Intelligent water and fertilizer integrated control method and device, electronic equipment and storage medium

Abstract

The invention relates to an intelligent water and fertilizer integrated control method and device, electronic equipment and a storage medium, and belongs to the technical field of data acquisition and processing. The method comprises the following steps: firstly, collecting various reaction data reflecting the water-fertilizer amount to form a dependent variable sequence and various influence data influencing the water-fertilizer amount to form an independent variable sequence; secondly, performing feature analysis on point pairs formed between change key points in the independent variable sequence and the dependent variable sequence to determine a target point pair sequence capable of accurately representing the lagging influence relation of the independent variable on the dependent variable, so as to accurately represent the lagging degree of the influence of the independent variable on the dependent variable and the real-time change condition of the lagging degree; therefore, the optimal time window length when the independent variable predicts the dependent variables at different moments is determined, the accuracy of water and fertilizer prediction is improved, and finally the fertilization control precision of water and fertilizer is improved.

Description

Technical Field

[0001] This application relates to the field of data acquisition and processing technology, and in particular to an intelligent integrated water and fertilizer control method and device, electronic equipment, and storage medium. Background Technology

[0002] Integrated water and fertilizer management technology mixes soluble fertilizers into irrigation water in a specific ratio to synergistically deliver water and nutrients, aiming to simultaneously meet the water and fertilizer requirements of crops and improve resource utilization efficiency. At different stages of crop growth, especially during rapid growth periods or under extreme climatic conditions, the rates of change in water metabolism and nutrient absorption can vary significantly. To achieve precise water and fertilizer control, existing technologies often use neural network algorithms, such as Long Short-Term Memory (LSTM) networks, to predict subsequent water and fertilizer reserves based on multi-source data influencing water and fertilizer. This allows for the formulation of water and fertilizer control strategies by combining the actual water and fertilizer requirements corresponding to the crop's growth stage with the predicted water and fertilizer reserves, ensuring healthy crop growth and improving resource utilization efficiency.

[0003] However, in practical applications, factors such as soil temperature and wind speed may have different lag effects on water and fertilizer inventory. Existing models often use a fixed window length to acquire multi-source data on various factors affecting water and fertilizer, making it difficult to effectively capture the time dependence between different data sources and water and fertilizer inventory. Especially when the lag effects are large and variable, setting a fixed window length can lead to poor prediction accuracy due to improper window selection or untimely updates. Therefore, how to dynamically optimize the window length based on the lag effects of each data source and the time characteristics of the prediction requirements has become a key technical challenge to improve the accuracy of water and fertilizer prediction. Summary of the Invention

[0004] In view of this, embodiments of this application provide an intelligent integrated water and fertilizer control method and device, electronic device, and storage medium to solve the technical problem of insufficient control accuracy of water and fertilizer in current water and fertilizer control strategies.

[0005] Firstly, a method for intelligent integrated water and fertilizer control is provided, the method comprising: Sample multiple data affecting water and fertilizer storage and record each data sequence as an independent variable sequence. Sample each data sequence reflecting water and fertilizer storage and record each data sequence as a dependent variable sequence. Determine the key points of change based on the local change characteristics of each sampling point in any sequence. Each key point in any sequence of independent variables is paired with each key point in any sequence of dependent variables that is sampled after the key point in any sequence of independent variables, thus forming a key point pair set for each key point in any sequence of independent variables. Starting from the first key point in any sequence of independent variables, each key point pair in the current key point pair set is used as a base pair. In the key point pair set of the next key point in any sequence of independent variables, the key point pair with the highest similarity in local change similarity and temporal difference between the two key points in the base pair is determined as a matching pair. The matching pair is used as a new base pair, and the process of determining the matching pair is repeated until all key points in any sequence of independent variables are traversed, thus obtaining a matching pair sequence for each key point pair. The sampling time of the key point in any sequence of dependent variables in the matching pair is after the sampling time of the key point in any sequence of independent variables in the base pair. The credibility of the matching point pair sequence of any key point pair is determined by the similarity between any two adjacent key point pairs in the matching point pair sequence of any key point pair and the local change similarity between two changing key points within each key point pair. The matching point pair sequence corresponding to the key point pair sequence with the highest credibility in the key point pair sequence of the first changing key point is taken as the target point pair sequence. The time window length for predicting the reflection data corresponding to any dependent variable sequence based on the influence data corresponding to any independent variable sequence is determined by the target point pair sequence. Fertilization control is then carried out after predicting the water and fertilizer inventory based on the time window length.

[0006] Secondly, a smart water and fertilizer integrated control device is provided, the device comprising: The data key point determination module is used to sample various impact data affecting water and fertilizer storage and record each impact data sequence as an independent variable sequence, sample each response data reflecting water and fertilizer storage and record each response data sequence as a dependent variable sequence, and determine the key change points based on the local change characteristics of each sampling point in any sequence; The data key point matching module is used to form key point pairs by pairing any changing key point in any independent variable sequence with each changing key point in any dependent variable sequence after the sampling time of the changing key point, thus obtaining a set of key point pairs for the changing key point. Starting from the first changing key point in the independent variable sequence, using any key point pair in the current set of key point pairs as the base pair, the module determines the key point pair with the highest similarity in local change similarity and temporal difference between the two changing key points within the base pair in the set of key point pairs for the next changing key point as the matching pair. The matching pair is used as the new base pair, and the process of determining the matching pair is repeated until all changing key points in the independent variable sequence are traversed, resulting in a sequence of matching pairs for the given key point pairs. The sampling time of the changing key point in the matching pair belonging to the changing key point in the dependent variable sequence is after the sampling time of the changing key point in the base pair belonging to the changing key point in the independent variable sequence. The water and fertilizer prediction and fertilization module is used to determine the credibility of the matching point pair sequence of any key point pair based on the similarity between any two adjacent key point pairs in the matching point pair sequence of any key point pair and the local change similarity between two changing key points within each key point pair. The matching point pair sequence corresponding to the key point pair sequence with the highest credibility in the key point pair sequence of the first changing key point is taken as the target point pair sequence. The target point pair sequence is used to determine the time window length when predicting the reflection data corresponding to any dependent variable sequence based on the influence data corresponding to any independent variable sequence. Fertilization control is then performed after determining the predicted water and fertilizer inventory based on the time window length.

[0007] Thirdly, embodiments of this application provide an electronic device, which includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the intelligent water and fertilizer integration control method as described in the first aspect.

[0008] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the intelligent water and fertilizer integration control method as described in the first aspect.

[0009] The advantages of this application compared to the prior art are: This application collects various response data on water and fertilizer amounts to form a dependent variable sequence and various data affecting water and fertilizer amounts to form an independent variable sequence. Then, it performs feature analysis on the point pairs formed between the key points of change in the independent variable sequence and the dependent variable sequence to determine the target point pair sequence that can accurately represent the lagged influence of the independent variable on the dependent variable. This accurately represents the degree of lag in the influence of the independent variable on the dependent variable and the real-time changes in this lag degree. In this way, it determines the optimal time window length for predicting the dependent variable at different times, improves the accuracy of water and fertilizer prediction, and ultimately improves the precision of water and fertilizer application control. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a schematic diagram of the application environment of an intelligent water and fertilizer integration control method provided in Embodiment 1 of this application; Figure 2 This is a flowchart illustrating an intelligent water and fertilizer integration control method provided in Embodiment 2 of this application; Figure 3 This is a schematic diagram of key changes in the independent and dependent variables provided in Embodiment 2 of this application; Figure 4 This is a schematic diagram of the structure of an intelligent water and fertilizer integrated control device provided in Embodiment 3 of this application; Figure 5 This is a schematic diagram of the structure of an electronic device provided in Embodiment 4 of this application. Detailed Implementation

[0012] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0013] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0014] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0015] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0016] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0017] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0018] It should be understood that the sequence number of each step in the following embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0019] To illustrate the technical solution of this application, specific embodiments are described below.

[0020] Embodiment 1 of this application provides an intelligent integrated water and fertilizer control method, which is applied in, for example... Figure 1In this application environment, the client and server communicate with each other. Clients include, but are not limited to, handheld computers, desktop computers, laptops, ultra-mobile personal computers (UMPCs), netbooks, cloud terminal devices, and personal digital assistants (PDAs). The server can be implemented using a standalone server or a server cluster consisting of multiple servers. Sensing devices are deployed in the crop growth environment to collect various data affecting water and fertilizer levels, as well as various response data reflecting water and fertilizer levels.

[0021] See Figure 2 This is a flowchart illustrating an intelligent water and fertilizer integration control method provided in Embodiment 2 of this application. This control method can be applied to... Figure 1 The server-side components connect to relevant databases and rule bases to retrieve data. These electronic devices can also connect to clients to collect data and control commands sent by users, and connect to sensing devices to acquire various types of data.

[0022] See Figure 2 This is a flowchart illustrating an intelligent water and fertilizer integration control method provided in Embodiment 2 of this application. Figure 2 As shown, the control method may include the following steps: Step S201: Sample various impact data affecting water and fertilizer storage and record each impact data sequence as an independent variable sequence; sample each response data reflecting water and fertilizer storage and record each response data sequence as a dependent variable sequence; determine the key change points based on the local change characteristics of each sampling point in any sequence.

[0023] To achieve accurate prediction of water and fertilizer levels, it is necessary to sample time-series data related to water and fertilizer using sensing devices composed of sensors and monitoring equipment. Water and fertilizer levels are reflected in two aspects: water content, represented by soil moisture data in the crop growth environment, and fertilizer content, represented by the concentration data of fertilizer elements (such as nitrogen, phosphorus, and potassium) in the soil in the crop growth environment. The corresponding soil moisture data and soil fertilizer element (such as nitrogen, phosphorus, and potassium) concentration data constitute the reaction data reflecting water and fertilizer levels. Since reaction data reflects water and fertilizer levels, it is a dependent variable in water and fertilizer prediction. Correspondingly, various reaction data are arranged into a reaction data sequence according to their data type during the sampling process and recorded as the dependent variable sequence.

[0024] Meanwhile, water and fertilizer reserves are affected by a variety of data, including ambient temperature, soil temperature, wind speed, ambient humidity, light intensity, and meteorological data such as precipitation. These data will affect the crop growth environment and thus influence the above-mentioned response data, ultimately affecting water and fertilizer reserves. Therefore, they are considered data that affect water and fertilizer reserves and are independent variables in water and fertilizer prediction. Correspondingly, various influencing data form an influencing data sequence according to their data type during the sampling process and are recorded as an independent variable sequence.

[0025] In addition, crop growth status data can be collected using image sensors, and the crop growth stage can be obtained through image recognition methods. This information can then be used to compare the required water and fertilizer amounts for each crop growth stage after the predicted water and fertilizer levels have been determined, so that fertilizer control can be implemented. Obtaining crop growth stages through image recognition methods is an existing technology and will not be elaborated on here.

[0026] In the process of intelligent water and fertilizer control, the accuracy of water and fertilizer demand prediction directly affects the rationality of irrigation and fertilization strategies. If the prediction deviation is large, it can easily lead to excessive or insufficient water and fertilizer application, causing problems such as root hypoxia, soil salinization, and water stress, which in turn affect crop growth and yield.

[0027] Existing methods typically utilize models such as Long Short-Term Memory (LSTM) networks to predict water and fertilizer levels from multi-source time-series data on meteorology, soil, and crop growth. However, these input variables generally exhibit a lag effect with respect to changes in water and fertilizer levels; that is, the impact is not immediately apparent but rather gradually transmitted over time. For example, an increase in rainfall requires several hours to days of infiltration and diffusion to significantly alter soil moisture, and increased evaporation caused by rising temperatures usually only manifests as an increase in water demand after a certain period. The effects of fertilization on soil conductivity and crop nutrient uptake are even more likely to lag by several days.

[0028] Because the lag effects and variation characteristics of different variables vary, the window length of different data sources should be dynamically adjusted according to their lag characteristics during water and fertilizer demand forecasting. However, existing LSTM models typically use a fixed window length or preset window lengths based on human experience, which has significant limitations. Specifically, a window that is too short may not cover the entire lag period, leading to information loss and underfitting; while a window that is too long will introduce too much historical data, increasing model complexity, computational burden, and potentially causing overfitting, thus affecting the model's prediction accuracy. Therefore, fixed window lengths or untimely window updates cannot effectively adapt to the different lag effects of various input variables and their dynamic changes, thus affecting the accuracy of water and fertilizer forecasting.

[0029] To address this issue, this embodiment aims to design an adaptive window mechanism that can automatically adjust the window length based on the differences in the lag characteristics of the influence of each input variable on water and fertilizer content in the input model. This will accurately capture the true impact of variables on water and fertilizer demand, improve prediction accuracy, and enhance the robustness of the model.

[0030] Therefore, it is necessary to first find the key points in the sequences of independent and dependent variables that have the ability to effectively represent their respective changes. On the one hand, this will reduce the amount of computation and improve computational efficiency, and on the other hand, improve the accuracy of determining the constraint relationship between independent and dependent variables.

[0031] To ensure the consistency and reliability of the multi-source time series data, i.e., the independent variable series and the dependent variable series mentioned above, a unified preprocessing process is first performed on each time series data. Gaussian filtering is used to reduce sensor noise while preserving the main trends. Simultaneously, spline interpolation is used to align the time of data with inconsistent sampling frequencies, ensuring uniformity across the data series on the time scale. Furthermore, these data are standardized, such as through normalization, to avoid errors caused by different data measurement scales. This step effectively reduces errors caused by noise interference and frequency differences, and standardizes the data, laying a stable data foundation for subsequent variable lag analysis and window adaptive mechanisms.

[0032] After smoothing and aligning the time-series data, local maxima in each sequence are identified using a peak detection algorithm, and the interval between adjacent peaks is calculated. This interval reflects the natural variation cycle of the sequence. Based on the natural variation cycle of the sequence, the analysis time for local feature analysis of the sampling points in the sequence is determined. Then, based on the determined analysis time, the magnitude of the local change at each sampling point in the sequence and the stability of this local change process are determined, thereby determining its likelihood as a "key point of change." This allows for the selection of moments that truly represent significant changes in the variable. In other words, key points of change are determined based on the local change characteristics of each sampling point in any sequence, including: Determine the peak interval between any two adjacent peak points in any sequence, and use half of the minimum peak interval as the analysis time for any sequence; The coefficient of variation of the second-order difference sequence of the data segment to the left of any sampling point in any sequence under the analysis time is calculated and denoted as the first coefficient of variation. The coefficient of variation of the second-order difference sequence of the data segment to the right of any sampling point under the analysis time is calculated and denoted as the second coefficient of variation. The mean slope of each sampling point in the data segment to the left of any sampling point under the analysis time is calculated and denoted as the first mean slope. The mean slope of each sampling point in the data segment to the right of any sampling point under the analysis time is calculated and denoted as the second mean slope. The absolute value of the difference between the mean of the first slope and the mean of the second slope is recorded as the abrupt change value. The mean of the first coefficient of variation and the second coefficient of variation is recorded as the stability of change value. The criticality of change of any sampling point is determined based on the abrupt change value and the stability of change value. The criticality of change is directly proportional to the abrupt change value and inversely proportional to the stability of change value. Sampling points whose change criticality is greater than that of the adjacent sampling points on the left and right sides are recorded as initial screening critical points. All the initial screening critical points are clustered according to the change criticality with a target cluster size of 2. Each initial screening critical point in the cluster with a larger mean change criticality is taken as the change critical point in any sequence.

[0033] The identified key points can be used to align the key nodes of change between different variables or different sequences, thereby realizing the lag relationship between the independent variable sequence and the dependent variable sequence regarding water and fertilizer inventory based on the key nodes of change, providing a direct basis for setting the adaptive window length.

[0034] Furthermore, as a preferred embodiment, the criticality of the change is:

[0035] in, This indicates the criticality of the change at the i-th sampling point in the current sequence. This represents the average slope of all sampling points in the data segment to the left of the i-th sampling point, within the analysis duration. This represents the average slope of all sampling points in the data segment under the analysis duration to the right of the i-th sampling point. This indicates taking the absolute value, and e represents the natural constant. Let represent the coefficient of variation of the second-order difference sequence of the data segment to the left of the i-th sampling point under the analysis duration. Let represent the coefficient of variation of the second-order difference sequence of the data segment to the right of the i-th sampling point, under the analysis duration. The coefficient of variation of the second-order difference sequence of the data segment indicates the stability of its changes. If the stability of the changes on both sides of the sampling point is high, and the absolute value of the difference between the mean slopes on both sides of the sampling point is large, that is, if the degree of local mutation at the sampling point is high, then it is likely that the data point is a key point of change. The absolute value of the difference between the mean slope of each data point to the left and the mean slope of each data point to the right of the i-th sampling point in the sequence being analyzed, with the i-th sampling point as the center, is used to represent the degree of difference in the trend of change on both sides of the i-th sampling point, thereby characterizing the degree of local mutation at the sampling point. This represents the mean of the coefficients of variation of the second-order difference sequences of the data points on the left and right sides of the i-th sampling point in the analyzed sequence within its corresponding analysis range, centered at i. The smaller this value, the better. The larger the value, the more stable the change process on both sides of the data point is. However, if the change trend on both sides of the data point is greater, that is, the degree of abrupt change is greater, then the data point is more likely to be a key point of change.

[0036] The above method yields the key change score for any sampling point in any sequence. Then, iterate through each sampling point in the sequence sequentially. If a sampling point satisfies... Greater than its two adjacent sampling points If the value is high, then that location is more likely to be a key point of change, and then it is selected as the initial key point.

[0037] This allows us to obtain initial screening key points in any sequence. However, while these key points represent local variations, the criticality of these variations only reflects local characteristics. Therefore, it is necessary to compare all initial screening key points in the sequence across the entire sequence. For any sequence, the initial screening key points are then sorted according to their corresponding... The values ​​are sorted sequentially, either ascending or descending; this embodiment uses descending order. Then, the sorted initial screening key points of any sequence are clustered using the K-means clustering algorithm, where K is 2, resulting in two clusters. These two clusters represent the initial screening key points with relatively strong and weak changes in the sequence. Further selection of clusters is then performed. Clusters with larger mean values ​​represent key initial screening points with relatively strong changes in the sequence, and these are used as key change points in the current sequence for subsequent analysis.

[0038] Step S202: For any key point of change in any independent variable sequence, construct key point pairs with each key point of change in any dependent variable sequence after the sampling time of that key point of change, thus obtaining a set of key point pairs for that key point of change. Starting from the first key point of change in any independent variable sequence, use any key point pair in the current set of key point pairs as a base pair. In the set of key point pairs for the next key point of change, determine the key point pair with the highest similarity in local change similarity and temporal difference between the two key points of change within the base pair as a matching pair. Repeat the process of determining matching pairs using the matching pair as a new base pair until all key points of change in any independent variable sequence are traversed, thus obtaining a sequence of matching pairs for any key point pair. The sampling time of the key point of change in any dependent variable sequence in the matching pair is after the sampling time of the key point of change in any independent variable sequence in the base pair.

[0039] The above methods yield key points of change for any sequence. When obtaining the relationship between any independent variable sequence and dependent variable sequence, for example, analyzing the lagged effect of changes in the independent variable (environmental temperature) on the dependent variable (soil moisture), the two sequences should first be time-series aligned. Since the influence of the independent variable on the dependent variable is lagged, a key point of change in the independent variable sequence will only form a lagged effect relationship with each key point of change in the dependent variable sequence in the time period following the sampling time of that key point of change. Therefore, the key point of change in the independent variable and each key point of change in the dependent variable in the time period following the sampling time of that key point of change in the independent variable are sequentially paired to form a key point pair, which serves as the key point pair set for that arbitrary key point of change in the independent variable. This set is used to complete the subsequent lag relationship analysis between the independent and dependent variable sequences. A schematic diagram of each key point of change in the independent and dependent variables is shown below. Figure 3 As shown, the horizontal axis represents the time series, and the vertical fluctuations represent the magnitude of the values ​​of the independent variable and the dependent variable at each sampling point, respectively.

[0040] like Figure 3 As shown, if we analyze the lag relationship between the independent variable target keypoint 1 and the dependent variable, we select independent variable keypoint 1 and any keypoint in the dependent variable located after the sampling time of keypoint 1 for matching analysis. That is, keypoint 1 is paired with keypoint a, keypoint b, keypoint c, keypoint d, and keypoint e respectively to form keypoint pairs, resulting in the set of keypoint pairs for keypoint 1 as follows: When analyzing the lag relationship between the independent variable target keypoint 2 and the dependent variable, any keypoint in the dependent variable after the sampling time of independent variable keypoint 2 is selected for matching analysis. That is, keypoint 2 is paired with keypoints b, c, d, and e respectively to form keypoint pairs, resulting in the following set of keypoint pairs for keypoint 2: Similarly, the set of key point pairs for independent variable key point 3 is: The set of keypoint pairs for independent variable keypoint 4 is as follows: Matching analysis is performed between key points of the independent variable and any key point of the dependent variable after that point in time. This is because changes in the dependent variable are lagged; therefore, only key points of the independent variable and any key point of the dependent variable after that point in time are likely related to changes in the dependent variable at that time. Matching analysis is also performed between key points of the independent variable and any key point of the dependent variable after that point in time—that is, all key points. This is because changes in the dependent variable may be caused by other independent variables, which may also have key points of change at that time. Therefore, in cases of uncertainty, full matching is necessary to improve the comprehensiveness of the lag analysis.

[0041] It is easy to understand that the lagged effect of the independent variable on the dependent variable always exists. Although the degree of lag varies at different times, it remains generally stable. Therefore, for any keypoint pair in the keypoint pair set of keypoint 1, there must exist a unique keypoint pair in the keypoint pair set of the next keypoint 2 that is most similar to that keypoint pair in both lag time and impact characteristics. Thus, the most similar keypoint pairs can be found sequentially in the keypoint pair sets corresponding to keypoint 2, keypoint 3, and so on up to keypoint 4, forming a sequence of matching keypoint pairs corresponding to any of the aforementioned keypoint pairs. The lag time and impact characteristics can be reflected by the similarity and temporal differences in local changes between the two changing keypoints within a keypoint pair, including: The key point in the current key point pair that belongs to the independent variable sequence is recorded as the previous key point, and the key point in the current key point pair that belongs to the dependent variable sequence is recorded as the next key point. The data segment between the previous key point and the next adjacent key point in its independent variable sequence is recorded as the first local sequence, and the data segment between the next key point and the next adjacent key point in its dependent variable sequence is recorded as the second local sequence. After aligning the starting sampling points of the first local sequence and the second local sequence, the Spearman rank correlation coefficient between the overlapping parts of the first local sequence and the second local sequence is determined. Based on the Spearman rank correlation coefficient, the local change similarity between the two changing key points inside the current key point is determined. Calculate the time difference between the previous key point and the next key point with respect to the sampling time, and determine the time difference between the current key point and the two changing key points inside it based on the time difference.

[0042] For example, in determining the matching relationship between the independent variable sequence and the dependent variable sequence, key points are used to... For example, the process of determining the similarity and temporal differences in local changes between two internal key points of this key point is as follows: Keypoint 1 and its adjacent target keypoint 2 are selected as segmentation points. The local sequence corresponding to keypoint 1 is extracted and denoted as the first local sequence. Similarly, the local sequence corresponding to keypoint a is extracted using keypoints a and b as segmentation points and denoted as the second local sequence. Keypoint 1 and keypoint a are then aligned. The intersection of the first and second local sequences is selected as the analysis segment, and the Spearman rank correlation coefficient between the first and second local sequences in the intersection is calculated. This Spearman rank correlation coefficient ranges from -1 to 1, where 1 represents perfect positive correlation, -1 represents perfect negative correlation, and 0 represents no correlation. It measures the relationship between the ranks of two variables without requiring a linear relationship. The Spearman rank correlation coefficient is used to determine the similarity of local changes between keypoint 1 and keypoint a. Simultaneously, the time difference between the sampling time of keypoint 1 and keypoint a is calculated to determine the temporal difference between these two changing keypoints.

[0043] By performing the above operations, the similarity of local changes and temporal differences between the two changing keypoints within each keypoint pair in each keypoint pair set can be determined.

[0044] Because it's uncertain which keypoint pair in each set is the most accurate, but it's easy to understand that the lagged effect of the independent variable on the dependent variable is constant and remains stable, we can make hypothetical inferences. Assume the target keypoint 1 corresponds to the following set of keypoint pairs: middle To determine the keypoint pairs that most accurately represent the matching relationship between the independent and dependent variables, the set of keypoint pairs corresponding to keypoint 2 is... In the middle, there will inevitably be a point that corresponds to the key point. The most accurate matching relationship is that of keypoint pairs. Since the influence of independent variables and lag characteristics usually result in a certain degree of similarity within a short timeframe, keypoint 2 corresponds to a keypoint pair in the keypoint pair set. When the similarity is at its maximum, it means that the key point 2 corresponds to a key point pair in the set of key point pairs, which is likely an accurate match.

[0045] It is important to emphasize that, because the lag in the influence relationship between the independent and dependent variables always exists, the key point pair set in key point 2, and the key point pair... The keypoint pair with the highest similarity, that is, the keypoint pair that conforms to the characteristics of the true lagged influence of the dependent variable on the independent variable, can only be the keypoint pair formed by each dependent variable keypoint after the sampling time of keypoint a and keypoint 2. Or refer to further... Figure 3 For example, in the set of keypoint pairs for keypoint 2, there are keypoint pairs... The key pair with the highest similarity, that is, the key pair that meets the characteristics of the real lagged effect of the dependent variable on the independent variable, can only occur in key pairs. Or the key points are correct. It is impossible for it to originate from key points. Or the key points are correct. The reason lies in the key points Or the key points are correct. The sampling time of key points b or c in the change of the dependent variable sequence is not at a key point pair. The sampling time is after the key point c of the change in the dependent variable sequence.

[0046] Determining the similarity includes: Using any key point pair in the set of key point pairs of the next change key point as the comparison pair of the base point pair, calculate the normalized value of the absolute value of the difference between the local change similarity between the two change key points in the base point pair and the local change similarity between the two change key points in the comparison pair, and denot it as the first difference degree. The normalized value of the absolute value of the difference between the temporal difference between the two key points of change within the base point pair and the temporal difference between the two key points of change within the comparison point pair is denoted as the second difference degree. The similarity between the base point pair and the comparison point pair is determined based on the first difference degree and the second difference degree, and the similarity degree is inversely proportional to both the first difference degree and the second difference degree.

[0047] Furthermore, as a preferred embodiment, the similarity is:

[0048] in, This represents the similarity between the r-th keypoint pair in the set of keypoint pairs for the u-th keypoint in the sequence of independent variables and the j-th keypoint pair in the set of keypoint pairs for the v-th keypoint in the sequence of independent variables. `norm` represents the normalization function, such as linear normalization, norm normalization, etc. The Spearman rank correlation coefficient represents the local variation similarity between two changing keypoints within the set of keypoint pairs of the u-th keypoint in the sequence of independent variables. This represents the local variation similarity between two changing keypoints within the set of keypoint pairs containing the v-th keypoint in the sequence of independent variables. This represents the temporal difference between two changing keypoints within the r-th keypoint pair in the set of keypoint pairs of the u-th keypoint in the sequence of independent variables. This represents the temporal difference between two changing keypoints within the set of keypoint pairs of the v-th keypoint in the sequence of independent variables. This represents the difference between the local change similarity between two changing keypoints within the j-th keypoint pair in the set of keypoint pairs for the v-th keypoint in the independent variable sequence and the local change similarity between two changing keypoints within the r-th keypoint pair in the set of keypoint pairs for the u-th keypoint in the independent variable sequence. The smaller the value of this difference, the greater the similarity between the two keypoint pairs; similarly, This indicates the magnitude of the temporal difference between these two keypoint pairs. A smaller value indicates a higher similarity between the two keypoint pairs. Therefore, The larger the value, the greater the similarity between the two keypoint pairs.

[0049] Through the above, it is possible to use Assuming a base pair of reference points, the set of point pairs corresponding to key point 2. In the process of determining each key point pair and key point pair The similarity between them is used to select the set of point pairs corresponding to key point 2. Middle and key points The keypoint pair with the highest similarity between them is used as the matching point pair, and if it is... So, based on the assumption Yes, it is accurate. Since the influence of independent variables on dependent variables and their lag times often exhibit similarities and continuity, then in this case... That should also be the most accurate, followed by For reference, that is, as a new base pair, the operation of finding the most similar matching pair is repeated in the keypoint pair set of the next keypoint 3. This operation is repeated iteratively until the keypoint pair sets corresponding to all keypoints in the independent variable sequence are traversed. This will give us the keypoint pairs selected from the keypoint pair set of keypoint 1. The corresponding sequence of matching point pairs, which represents the sequence of matching point pairs assumed to be keypoint pairs. This is the most accurate representation of the relationship between the independent and dependent variables when the condition is accurate.

[0050] For the set of key point pairs of independent variable key point 1 Each key point pair will correspond to a matching point pair sequence. Therefore, for the currently analyzed independent variable sequence and dependent variable sequence, a set of matching point pair sequences will be obtained.

[0051] Step S203: Determine the credibility of the matching point pair sequence of any key point pair based on the similarity between any two adjacent key point pairs in the matching point pair sequence of any key point pair and the local change similarity between two changing key points within each key point pair. Take the matching point pair sequence corresponding to the key point pair sequence with the highest credibility in the key point pair set of the first changing key points as the target point pair sequence. Determine the time window length for predicting the reflection data corresponding to any dependent variable sequence based on the influence data corresponding to any independent variable sequence using the target point pair sequence. After determining the predicted water and fertilizer inventory based on the time window length, perform fertilization control.

[0052] Since the set of keypoint pairs for keypoint 1 is unknown. Which key point truly and accurately represents the influence relationship between the independent and dependent variables? Therefore, it's easy to understand that within the set of matched point pairs obtained from the preceding steps, there will only be one optimal matched point pair sequence that accurately represents the relationship between the independent and dependent variables. Thus, it's necessary to determine the reliability of each matched point pair sequence in the set to select and determine the optimal matched point pair sequence.

[0053] Since the optimal matching point pair sequence can accurately characterize the influence relationship between the independent variable and the dependent variable, and the influence relationship between the independent variable and the dependent variable is relatively stable as mentioned above, the reliability of the matching point pair sequence can be determined by measuring the consistency of the similarity of all key point pairs within each matching point pair sequence and the magnitude of the similarity of local changes between two changing key points within each key point pair. This includes: In the matching point pair sequence of any key point pair, select any key point pair as the target key point pair. Calculate the similarity between the target key point pair and its two adjacent key point pairs in the matching point pair sequence of any key point pair. Subtract the two similarities and take the absolute value as the local stability of the similarity of the target key point pair. Calculate the sum of the local stability of the similarity of all key point pairs in the matching point pair sequence of any key point pair as the first sub-confidence of the matching point pair sequence of any key point pair. Calculate the mean of the local change similarity between two changing key points within all key point pairs in the matching point pair sequence of any key point pair, and use it as the second sub-credibility of the matching point pair sequence of any key point pair. The credibility of the matching point pair sequence of any key point pair is determined based on the first sub-credibility and the second sub-credibility, wherein the credibility is inversely proportional to the first sub-credibility and directly proportional to the second sub-credibility.

[0054] Furthermore, as a preferred option, the confidence level is:

[0055] in, This represents the confidence level of the s-th matching point pair in the sequence. This represents an exponential function with base e. This represents the number of matching pairs in the sequence of the s-th matching pair, and norm represents the normalization function, such as linear normalization, norm normalization, etc. This represents the similarity between the (x-1)th keypoint pair and the xth keypoint pair in the s-th matching point pair sequence. This represents the similarity between the x-th keypoint pair and the (x+1)-th keypoint pair in the s-th matching point pair sequence. This indicates taking the absolute value. This represents the mean absolute value of the local variation similarity between the two changing key points within each key point pair in the s-th matching point pair sequence, which is also the mean absolute value of the Spearman rank correlation coefficient between the two changing key points within each key point pair in the s-th matching point pair sequence. The larger this value is, the stronger the overall correlation of the current s-th matching sequence, and the higher the credibility coefficient of the sequence. This represents the sum of the absolute differences in similarity between any two adjacent keypoint pairs in the s-th matching point pair sequence. The smaller this value, the higher the reliability and credibility of the current matching point pair sequence as a correct matching point pair sequence.

[0056] This allows us to determine the confidence level of each matching point pair sequence in the set. Selecting the matching point pair sequence with the highest confidence level as the target point pair sequence allows us to determine the time window length for predicting the reflected data corresponding to any dependent variable sequence based on the influence data corresponding to any independent variable sequence, including: Based on the sampling time difference between two changing key points within each key point pair in the target point pair sequence, determine the lag time corresponding to each key point pair in the target point pair sequence; based on the lag time corresponding to each key point pair in the target point pair sequence, determine the lag time sequence between the influence data corresponding to any independent variable sequence and the reflection data corresponding to any dependent variable sequence. When predicting the data corresponding to the dependent variable sequence after the target time using the impact data corresponding to the impact data corresponding to any independent variable sequence, the time window length for selecting the impact data corresponding to any independent variable sequence is determined according to the lag time series. The time window length is equal to the product of the sum of a constant 1 and a window adjustment parameter multiplied by the lag time corresponding to the target time in the lag time series. The window adjustment parameter is the normalized value of the negative of the maximum confidence value of the matching point pair sequence in the key point pair set of the first change key point.

[0057] Furthermore, as a preferred embodiment, the length of the time window is:

[0058] in, This refers to the length of the time window used when selecting the data corresponding to the influence data of any independent variable sequence z to predict the data of any dependent variable sequence after the target time n, based on the influence data corresponding to any independent variable sequence z. This represents the time difference between the target time n of any dependent variable sequence on the lag time series between the influencing data corresponding to any independent variable sequence z and the reflecting data corresponding to any dependent variable sequence. In other words, it represents the time difference between the time value of any independent variable sequence and the time value of any dependent variable sequence when the time value of the dependent variable sequence is n on the lag time series. Represents the normalization function. This represents the maximum confidence level of the matching point pairs in the set of matching point pairs formed between any independent variable sequence z and any dependent variable sequence. This value indirectly indicates the stability of the influence relationship between the independent variable and the dependent variable and the reliability of the analysis. Therefore, the smaller this value, the higher the estimated lag time. This requires a slight increase to improve fault tolerance. This serves as a window adjustment parameter when determining the length of the time window.

[0059] Therefore, the length of the time window for any influencing data can be determined when predicting the value of any response data at the time to be predicted based on any influencing data. This allows for the determination of fertilization control after predicting water and fertilizer stock, including: When predicting the water and fertilizer reserves after the target time, the sampled data on each impact and the time window length corresponding to the prediction of each impact on each response data are input into the Long Short-Term Memory network to determine the predicted water and fertilizer reserves after the target time. The predicted water and fertilizer reserves are compared with the rainfall at the target time and the fertilizer requirements corresponding to the crop growth stage at the target time to complete the fertilization control at the target time.

[0060] Specifically, to achieve accurate water and fertilizer inventory prediction, this embodiment uses the time window length of each influencing data and its dynamic changes in response to each type of data at different prediction times as input to a Long Short-Term Memory (LSTM) network to predict the water and fertilizer inventory at the predicted time. During the prediction process, weather forecast data, especially precipitation, is obtained through a meteorological platform interface as a crucial basis for irrigation decisions. Specifically, the system calculates the actual amount of irrigation water needed by subtracting the precipitation during the forecast period from the predicted water requirement under the predicted humidity. Regarding fertilizer application, the system analyzes the crop's growth stage using image recognition technology to obtain the fertilizer data required by the crop. Then, the predicted fertilizer inventory is compared with the actual fertilizer requirement of the crop: if the predicted fertilizer inventory at a future time is greater than the requirement for the corresponding crop growth stage, no fertilization is needed; otherwise, the required fertilizer amount is calculated based on the difference, thus achieving precise fertilization control.

[0061] This intelligent prediction and control mechanism can significantly improve the precision of water and fertilizer control, effectively avoid water and fertilizer waste, improve the efficiency of water and fertilizer resource utilization, and ensure the healthy growth of crops.

[0062] Corresponding to the method in the above embodiments, Figure 4 A structural block diagram of an intelligent water and fertilizer integration control device according to Embodiment 3 of this application is shown. This control device is applied to an electronic device, which connects to a target database through a preset application programming interface (API). When the target database is driven to execute corresponding tasks, corresponding task logs are generated, which can be collected via an API. For ease of explanation, only the parts relevant to the embodiments of this application are shown.

[0063] See Figure 4 The control device includes: The data key point determination module 31 is used to sample various impact data affecting water and fertilizer storage and record each impact data sequence as an independent variable sequence, sample each response data reflecting water and fertilizer storage and record each response data sequence as a dependent variable sequence, and determine the key change points based on the local change characteristics of each sampling point in any sequence. The data key point matching module 32 is used to form key point pairs by pairing any changing key point in any independent variable sequence with each changing key point in any dependent variable sequence after the sampling time of the changing key point, thus obtaining a set of key point pairs for any changing key point. Starting from the first changing key point in any independent variable sequence, using any key point pair in the current set of key point pairs as the base pair, the module determines the key point pair with the highest similarity in local change similarity and temporal difference between the two changing key points within the base pair in the set of key point pairs for the next changing key point as the matching pair. The matching pair is used as the new base pair, and the process of determining the matching pair is repeated until all changing key points in any independent variable sequence are traversed, thus obtaining a sequence of matching pairs for any key point pair. The sampling time of the changing key point in the matching pair belonging to the changing key point in the dependent variable sequence is after the sampling time of the changing key point in the base pair belonging to the changing key point in the independent variable sequence. The water and fertilizer prediction and fertilization module 33 is used to determine the credibility of the matching point pair sequence of any key point pair based on the similarity between any two adjacent key point pairs in the matching point pair sequence of any key point pair and the local change similarity between two changing key points within each key point pair. The matching point pair sequence corresponding to the key point pair sequence with the highest credibility in the key point pair sequence of the first changing key point is taken as the target point pair sequence. The target point pair sequence is used to determine the time window length when predicting the reflection data corresponding to any dependent variable sequence based on the influence data corresponding to any independent variable sequence. Fertilization control is performed after determining the predicted water and fertilizer inventory based on the time window length.

[0064] It should be noted that the information interaction and execution process between the above modules are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, which will not be repeated here.

[0065] Figure 5 This is a schematic diagram of the structure of an electronic device provided in Embodiment 4 of this application. Figure 5 As shown, the electronic device of this embodiment includes: at least one processor ( Figure 5The diagram shows only one of the following: a memory and a computer program stored in the memory and capable of running on at least one processor. When the processor executes the computer program, it implements the steps described in the above-described intelligent water and fertilizer integration control method embodiment.

[0066] The electronic device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that... Figure 5 This is merely an example of an electronic device and does not constitute a limitation on electronic devices. Electronic devices may include more or fewer components than shown, or may combine certain components or different components.

[0067] The processor referred to can be a CPU, but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0068] Memory includes readable storage media, internal memory, etc., wherein internal memory can be the RAM of an electronic device, providing an environment for the operation of the operating system and computer-readable instructions stored in the readable storage media. The readable storage media can be the hard drive of an electronic device, or in other embodiments, it can be an external storage device of the electronic device, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, memory can include both internal storage units and external storage devices of the electronic device. Memory is used to store the operating system, applications, bootloader, data, and other programs, such as program code for computer programs. Memory can also be used to temporarily store data that has been output or will be output.

[0069] Those skilled in the art will understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this invention. The specific working process of the units and modules in the above device can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here. If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention can implement all or part of the processes in the above embodiments by instructing related hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the above method embodiments. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. A computer-readable medium can include at least: any entity or device capable of carrying computer program code, a recording medium, a computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0070] The present invention can implement all or part of the processes in the above embodiments of the method, or it can be accomplished by a computer program product. When the computer program product is run on an electronic device, the electronic device executes the steps in the above method embodiments.

[0071] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0072] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0073] In the embodiments provided by this invention, it should be understood that the disclosed devices / electronic devices and methods can be implemented in other ways. For example, the device / electronic device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0074] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0075] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A smart water and fertilizer integrated control method, characterized in that, The method includes: Sample multiple data affecting water and fertilizer storage and record each data sequence as an independent variable sequence. Sample each data sequence reflecting water and fertilizer storage and record each data sequence as a dependent variable sequence. Determine the key points of change based on the local change characteristics of each sampling point in any sequence. Each key point in any sequence of independent variables is paired with each key point in any sequence of dependent variables that is sampled after the key point in any sequence of independent variables, thus forming a key point pair set for each key point in any sequence of independent variables. Starting from the first key point in any sequence of independent variables, each key point pair in the current key point pair set is used as a base pair. In the key point pair set of the next key point in any sequence of independent variables, the key point pair with the highest similarity in local change similarity and temporal difference between the two key points in the base pair is determined as a matching pair. The matching pair is used as a new base pair, and the process of determining the matching pair is repeated until all key points in any sequence of independent variables are traversed, thus obtaining a matching pair sequence for each key point pair. The sampling time of the key point in any sequence of dependent variables in the matching pair is after the sampling time of the key point in any sequence of independent variables in the base pair. The credibility of the matching point pair sequence of any key point pair is determined by the similarity between any two adjacent key point pairs in the matching point pair sequence of any key point pair and the local change similarity between two changing key points within each key point pair. The matching point pair sequence corresponding to the key point pair sequence with the highest credibility in the key point pair sequence of the first changing key point is taken as the target point pair sequence. The time window length for predicting the reflection data corresponding to any dependent variable sequence based on the influence data corresponding to any independent variable sequence is determined by the target point pair sequence. Fertilization control is then carried out after predicting the water and fertilizer inventory based on the time window length.

2. The intelligent water and fertilizer integration control method according to claim 1, characterized in that, The process of determining key change points based on the local change characteristics of each sampling point in any sequence includes: Determine the peak interval between any two adjacent peak points in any sequence, and use half of the minimum peak interval as the analysis time for any sequence; The coefficient of variation of the second-order difference sequence of the data segment to the left of any sampling point in any sequence under the analysis time is calculated and denoted as the first coefficient of variation. The coefficient of variation of the second-order difference sequence of the data segment to the right of any sampling point under the analysis time is calculated and denoted as the second coefficient of variation. The mean slope of each sampling point in the data segment to the left of any sampling point under the analysis time is calculated and denoted as the first mean slope. The mean slope of each sampling point in the data segment to the right of any sampling point under the analysis time is calculated and denoted as the second mean slope. The absolute value of the difference between the mean of the first slope and the mean of the second slope is recorded as the abrupt change value. The mean of the first coefficient of variation and the second coefficient of variation is recorded as the stability of change value. The criticality of change of any sampling point is determined based on the abrupt change value and the stability of change value. The criticality of change is directly proportional to the abrupt change value and inversely proportional to the stability of change value. Sampling points whose change criticality is greater than that of the adjacent sampling points on the left and right sides are recorded as initial screening critical points. All the initial screening critical points are clustered according to the change criticality with a target cluster size of 2. Each initial screening critical point in the cluster with a larger mean change criticality is taken as the change critical point in any sequence.

3. The intelligent water and fertilizer integration control method according to claim 1, characterized in that, Determining the similarity and temporal differences in local changes between two key points of change within the aforementioned context includes: The key point in the current key point pair that belongs to the independent variable sequence is recorded as the previous key point, and the key point in the current key point pair that belongs to the dependent variable sequence is recorded as the next key point. The data segment between the previous key point and the next adjacent key point in its independent variable sequence is recorded as the first local sequence, and the data segment between the next key point and the next adjacent key point in its dependent variable sequence is recorded as the second local sequence. After aligning the starting sampling points of the first local sequence and the second local sequence, the Spearman rank correlation coefficient between the overlapping parts of the first local sequence and the second local sequence is determined. Based on the Spearman rank correlation coefficient, the local change similarity between the two changing key points inside the current key point is determined. Calculate the time difference between the previous key point and the next key point with respect to the sampling time, and determine the time difference between the current key point and the two changing key points inside it based on the time difference.

4. The intelligent water and fertilizer integration control method according to claim 1 or 3, characterized in that, Determining the similarity includes: Using any key point pair in the set of key point pairs of the next change key point as the comparison pair of the base point pair, calculate the normalized value of the absolute value of the difference between the local change similarity between the two change key points in the base point pair and the local change similarity between the two change key points in the comparison pair, and denot it as the first difference degree. The normalized value of the absolute value of the difference between the temporal difference between the two key points of change within the base point pair and the temporal difference between the two key points of change within the comparison point pair is denoted as the second difference degree. The similarity between the base point pair and the comparison point pair is determined based on the first difference degree and the second difference degree, and the similarity degree is inversely proportional to both the first difference degree and the second difference degree.

5. The intelligent water and fertilizer integration control method according to claim 1, characterized in that, Determining the credibility of the matching point pair sequence for any key point pair includes: In the matching point pair sequence of any key point pair, select any key point pair as the target key point pair. Calculate the similarity between the target key point pair and its two adjacent key point pairs in the matching point pair sequence of any key point pair. Subtract the two similarities and take the absolute value as the local stability of the similarity of the target key point pair. Calculate the sum of the local stability of the similarity of all key point pairs in the matching point pair sequence of any key point pair as the first sub-confidence of the matching point pair sequence of any key point pair. Calculate the mean of the local change similarity between two changing key points within all key point pairs in the matching point pair sequence of any key point pair, and use it as the second sub-credibility of the matching point pair sequence of any key point pair. The credibility of the matching point pair sequence of any key point pair is determined based on the first sub-credibility and the second sub-credibility, wherein the credibility is inversely proportional to the first sub-credibility and directly proportional to the second sub-credibility.

6. The intelligent water and fertilizer integration control method according to claim 1, characterized in that, The determination of the time window length for predicting the reflection data corresponding to any dependent variable sequence using the influence data corresponding to any independent variable sequence includes: Based on the sampling time difference between two changing key points within each key point pair in the target point pair sequence, determine the lag time corresponding to each key point pair in the target point pair sequence; based on the lag time corresponding to each key point pair in the target point pair sequence, determine the lag time sequence between the influence data corresponding to any independent variable sequence and the reflection data corresponding to any dependent variable sequence. When determining the data after the target time of the reflection data corresponding to any dependent variable sequence based on the impact data corresponding to any independent variable sequence according to the lag time series, the time window length used for the impact data corresponding to any independent variable sequence; when predicting the data after the target time of the reflection data corresponding to any dependent variable sequence based on the impact data corresponding to any independent variable sequence, the time window length for selecting the impact data corresponding to any independent variable sequence is determined according to the lag time series. The time window length is equal to the product of the sum of a constant 1 and a window adjustment parameter multiplied by the lag time corresponding to the target time in the lag time series. The window adjustment parameter is the normalized value of the negative of the maximum confidence value of the matching point pair sequence in the key point pair set of the first change key point.

7. The intelligent water and fertilizer integration control method according to claim 6, characterized in that, The process of determining and controlling fertilization after predicting water and fertilizer reserves includes: When predicting the water and fertilizer reserves after the target time, the sampled data on each impact and the time window length corresponding to the prediction of each impact on each response data are input into the Long Short-Term Memory network to determine the predicted water and fertilizer reserves after the target time. The predicted water and fertilizer reserves are compared with the rainfall at the target time and the fertilizer requirements corresponding to the crop growth stage at the target time to complete the fertilization control at the target time.

8. An intelligent integrated water and fertilizer control device, characterized in that, The device includes: The data key point determination module is used to sample various impact data affecting water and fertilizer storage and record each impact data sequence as an independent variable sequence, sample each response data reflecting water and fertilizer storage and record each response data sequence as a dependent variable sequence, and determine the key change points based on the local change characteristics of each sampling point in any sequence; The data key point matching module is used to form key point pairs by pairing any changing key point in any independent variable sequence with each changing key point in any dependent variable sequence after the sampling time of the changing key point, thus obtaining a set of key point pairs for the changing key point. Starting from the first changing key point in the independent variable sequence, using any key point pair in the current set of key point pairs as the base pair, the module determines the key point pair with the highest similarity in local change similarity and temporal difference between the two changing key points within the base pair in the set of key point pairs for the next changing key point as the matching pair. The matching pair is used as the new base pair, and the process of determining the matching pair is repeated until all changing key points in the independent variable sequence are traversed, resulting in a sequence of matching pairs for the given key point pairs. The sampling time of the changing key point in the matching pair belonging to the changing key point in the dependent variable sequence is after the sampling time of the changing key point in the base pair belonging to the changing key point in the independent variable sequence. The water and fertilizer prediction and fertilization module is used to determine the credibility of the matching point pair sequence of any key point pair based on the similarity between any two adjacent key point pairs in the matching point pair sequence of any key point pair and the local change similarity between two changing key points within each key point pair. The matching point pair sequence corresponding to the key point pair sequence with the highest credibility in the key point pair sequence of the first changing key point is taken as the target point pair sequence. The target point pair sequence is used to determine the time window length when predicting the reflection data corresponding to any dependent variable sequence based on the influence data corresponding to any independent variable sequence. Fertilization control is then performed after determining the predicted water and fertilizer inventory based on the time window length.

9. An electronic device, characterized in that, The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the intelligent water and fertilizer integration control method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the intelligent water and fertilizer integration control method as described in any one of claims 1 to 7.

Images