DBSCAN clustering method is used, which is an existing technology; its specific process is not detailed here. The number of temporarily optimal operation parameters in each parameter cluster is counted and marked as the number of temporarily optimal parameters. If the number of temporarily optimal parameters in a parameter cluster is greater than 1, the nearest operation parameter is determined, and the simplexes corresponding to temporarily optimal operation parameters that are not the nearest operation parameter are marked as repulsive simplexes. Repulsion operations are then performed on the centroids of each repulsive simplex. If the number of temporarily optimal parameters in a parameter cluster is equal to 1, no operation is performed. The nearest operation parameter is determined as follows: the mean of the temporarily optimal operation parameters in the parameter cluster is calculated to obtain the central operation parameter; the Euclidean distance between each group of temporarily optimal operation parameters and the central operation parameter is calculated and marked as the parameter distance; the temporarily optimal operation parameter with the smallest parameter distance is taken as the nearest operation parameter. The expression for the exclusion operation is: ; In the formula, This represents the centroid after the repulsion operation. Indicates the centroid before the rejection operation. Indicates the strength of repulsion. Indicates the parameters of the most recent job. This represents the Euclidean distance between the centroid before the rejection operation and the nearest operational parameter. This represents the zero-prevention constant, used to prevent the denominator from being zero; both the repulsion strength and the zero-prevention constant are preset by those skilled in the art based on the actual situation. The centroid is calculated as follows: the candidate operation parameter with the largest irrigation deviation in the simplex is marked as the worst operation parameter, and the mean value of all candidate operation parameters in the simplex other than the worst operation parameter is processed to obtain the centroid corresponding to the simplex.

[0043] The criteria for determining whether a simplex curve converges include: Perform geometric convergence and functional convergence tests on the simplex separately; if geometric convergence or functional convergence is satisfied, the simplex is considered convergent; if neither geometric convergence nor functional convergence is satisfied, the simplex is considered non-convergent. The geometric convergence determination process is as follows: calculate the Euclidean distance between each group of candidate operation parameters that are not temporary optimal operation parameters in the simplex and the temporary optimal operation parameter, and mark them as geometric distances; mark the largest geometric distance as the maximum distance and compare it with the preset geometric convergence threshold; if the maximum distance is greater than or equal to the geometric convergence threshold, it is determined that geometric convergence is not satisfied; if the maximum distance is less than the geometric convergence threshold, it is determined that geometric convergence is satisfied. The process of determining function convergence is as follows: Calculate the irrigation deviation for each group of candidate operation parameters in the simplex and mark it as the judgment deviation; mark the judgment deviation corresponding to the temporarily optimal operation parameter as the baseline deviation; subtract the baseline deviation from the judgment deviation corresponding to each group of candidate operation parameters that are not temporarily optimal operation parameters, and take the absolute value to obtain the deviation difference; mark the largest deviation difference as the maximum difference and compare it with the preset function convergence threshold; if the maximum difference is greater than or equal to the function convergence threshold, it is determined that the function does not converge; if the maximum difference is less than the function convergence threshold, it is determined that the function convergence is satisfied. It should be noted that both the geometric convergence threshold and the function convergence threshold are preset by those skilled in the art based on the variable range, function magnitude, and actual accuracy requirements.

[0044] This embodiment divides the target farmland into multiple unit farmlands and collects crop images, multispectral reflectance, and other multi-source crop phenotypic data from each unit farmland in real time. Combined with deep learning technology, it accurately identifies the crop type and growth stage of each unit farmland, providing reliable foundational data for subsequent precise water requirement assessment and stratified irrigation. Based on the dynamic coupling analysis of crop growth needs and soil moisture status, it quantitatively assesses the water deficit degree of each unit farmland and determines irrigation depth and total irrigation amount by acquiring three-dimensional root distribution data, achieving precise irrigation that matches the actual water requirements of the crops. Numerical simulation predicts the spatiotemporal distribution of water in the soil, and a multi-stage screening and improved multi-starting-point simplex method is used for continuous optimization search. A clustering guidance mechanism avoids local optima, ensuring the finding of the globally optimal irrigation scheme, thereby minimizing irrigation deviation and improving the efficiency and uniformity of irrigation water use. This achieves differentiated precision irrigation at the unit farmland level, significantly improving water resource utilization efficiency, reducing water waste, and ensuring that crops receive the most suitable water supply at different growth stages, thereby improving crop yield and quality and providing effective technical support for the sustainable development of smart agriculture.

[0045] Example 2

[0046] This application also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memories store computer-readable code, which, when executed by the one or more processors, can perform a smart agricultural irrigation control system based on multi-source data fusion as described above.

[0047] The methods or systems according to the embodiments of this application can also be implemented using the architecture of the electronic device shown in this application. The electronic device may include a bus, one or more CPUs, ROM, RAM, a communication port connected to a network, input / output, a hard disk, etc. The storage device in the electronic device, such as a ROM or hard disk, may store a smart agricultural irrigation control system based on multi-source data fusion provided in this application. Furthermore, the electronic device may also include a user interface. Of course, the architecture shown in this application is merely exemplary; when implementing different devices, one or more components in the electronic device shown in this application may be omitted according to actual needs.

[0048] Example 3

[0049] One embodiment of this application discloses a computer-readable storage medium. The computer-readable storage medium stores computer-readable instructions. When the computer-readable instructions are executed by a processor, they can perform a smart agricultural irrigation control system based on multi-source data fusion according to an embodiment of this application, as described with reference to the above figures. The storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0050] Furthermore, according to embodiments of this application, the processes described above with reference to the schematic diagrams can be implemented as computer software programs. For example, this application provides a non-transitory machine-readable storage medium storing machine-readable instructions that can be executed by a processor to perform instructions corresponding to the method steps provided in this application, such as a smart agricultural irrigation control system based on multi-source data fusion. When this computer program is executed by a central processing unit (CPU), it performs the functions defined in the method of this application.

[0051] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0052] All formulas in this manual are dimensionless and calculated numerically. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0053] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A smart agricultural irrigation control system based on multi-source data fusion, characterized in that, include: The data acquisition module is used to evenly divide the target farmland into multiple unit farmlands and collect multi-source crop phenotypic data of each unit farmland in real time. The crop identification module is used to fuse and analyze multi-source crop phenotypic data to accurately identify the crop type and growth stage of each unit of farmland. The moisture monitoring module is used to acquire the soil moisture status of each unit of farmland and perform coupled calculations with crop type and growth stage to quantitatively assess the degree of water deficit in each unit of farmland. The hydraulic calculation module is used to obtain the three-dimensional distribution of crop roots in each unit of farmland, determine the root characteristic parameters of each unit of farmland based on the three-dimensional distribution of crop roots, and dynamically calculate the irrigation hydraulic parameters of each unit of farmland according to the root characteristic parameters and the degree of water deficit. The irrigation control module is used to determine the irrigation operation parameters of each unit of farmland based on the irrigation hydraulic parameters of each unit of farmland and in combination with the pre-built water transport model through numerical solution, and to precisely control the irrigation equipment according to the irrigation operation parameters.

2. The intelligent agricultural irrigation control system based on multi-source data fusion according to claim 1, characterized in that, Multi-source crop phenotypic data includes crop image data and multispectral reflectance data; The content of accurately identifying the crop type and growth stage of farmland in each unit includes: The crop image data of each unit of farmland is input into the trained crop recognition model, and the type label of each unit of farmland is output. Based on the type label, the crop type corresponding to each unit of farmland is obtained. The crop image data and type labels of each unit of farmland are integrated to obtain the labeled data of each unit of farmland; the labeled data of each unit of farmland are input into the trained stage recognition model to output the predicted probability set corresponding to each unit of farmland; based on the real-time collected multispectral reflectance data, the stage probability set corresponding to each unit of farmland is calculated in sequence. The predicted probability set includes the predicted probabilities of different growth stages corresponding to crop types, and the stage probability set includes the stage probabilities of different growth stages corresponding to crop types. Different probability weights are set for the predicted probability and the stage probability. The predicted probability and the stage probability are weighted and summed for the same unit farmland and the same growth stage to obtain the comprehensive probability of each unit farmland for different growth stages. The comprehensive probabilities corresponding to the same unit farmland are compared, and the growth stage with the highest comprehensive probability is taken as the growth stage of the corresponding unit farmland.

3. The intelligent agricultural irrigation control system based on multi-source data fusion according to claim 2, characterized in that, The content of the stage probability set corresponding to each unit of farmland is calculated sequentially, including: The multispectral historical data corresponding to each unit of farmland is obtained, and combined with the real-time collected multispectral reflectance data, the spectral variation characteristics of each unit of farmland under different bands are calculated in turn. A set of typical features is pre-defined, which includes typical variation features of different growth stages corresponding to different crop types under different wavelengths; based on the crop type corresponding to each unit of farmland, the typical variation features of each unit of farmland are obtained; based on the spectral variation features and typical variation features of each unit of farmland, the feature similarity of different growth stages corresponding to each unit of farmland under different wavelengths is calculated in turn. Different bands are assigned corresponding band weights in sequence, and the feature similarity of each unit farmland corresponding to the same growth stage is calculated by weighted summation based on the band weights to obtain the overall similarity of each unit farmland corresponding to different growth stages; the overall similarity of each unit farmland corresponding to different growth stages is then processed by Softmax normalization to obtain the stage probability set of each unit farmland.

4. The intelligent agricultural irrigation control system based on multi-source data fusion according to claim 3, characterized in that, The quantitative assessment of the water deficit degree of each unit of farmland includes: Each unit of farmland is divided into multiple soil layers, each corresponding to a depth range. The field water holding capacity and permanent wilting point of each unit of farmland corresponding to different soil layers are obtained. Combined with the soil moisture status of each unit of farmland, the effective water content of each unit of farmland corresponding to different soil layers is calculated sequentially. Based on the crop type and growth stage corresponding to each unit of farmland, the crop coefficient and crop water sensitivity coefficient corresponding to each unit of farmland are obtained from the preset coefficient set; the reference evapotranspiration corresponding to the target farmland is calculated, and combined with the crop coefficient corresponding to each unit of farmland, the actual evapotranspiration of the crop corresponding to each unit of farmland is calculated; based on the effective water content of different soil layers corresponding to each unit of farmland, the water deficit ratio of different soil layers corresponding to each unit of farmland is calculated; the water deficit ratio of different soil layers corresponding to each unit of farmland, the actual evapotranspiration of the crop, and the crop water sensitivity coefficient are multiplied in sequence to obtain the degree of water deficit of different soil layers corresponding to each unit of farmland.

5. A smart agricultural irrigation control system based on multi-source data fusion according to claim 4, characterized in that, Root system characteristic parameters include horizontal expansion index, root depth index, and root density index; The content for determining the horizontal expansion index per unit of farmland includes: From the three-dimensional distribution of crop roots, the three-dimensional coordinates and unit length corresponding to each root point are obtained sequentially, and the root point corresponding to the base of the crop stem is marked as the center point; the Euclidean distance between each root point and the center point is calculated sequentially and marked as the root distribution distance; all root distribution distances are sorted from smallest to largest to generate a distance sequence; the number of root points is counted to obtain the number of units; the product of the number of units and the preset effective root coefficient is calculated to obtain the effective number; the distance sequence is then ranked... The root distribution distance at each location is used as a horizontal expansion index; where... This is the effective quantity.

6. A smart agricultural irrigation control system based on multi-source data fusion according to claim 5, characterized in that, The content for determining the root depth index per unit of farmland includes: Add up the unit lengths corresponding to all root points sequentially to obtain the total root length; sort the depth coordinates in the three-dimensional coordinates corresponding to all root points from smallest to largest to generate a depth sequence; Based on the depth sequence, calculate the cumulative root length corresponding to each root system point in sequence; calculate the ratio of each cumulative root length to the total root system length to obtain the root length percentage corresponding to each root system point; calculate the percentage difference corresponding to each root system point based on the root length percentage and the effective root system coefficient; take the depth coordinate corresponding to the root system point with the smallest percentage difference as the root system depth index. The determination of the root density index per unit area of ​​farmland includes: From the three-dimensional distribution of crop roots, the root points corresponding to each soil layer are obtained sequentially; based on the unit length of the root points corresponding to each soil layer, the regional root length corresponding to each soil layer is calculated; based on the depth range of each soil layer, the soil thickness corresponding to each soil layer is calculated sequentially; based on the soil thickness and unit area corresponding to each soil layer, the regional volume of each soil layer is calculated; based on the regional root length and regional volume corresponding to each soil layer, the root density index of each soil layer is calculated.

7. A smart agricultural irrigation control system based on multi-source data fusion according to claim 6, characterized in that, Irrigation hydraulic parameters include total irrigation volume and irrigation depth; The dynamic calculation of irrigation hydraulic parameters for each unit of farmland includes: The water deficit degree, soil thickness, and unit area of ​​different soil layers in the same unit of farmland are multiplied sequentially to obtain the stratified irrigation amount for each soil layer in each unit of farmland. The root density index of different soil layers in the same unit of farmland is averaged to obtain the average root density of each unit of farmland. The ratio of the root density index to the average root density of each soil layer in each unit of farmland is calculated to obtain the root factor for each soil layer in each unit of farmland. Based on the root factor, the stratified irrigation amount of different soil layers in each unit of farmland is weighted and summed to obtain the total irrigation amount for each unit of farmland. The root depth index of each unit of farmland is used as the corresponding irrigation depth.

8. A smart agricultural irrigation control system based on multi-source data fusion according to claim 7, characterized in that, The steps for determining irrigation operation parameters for a unit of farmland include: Step S1: Set according to the total irrigation amount Preliminary operating parameters for the group; Step S2: Based on the unit area and irrigation depth, from Selected from the initial operation parameters of the group Group candidate job parameters, and Group candidate job parameters are divided into Group of candidate parameter sets; Step S3: Select a set of candidate operation parameters from each set of candidate parameters as representative operation parameters, and filter a set of candidate parameters as the optimal parameter set based on the representative operation parameters; Step S4: Perform continuous optimization search on the candidate job parameters in the optimal parameter set to obtain a set containing... The optimal parameter range for the group of candidate job parameters; Step S5: Calculate the irrigation deviation for each group of candidate operation parameters in the optimal parameter range in turn, and select the candidate operation parameter with the smallest irrigation deviation as the irrigation operation parameter for the corresponding unit of farmland.

9. A smart agricultural irrigation control system based on multi-source data fusion according to claim 8, characterized in that, In step S1, set The preliminary operation parameters for the group include: A preset flow rate range is established. A value is randomly selected from this range as the initial flow rate, and the ratio of the total irrigation volume to the initial flow rate is calculated to obtain the initial time. Based on the initial flow rate and initial time, a set of initial operating parameters is set. This process is repeated until a total of [number missing] parameters are set. Preliminary operating parameters for the group; In step S2, the following are selected: The parameters for the group of candidate jobs include: Obtain the saturated and unsaturated hydraulic conductivity, calculate the product of the saturated hydraulic conductivity and the unit area to obtain the upper limit of flow; calculate the lower limit of flow based on the irrigation depth, the unsaturated hydraulic conductivity, and the preset maximum irrigation time; construct a flow range based on the upper and lower limits of flow; compare the preliminary flow in each set of preliminary operating parameters with the flow range in turn; if the preliminary flow is within the flow range, mark the corresponding preliminary operating parameter as a candidate operating parameter.

10. A smart agricultural irrigation control system based on multi-source data fusion according to claim 9, characterized in that, In step S4, the content of obtaining the optimal parameter range includes: The candidate job parameters in the optimal parameter set are randomly divided into... For each set of optimizations, the simplex method is used to perform continuous optimization search to obtain the simplex corresponding to each set of optimizations. Each simplex is iteratively optimized sequentially. After each iteration, the provisional optimal parameters for each simplex are obtained, and clustering is performed on the provisional optimal parameters for all simplexes. When a simplex converges, the iterative optimization process for that simplex ends. When all simplexes converge, the provisional optimal parameters for each simplex are obtained, and the irrigation deviation corresponding to each set of provisional optimal parameters is calculated. The simplex with the smallest irrigation deviation is selected. The optimal operating parameters are integrated to form the best parameter range; The clustering guidance for all temporary optimal operation parameters corresponding to the simplex includes: Cluster the temporary optimal operation parameters corresponding to each simplex to obtain multiple parameter clusters; count the number of temporary optimal operation parameters in each parameter cluster and mark them as the number of temporary optimal operation parameters; if the number of temporary optimal operation parameters in a parameter cluster is greater than 1, determine the corresponding nearest operation parameter, and mark the simplex corresponding to the temporary optimal operation parameters that are not the nearest operation parameters as repulsion simplexes, and perform repulsion operations on the centroids corresponding to each repulsion simplex in turn.

Images