Dynamic regulation method for maintaining soil nutrient balance in watermelon cultivation

Abstract

The application provides a dynamic regulation method for maintaining the stability of soil nutrients for watermelon cultivation, and relates to the technical field of soil nutrient regulation, which comprises the following steps: collecting soil nutrient parameters of multiple monitoring points in a target soil area of continuous cropping watermelon, obtaining a nutrient parameter array, and calculating average nutrient parameters; randomly generating a first regulation scheme, performing regulation prediction, obtaining first regulation nutrient parameters, performing stability evaluation, and obtaining stability fitness; obtaining continuous cropping parameters of the target soil area, and configuring an optimization step; analyzing the nutrient error amplitude, obtaining element antagonistic coefficients, compensating the optimization step, obtaining an optimized step set, adjusting the first regulation scheme, obtaining a second regulation scheme set, and continuously performing regulation optimization to obtain an optimal regulation scheme, and regulating the target soil area. The technical problem of low regulation accuracy of the traditional soil nutrient regulation method in the prior art, and even over-regulation leading to deterioration of the soil nutrient state is solved.

Description

Technical Field

[0001] This invention relates to the field of soil nutrient regulation, and more particularly to a dynamic regulation method for maintaining soil nutrient stability in watermelon cultivation. Background Technology

[0002] During continuous watermelon cultivation, due to the selective absorption of nutrients by a single crop over a long period and the accumulation of root exudates, key nutrients such as potassium and nitrogen in the soil are prone to supply and demand imbalances, as well as antagonistic effects between elements. This leads to a continuous decline in soil nutrient stability, directly affecting the growth, yield, and quality of watermelons.

[0003] However, traditional soil nutrient regulation methods rely heavily on empirical fertilization schemes, lacking the ability to perceive and adapt to dynamic changes in soil nutrients in real time. They cannot accurately predict the actual effects of regulation measures, nor can they take into account complex factors such as continuous cropping time and differences in nutrient spatial distribution. This can easily lead to a disconnect between regulation schemes and actual soil needs, making it difficult to effectively improve nutrient imbalances. In fact, over-regulation may even exacerbate antagonistic effects between elements, leading to further deterioration of soil nutrient status.

[0004] Therefore, there is an urgent need for a method for dynamic regulation of soil nutrients in watermelon cultivation that integrates intelligent prediction and dynamic optimization mechanisms. Summary of the Invention

[0005] This invention addresses the technical problems of low adjustment accuracy or even over-adjustment leading to deterioration of soil nutrient status in traditional soil nutrient regulation methods in the prior art, and provides a dynamic regulation method for maintaining soil nutrient stability in watermelon cultivation.

[0006] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: This invention provides a dynamic regulation method for maintaining soil nutrient stability in watermelon cultivation, comprising: Soil nutrient parameters were collected from multiple monitoring points within the target soil area of ​​continuously cropped watermelons to obtain a nutrient parameter array. The average nutrient parameters were calculated, where each nutrient parameter includes the content of potassium and nitrogen. A first regulation scheme for soil regulation is randomly generated, and the regulation is predicted in combination with the average nutrient parameters to obtain the first regulation nutrient parameters. A stability assessment is then conducted to obtain the stability fitness. Obtain the continuous cropping parameters for the target soil region and configure the optimization step size; The nutrient error range of the nutrient parameter array is analyzed, and the element antagonism coefficient is obtained. The optimization step size is compensated to obtain an optimization step size set. The first adjustment scheme is adjusted to obtain a second adjustment scheme set. The adjustment and optimization are continued to obtain the optimal adjustment scheme and the target soil area is adjusted.

[0007] The beneficial effects of this invention are: Compared to existing technologies, this application first collects soil nutrient parameters from multiple monitoring points within the target soil area for continuously cropped watermelons, obtaining a nutrient parameter array and calculating the average nutrient parameters. Through scientific sampling and calculation, the random errors of single-point sampling are eliminated, obtaining benchmark data that represents the overall nutrient status of the target soil area for continuously cropped watermelons, providing a reliable basis for subsequent adjustment schemes. Secondly, a first adjustment scheme for soil regulation is randomly generated, and adjustment prediction is performed based on the average nutrient parameters to obtain the first regulated nutrient parameters. A stability assessment is then conducted to obtain the stability fitness. By randomly generating the first adjustment scheme and combining it with the current average soil nutrient parameters, the adjustment effect is predicted, and its proximity to the ideal state is quantified, providing a quantifiable starting point and evaluation standard for subsequent optimization. Thirdly, the continuous cropping parameters of the target soil area are obtained, and an optimization step size is configured to dynamically match the actual degree of soil degradation, balancing adjustment efficiency and accuracy, and providing a reasonable search basis for iterative optimization of subsequent schemes. Finally, the nutrient error range of the nutrient parameter array was analyzed, and the element antagonism coefficient was obtained. The optimization step size was compensated to obtain an optimized step size set. The first regulation scheme was adjusted to obtain a second regulation scheme set. The regulation and optimization were continued to obtain the optimal regulation scheme. The target soil area was regulated, which solved the problems of fixed step size and easy omission of optimal solution in traditional regulation. The regulation scheme can adapt to the actual differences in soil and avoid the risk of element interaction. Finally, the accuracy and efficiency of nutrient stabilization regulation were achieved.

[0008] Through the aforementioned technical solution, this application collects data on the spatial heterogeneity of nutrients in the target soil area for continuously cropped watermelons at multiple monitoring points, accurately predicts the regulation effect using machine learning models, dynamically configures and optimizes the step size based on the continuous cropping time, and compensates and calibrates the step size using sample error amplitude and element antagonism coefficient. Finally, the optimal solution is selected through iterative optimization. This effectively overcomes the interference of nitrogen and potassium antagonism on nutrient absorption, adapts to the nutrient patterns of continuously cropped soils changing with planting years, improves the adaptability of the regulation scheme to complex soil conditions, and reduces the risk of nutrient imbalance caused by monitoring errors or blind regulation through refined step size design and multi-dimensional parameter integration. Thus, the precision of soil nutrient regulation is improved, achieving an upgrade from empirical regulation to dynamic, precise, and intelligent regulation, ultimately realizing precise regulation of soil nutrients for continuously cropped watermelons. Attached Figure Description

[0009] Figure 1 This is a flowchart illustrating the dynamic regulation method for maintaining soil nutrient stability in watermelon cultivation provided by the present invention. Figure 2The flowchart illustrates the process of obtaining element antagonism coefficients by performing element antagonism analysis based on the average nutrient parameters in the dynamic regulation method for maintaining soil nutrient stability in watermelon cultivation provided by the present invention. Detailed Implementation

[0010] 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.

[0011] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0012] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0013] Example 1, as Figure 1 As shown, this embodiment of the invention provides a dynamic regulation method for maintaining soil nutrient stability in watermelon cultivation, including: S10: Collect soil nutrient parameters from multiple monitoring points within the target soil area of ​​continuously cropped watermelons, obtain a nutrient parameter array, and calculate the average nutrient parameters, where each nutrient parameter includes the content of potassium and nitrogen.

[0014] Potassium and nitrogen are essential nutrients for plant growth. However, in the process of continuous watermelon planting, due to the long-term absorption characteristics of a single crop, the supply and demand of nutrients such as potassium and nitrogen in the soil are prone to imbalance, which in turn leads to limited watermelon growth and reduced quality.

[0015] To address the aforementioned issues, this application collects soil nutrient parameters from multiple monitoring points within the target soil area of ​​continuously cropped watermelons, obtains a nutrient parameter array, and calculates the average nutrient parameters, where each nutrient parameter includes the content of potassium and nitrogen.

[0016] Specifically, step S10 in the method includes: Soil nutrient parameters were collected from multiple monitoring points within the target soil area of ​​continuously cropped watermelons to obtain a nutrient parameter array, where each nutrient parameter includes the content of potassium and nitrogen. Calculate the mean of the nutrient parameter array to obtain the average nutrient parameters.

[0017] In this embodiment, soil nutrient parameters are first collected from multiple monitoring points within the target soil area for continuously cropped watermelons to obtain a nutrient parameter array, wherein each nutrient parameter includes the content of potassium and nitrogen. Specifically, within the target soil area for continuously cropped watermelons, multiple monitoring points are set up according to the actual size of the target soil area, following a uniform distribution or gridding principle. Soil samples are collected from each monitoring point, and the soil nutrient parameters, including potassium and nitrogen content, are detected and recorded for that monitoring point. For example, using a grid-based sampling method, 3-5 monitoring points are evenly set up in the target soil area for continuously cropped watermelons. Soil samples from the 0-20cm topsoil layer are collected using a 2cm diameter spiral soil auger. Each monitoring point is sampled three times and the samples are combined into one soil sample. For potassium content, the available potassium can be determined using a neutral ammonium acetate extraction-flame photometry method. For nitrogen content, the sample can be extracted with 0.01mol / L CaCl2 solution, and nitrate nitrogen can be determined using a nitrate ion-selective electrode, while ammonium nitrogen can be determined using an ammonium ion-selective electrode. For higher precision data, some samples can be sent to a laboratory for total potassium measurement using an atomic absorption spectrophotometer and total nitrogen measurement using a Kjeldahl nitrogen analyzer. This results in soil nutrient parameters containing potassium and nitrogen content from each monitoring point, forming a nutrient parameter array. For example, if three monitoring points are set up, the nutrient parameter array can be represented as: [(potassium content 180 mg / kg, nitrogen content 140 mg / kg), (potassium content 200 mg / kg, nitrogen content 160 mg / kg), (potassium content 220 mg / kg, nitrogen content 160 mg / kg), (potassium content 220 mg / kg, nitrogen content 14 ... mg / kg, nitrogen content 180 mg / kg).

[0018] Next, calculate the mean of the nutrient parameter array to obtain the average nutrient parameters. For example, if the nutrient parameter array is: [(potassium content 180 mg / kg, nitrogen content 140 mg / kg), (potassium content 200 mg / kg, nitrogen content 160 mg / kg), (potassium content 220 mg / kg, nitrogen content 180 mg / kg)], calculate the mean potassium content and the mean nitrogen content respectively, and obtain the average potassium content = (180 + 200 + 220) / 3 = 200 mg / kg, and the average nitrogen content = (140 + 160 + 180) / 3 = 160 mg / kg, which are used as the average nutrient parameters (average potassium content 200 mg / kg, average nitrogen content 160 mg / kg).

[0019] In summary, compared to existing technologies, this application collects soil nutrient parameters from multiple monitoring points within the target soil area of ​​continuously cropped watermelons, obtains a nutrient parameter array, and calculates the average nutrient parameters. Thus, through scientific sampling and calculation, the random errors of single-point sampling are eliminated, and benchmark data representing the overall nutrient status of the target soil area for continuously cropped watermelons is obtained, providing a reliable basis for subsequent adjustment schemes.

[0020] S20: Randomly generate a first regulation scheme for soil adjustment, combine it with the average nutrient parameters to predict the adjustment, obtain the first regulation nutrient parameters, conduct a stability assessment, and obtain the stability fitness.

[0021] Traditional regulation methods lack a systematic quantitative evaluation system and data-driven prediction models. They have neither established a unified standard benchmark for soil nutrients nor been able to simulate the regulation effect through scientific algorithms. This makes it difficult to predict the expected regulation effect of the regulation scheme, and thus it is impossible to accurately judge the actual effect of the scheme on soil nutrient stability. Often, insufficient or excessive regulation will cause nutrient imbalance and affect the growth stability of watermelons.

[0022] To address the aforementioned issues, this application randomly generates a first adjustment scheme for soil regulation, combines the average nutrient parameters to predict the adjustment, obtains the first regulated nutrient parameters, conducts a stability assessment, and obtains the stability fitness.

[0023] Specifically, step S20 in the method includes: Obtain a space for soil conditioning schemes, wherein the space for soil conditioning schemes includes a potassium replenishment space and a nitrogen replenishment space; A first adjustment scheme for soil adjustment is randomly generated within the adjustment scheme space; Based on the first adjustment scheme and the average nutrient parameters, an adjustment prediction is made to obtain the first adjustment nutrient parameters; Obtain standard nutrient parameters and calculate the first homeostatic fitness of the first adjusted nutrient parameters.

[0024] In this embodiment, a soil conditioning scheme space is first obtained. This space includes a potassium supplementation space and a nitrogen supplementation space, which are sets of all possible potassium and nitrogen supplementation schemes. These spaces are used to constrain the rationality of the conditioning schemes and prevent excessive or insufficient supplementation, thus exceeding reasonable operating limits. For example, the potassium supplementation space can be determined based on the actual potassium deficiency in the target soil area for continuous watermelon cropping, the potassium requirement for watermelon growth, and the fertilization safety threshold, for example, 0-50 kg / mu. Similarly, the nitrogen supplementation space can be determined based on the actual nitrogen deficiency in the target soil area for continuous watermelon cropping, the nitrogen requirement for watermelon growth, and the fertilization safety threshold, for example, 0-30 kg / mu. Thus, the soil conditioning scheme space is obtained as follows: potassium supplementation space is 0-50 kg / mu, and nitrogen supplementation space is 0-30 kg / mu.

[0025] Secondly, the potassium and nitrogen supplementation amounts for soil adjustment are randomly generated within the adjustment scheme space and used as the first adjustment scheme. For example, 25 kg / mu is randomly selected from the potassium supplementation space (0-50 kg / mu) and 12 kg / mu is randomly selected from the nitrogen supplementation space (0-30 kg / mu) as the first adjustment scheme: potassium supplementation 25 kg / mu and nitrogen supplementation 12 kg / mu. In this way, the first adjustment scheme is randomly generated from the adjustment scheme space, avoiding the bias of human experience in the initial scheme and ensuring that subsequent optimization can select the optimal solution from a wider range of possibilities.

[0026] Next, adjustment predictions are made based on the first adjustment scheme and average nutrient parameters. A nutrient adjustment predictor built using machine learning predicts the potassium and nitrogen contents of the soil after adjustment according to the first scheme, serving as the first nutrient adjustment parameters. The core logic of the nutrient adjustment predictor is to quantify the relationship between the adjustment scheme and changes in soil nutrients based on historical data and machine learning. For example, if the average nutrient parameters are an average potassium content of 200 mg / kg and an average nitrogen content of 160 mg / kg, and the first adjustment scheme is to supplement potassium at 25 kg / mu and nitrogen at 12 kg / mu, the pre-trained nutrient adjustment predictor predicts that after adjustment, potassium may rise to 225 mg / kg and nitrogen to 170 mg / kg, resulting in the first nutrient adjustment parameters: (225 mg / kg, 170 mg / kg).

[0027] Finally, standard nutrient parameters are obtained, and the first stability fitness of the first adjusted nutrient parameter is calculated. The standard nutrient parameters represent the optimal potassium and nitrogen contents for watermelon growth, such as potassium at 230 mg / kg and average nitrogen at 175 mg / kg. For example, using the standard nutrient parameters as a benchmark, the similarity between the first adjusted nutrient parameter and the standard nutrient parameter is calculated as the first stability fitness. The smaller the difference between the first adjusted nutrient parameter and the standard nutrient parameter, the higher the similarity, and thus the higher the first stability fitness, indicating that the soil nutrients are closer to the ideal state after adjustment by the first adjusted nutrient parameter.

[0028] Specifically, the step of "making adjustment predictions based on the first adjustment scheme and average nutrient parameters to obtain the first adjustment nutrient parameters" includes: Based on historical watermelon soil conditioning data, a set of sample conditioning schemes and a set of sample nutrient parameters were collected. The conditioned nutrient parameters under different sample conditioning schemes and sample nutrient parameters were also collected to obtain a set of sample conditioned nutrient parameters. Construct a nutrient adjustment predictor based on machine learning; The nutrient adjustment predictor is trained under supervision using the set of sample adjustment schemes, the set of sample nutrient parameters, and the set of sample adjusted nutrient parameters until the training converges.

[0029] In this embodiment, several potassium and nitrogen supplementation schemes are first collected based on historical watermelon soil conditioning data, such as potassium supplementation of 10 kg / mu and nitrogen supplementation of 5 kg / mu, as a sample conditioning scheme set. Soil nutrient parameters before the sample conditioning scheme are extracted, such as potassium 170 mg / kg and nitrogen 130 mg / kg, as a sample nutrient parameter set. Then, regulated nutrient parameters under different sample conditioning schemes and sample nutrient parameter adjustments are collected, such as potassium 185 mg / kg and nitrogen 138 mg / kg after the sample conditioning scheme, to obtain a sample regulated nutrient parameter set.

[0030] Secondly, a nutrient adjustment predictor based on machine learning is constructed. For example, the nutrient adjustment predictor can adopt a three-layer fully connected neural network architecture, mainly composed of an input layer, hidden layers, and an output layer. The input layer contains four neurons, corresponding to the input features: potassium supplementation amount, nitrogen supplementation amount, soil potassium content before potassium supplementation, and soil nitrogen content before nitrogen supplementation. It is responsible for converting the original adjustment scheme and soil nutrient parameters into vector signals that the network can process. The hidden layer is designed as a three-layer progressive structure. The first layer contains 64 neurons and uses the ReLU activation function to extract the basic correlation of input features through nonlinear transformation, such as the interaction between potassium supplementation amount and current potassium content. The second layer contains 32 neurons and uses the ReLU activation function to further compress feature dimensions and enhance key information. The third layer contains 16 neurons and uses the ReLU activation function to achieve high-order feature fusion, capturing the complex mapping law between the adjustment scheme and soil nutrient changes. The output layer contains two neurons and uses a linear activation function to directly output the adjusted nutrient parameters: adjusted potassium content and nitrogen content. Finally, the nutrient adjustment predictor is trained under supervision using the sample conditioning scheme set, sample nutrient parameter set, and sample-adjusted nutrient parameter set until training converges. For example, the nutrient adjustment predictor can be trained using the following technical path: 1. Data preparation: Divide the sample conditioning scheme set, sample nutrient parameter set, and sample-adjusted nutrient parameter set into training, validation, and test sets according to a ratio of 7:1.5:1.5. The training set is used for model parameter learning, the validation set is used to monitor overfitting risk, and the test set is used for final performance evaluation. 2. Model Training: Using the sample conditioning scheme and sample nutrient parameters in the training set as input features, and the sample conditioning nutrient parameters as supervision labels, the Adam optimizer (initial learning rate 0.001) is used for iterative training by minimizing the mean squared error (MSE) loss function. During training, the validation set loss is calculated after each iteration. When the validation set loss does not decrease for 30 consecutive iterations (or a preset number of iterations), the model is considered to have converged and training is stopped. At the same time, to avoid overfitting, a 5% Dropout mechanism can be introduced in the hidden layer, and an early stopping strategy is used to finally obtain the trained nutrient conditioning predictor.

[0031] Furthermore, the step of "obtaining standard nutrient parameters and calculating the first homeostatic fitness of the first adjusted nutrient parameters" includes: Obtain standard nutrient parameters; Calculate the similarity between the first adjusted nutrient parameter and the standard nutrient parameter to obtain the first homeostatic fitness.

[0032] In this embodiment, standard nutrient parameters are first obtained. Specifically, the standard nutrient parameters are the optimal baseline values ​​of potassium and nitrogen content in the soil, which determine the target optimization direction of the adjustment scheme and can be determined in combination with the growth requirements of watermelons and soil characteristics. For example, field trials at different watermelon growth stages (such as seedling stage, fruit setting stage, and maturity stage) can be conducted to determine the soil potassium and nitrogen content ranges that maximize watermelon yield and achieve optimal quality (such as sugar content and taste). For instance, experiments have shown that the optimal available potassium content during the watermelon fruit setting stage is 180–220 mg / kg, and the available nitrogen content is 140–160 mg / kg. These can be used as general standard nutrient parameters. These parameters can then be localized based on the soil type of the target soil region. This is because different soil types (such as sandy soil and clay loam) have different nutrient retention capacities. Therefore, the general standard nutrient parameters need to be fine-tuned according to the actual soil type. For example, sandy soil has a weak fertilizer retention capacity, so the standard nitrogen content can be appropriately increased to avoid rapid loss, while clay loam can be appropriately decreased to prevent accumulation. If field trial data is lacking, recommended nutrient values ​​for watermelon cultivation from agricultural industry standards or authoritative research can be cited to ensure the scientific validity of the standard nutrient parameters.

[0033] Secondly, the similarity between the first adjusted nutrient parameter and the standard nutrient parameter is calculated to obtain the first stability fitness. The higher the similarity between the first adjusted nutrient parameter and the standard nutrient parameter, the closer the first adjustment scheme is to the optimal state of soil nutrients after adjustment. That is, the first stability fitness can evaluate the adjustment effect of the first adjusted nutrient parameter; the higher the first stability fitness, the better the adjustment effect of the first adjusted nutrient parameter. For example, the similarity between the first adjusted nutrient parameter and the standard nutrient parameter can be calculated using the difference rate method or the Euclidean distance method. For instance, if the first adjusted nutrient parameter is potassium content of 225 mg / kg and nitrogen content of 170 mg / kg, and the standard nutrient parameter is standard potassium content of 220 mg / kg and standard nitrogen content of 165 mg / kg, the first stability fitness can be calculated using the difference rate method: calculate the absolute difference rate between the first adjusted nutrient parameter and the standard nutrient parameter respectively, and then take the average as the average difference rate. First stability fitness = 1 - average difference rate, where the absolute difference rate of potassium content = |potassium content - standard nitrogen content|. The absolute difference rate of potassium content is calculated as follows: || Potassium content| / || Standard potassium content. The average difference rate is calculated as follows: (Absolute difference rate of potassium content + Absolute difference rate of nitrogen content) / 2. For example, the absolute difference rate of potassium content is approximately 2.3% (|225-220| / 220), and the absolute difference rate of nitrogen content is approximately 3% (|170-165| / 165). Therefore, the average difference rate is (2.3%+3%) / 2 = 2.65%. The first stability fitness is 1 - 2.65% = 97.35%. The first stability fitness ranges from 0 to 1, and a higher first stability fitness is better.

[0034] In summary, compared to existing technologies, this application randomly generates a first regulation scheme for soil adjustment, combines it with the average nutrient parameters to predict the adjustment, obtains the first regulated nutrient parameters, performs a stability assessment, and obtains the stability fitness. Thus, by randomly generating a first regulation scheme, combining it with the current average soil nutrient parameters to predict the regulation effect, and quantifying its proximity to the ideal state, a quantifiable starting point and evaluation standard are provided for subsequent optimization.

[0035] S30: Obtain the continuous cropping parameters of the target soil area and configure the optimization step size.

[0036] The duration of continuous cropping is positively correlated with the state of soil nutrient imbalance. The longer the continuous cropping period, the more watermelons absorb potassium and nitrogen and accumulate root exudates, which leads to a decrease in soil potassium content and an imbalance in nitrogen form. In other words, the more severe the soil nutrient imbalance, the larger the step size is needed to quickly explore effective adjustment solutions, such as significantly increasing the amount of potassium and nitrogen supplemented, to avoid the optimization process being too lengthy and missing key adjustment directions due to too small a step size. When the continuous cropping period is shorter, the soil imbalance is milder, and precise adjustment can be achieved with a smaller step size, avoiding over-adjustment caused by a large step size.

[0037] To address the aforementioned issues, this application obtains the continuous cropping parameters of the target soil region and configures the optimization step size.

[0038] Specifically, step S30 in the method includes: Obtain the continuous cropping parameters of the target soil area, wherein the continuous cropping parameters include the continuous cropping time; The average continuous cropping time of watermelons is obtained, the ratio of the continuous cropping time to the average continuous cropping time is calculated, and the preset step size is adjusted and configured as the optimized step size. The optimized step size includes the percentage range of the adjustment scheme.

[0039] In this embodiment, the continuous cropping parameters of the target soil area are first obtained. These parameters include the continuous cropping time, which refers to the number of years or cycles during which watermelons are continuously grown in the target soil area, such as 3 years or 5 seasons. For example, the continuous cropping parameter for the target soil area is 3 years.

[0040] Secondly, the average continuous cropping time of watermelons is obtained, and the ratio of continuous cropping time to average continuous cropping time is calculated. The preset step size is then adjusted and configured as the optimized step size. The preset step size is the initial percentage range set in the adjustment scheme, which can be dynamically set according to actual conditions such as soil type, planting scale, and historical adjustment effects, for example, 1% or 2%. The optimized step size includes adjusting the percentage range of the adjustment scheme, such as 2% or 5%, which is used to control the search range of the scheme iteration. The larger the step size, the larger the adjustment range each time. For example, adjusting from potassium supplementation of 20 kg / mu to potassium supplementation of 21 kg / mu, the percentage adjustment range of the adjustment scheme is 5%. The smaller the step size, the more refined the adjustment. For example, adjusting from potassium supplementation of 20 kg / mu to 20.5 kg / mu, the percentage adjustment range of the adjustment scheme is 2.5%. For example, if the preset step size is 5%, the average continuous cropping time of watermelons in similar soils or regions is extracted from agricultural statistics, such as 2 years, as a benchmark for measuring the continuous cropping parameters of the target soil region. Then, the ratio of continuous cropping time to average continuous cropping time is calculated as 3 / 2 = 1.5, indicating that the continuous cropping intensity of the target soil region is higher than the average level and the soil nutrient imbalance is more serious. Finally, the ratio of continuous cropping time to average continuous cropping time is multiplied by the preset step size to obtain the optimized step size = 5% * 1.5 = 7.5%. That is, when adjusting the adjustment scheme in the future, the amount of potassium and nitrogen supplementation can be increased or decreased by 7.5% each time.

[0041] In summary, compared to existing technologies, this application obtains the continuous cropping parameters of the target soil region and configures the optimization step size. This allows the optimization step size to dynamically match the actual degree of soil degradation, balancing adjustment efficiency and accuracy, and providing a reasonable search basis for subsequent iterative optimization of the scheme.

[0042] S40: Analyze the nutrient error range of the nutrient parameter array, process and obtain the element antagonism coefficient, compensate the optimization step size to obtain an optimization step size set, adjust the first adjustment scheme to obtain a second adjustment scheme set, and continue to adjust and optimize to obtain the optimal adjustment scheme, and adjust the target soil area.

[0043] Errors in soil nutrient monitoring and antagonistic interactions between elements can reduce the accuracy of optimization schemes. For example, monitoring errors may lead to biased judgments about the actual nutrient status of the soil. If the optimization step size is set too large in this case, potential better solutions may be skipped during the iteration process. Furthermore, the existence of element antagonism makes the absorption relationship between potassium and nitrogen more complex, further increasing the optimization difficulty. Therefore, it is necessary to combine the element antagonism coefficient (reflecting the inhibitory strength of nitrogen on potassium) and the sample error amplitude (reflecting the spatial differences in soil nutrients) to dynamically compensate for the optimization step size, thereby improving the accuracy and comprehensiveness of the scheme search and reducing the risk of missing the optimal solution.

[0044] To address the aforementioned issues, this application analyzes the nutrient error range of the nutrient parameter array, processes and obtains the element antagonism coefficient, compensates the optimization step size to obtain an optimization step size set, adjusts the first adjustment scheme to obtain a second adjustment scheme set, and continues to perform adjustment and optimization to obtain the optimal adjustment scheme for adjusting the target soil region.

[0045] Specifically, step S40 in the method includes: Calculate the average error amplitude between the nutrient parameter array and the average sample parameters, and use it as the sample error amplitude; Based on the average nutrient parameters, elemental antagonism analysis was performed to obtain the elemental antagonism coefficient; The step size compensation coefficient is calculated based on the sample error amplitude and the element antagonism coefficient. The optimized step size is compensated using the step size compensation coefficient to obtain a compensation step size interval. All optimized step sizes within the compensation step size interval are selected to obtain an optimized step size set, wherein the minimum scale of the optimized step size is one percent. Using the optimized step size set, the first adjustment scheme is adjusted to obtain the second adjustment scheme set; Continue iterative optimization until the preset number of optimizations is reached, and output the adjustment scheme with the highest stability and fitness as the optimal adjustment scheme.

[0046] In this embodiment, the average error range between the nutrient parameter array and the average sample parameter is first calculated as the sample error range. The sample error range can reflect the degree of deviation between the nutrient parameters of each monitoring point in the target soil area and the average sample parameter. The larger the sample error range, the more uneven the nutrient distribution in the target soil area, and the more precise the step size adjustment is needed to cover the needs of different areas. For example, if the nutrient parameter array is: [(potassium content 180 mg / kg, nitrogen content 140 mg / kg), (potassium content 200 mg / kg, nitrogen content 160 mg / kg), (potassium content 220 mg / kg, nitrogen content 180 mg / kg)], and the average nutrient parameters are average potassium content 200 mg / kg and average nitrogen content 160 mg / kg, and the error ranges between the potassium content at the three monitoring points and the average potassium content are (|180-200|) / 200=10%, (|200-200|) / 200=0%, and (|220-200|) / 200=10%, respectively, then the average error range of potassium content = (10% + 0%). (12.5% ​​+ 0% + 12.5%) / 3 = 6.67%. The absolute differences between the nitrogen content and the average nitrogen content are (|140-160|) / 160 = 12.5%, (|160-160|) / 160 = 0%, and (|180-160|) / 160 = 12.5%, respectively. Therefore, the average error range of nitrogen content = (12.5% ​​+ 0% + 12.5%) / 3 = 8.3%. Then, calculate the mean of the average error range of potassium content and the average error range of nitrogen content, and get the sample error range = (6.67% + 8.3%) ≈ 7.5%. The larger the sample error range, the more uneven the distribution of soil nutrients, and the more refined the step size adjustment is needed to cover the needs of different areas.

[0047] Secondly, based on the average nutrient parameters, element antagonism analysis was conducted to obtain the element antagonism coefficient. Element antagonism refers to the mutual inhibition effect between certain nutrients in the soil. When the nitrogen content in the soil is too high, it will interfere with the absorption and utilization of potassium by the watermelon roots, so that even if the potassium content is sufficient, the plant cannot effectively obtain it, thus affecting growth.

[0048] Next, the step size compensation coefficient is calculated based on the sample error amplitude and the element antagonism coefficient, where the step size compensation coefficient = sample error amplitude + element antagonism coefficient. For example, if the sample error amplitude is 7.5% and the element antagonism coefficient is 4.5%, then the step size compensation coefficient = 7.5% + 2.5% = 10%.

[0049] Furthermore, a step size compensation coefficient is used to compensate the optimized step size, obtaining a compensated step size interval. All optimized step sizes within the compensated step size interval are selected to obtain the optimized step size set. The compensated step size interval = optimized step size * (1 ± step size compensation coefficient), and the minimum scale of the optimized step size is one percent. For example, if the step size compensation coefficient is 12% and the optimized step size is 7.5%, then the compensated step size interval = 7.5% * (1 ± 10%) = 6.75% to 8.25%. All optimized step sizes of 7% and 8% within this interval are selected with a precision of 1% as the optimized step size set.

[0050] Furthermore, an optimized step size set is used to adjust the first adjustment scheme to obtain a second adjustment scheme set. For example, if the first adjustment scheme is to supplement potassium at 25 kg / mu and nitrogen at 12 kg / mu, the first adjustment scheme is adjusted according to each optimized step size in the optimized step size set. For instance, when the optimized step size is 7%, the second adjustment scheme is to supplement potassium at 25*(1+7%) = 26.75 kg / mu and nitrogen at 12*(1+7%) = 12.84 kg / mu. When the optimized step size is 8%, the second adjustment scheme is to supplement potassium at 25*(1+8%) = 27 kg / mu and nitrogen at 12*(1+8%) = 12.96 kg / mu, thus forming the second adjustment scheme set.

[0051] Finally, iterative optimization continues until a preset number of optimizations is reached. The adjustment scheme with the highest stability fitness is output as the optimal adjustment scheme. For example, following the same method as generating the second adjustment scheme set, for each second adjustment scheme in the second adjustment scheme set, adjustment prediction is performed according to step S20 to obtain the adjustment nutrient parameters, stability assessment is performed to obtain the stability fitness, and then the second adjustment scheme is adjusted according to the aforementioned steps in S30 to obtain the third adjustment scheme set. The next round of iteration is then entered until a preset number of optimizations is reached (e.g., 50 rounds, which can be dynamically set by those skilled in the art according to actual needs). The adjustment scheme with the highest stability fitness is output as the optimal adjustment scheme.

[0052] Specifically, such as Figure 2 As shown, the step of "performing elemental antagonism analysis based on the average nutrient parameters to obtain elemental antagonism coefficients" includes: Based on historical watermelon planting soil conditioning data, a set of sample nutrient parameters was collected, and the extent to which nitrogen inhibited potassium absorption under different sample nutrient parameters was collected. The set of sample element antagonism coefficients was then labeled. Based on the set of sample nutrient parameters and the set of sample element antagonism coefficients, an element antagonism classifier is constructed. The potassium and nitrogen content within the average nutrient parameters is input into the element antagonism classifier, and the element antagonism coefficient is obtained by classification output.

[0053] In the embodiments of this application, such as Figure 2 As shown, firstly, based on historical soil conditioning data for watermelon cultivation, a set of sample nutrient parameters was collected, and the extent to which nitrogen inhibited potassium absorption under different sample nutrient parameters was collected, and the set of sample element antagonism coefficients was obtained. The extent to which nitrogen inhibited potassium absorption under different sample nutrient parameters was determined through a controlled field micro-plot experiment: firstly, different nitrogen and potassium content gradient combinations were designed to form multiple groups of sample nutrient parameter treatments, while a nitrogen-free control group was set up. Then, watermelon seedlings with uniform growth were planted, and soil potassium residue and plant potassium accumulation were measured during key growth periods to calculate the actual potassium absorption. Finally, using the control group as a baseline, the nitrogen inhibition amplitude of each treatment group was calculated using a formula, which serves as the extent to which nitrogen inhibited potassium absorption under different sample nutrient parameters. For example, the average potassium and nitrogen contents of different soils are collected from historical planting records, such as (180 mg / kg, 150 mg / kg), (200 mg / kg, 170 mg / kg), etc., as a set of sample nutrient parameters. The extent to which nitrogen inhibits potassium absorption under different sample nutrient parameters is collected, such as 15% and 17%, and the set of sample element antagonism coefficients is obtained by labeling.

[0054] Secondly, an element antagonism classifier is constructed based on the sample nutrient parameter set and the sample element antagonism coefficient set. For example, the element antagonism classifier can be constructed based on the random forest algorithm, mainly consisting of a feature input layer, a decision tree ensemble layer, and an output layer. It is specifically used to quantify the antagonistic strength of nitrogen against potassium in soil. The feature input layer contains two input features: potassium content and nitrogen content, which directly reflect the current nutrient baseline state of the soil. The decision tree ensemble layer consists of 100 CART regression trees forming the main body of the forest. Each tree is constructed in the following way: using the bootstrap sampling method, 70% of the samples (including replacement) are randomly selected from the historical sample set as training data for a single tree, and the remaining 30% are used as out-of-bag (OOB) data for internal validation. When a node splits, the split threshold is calculated based on the two input features (based on the principle of minimizing mean squared error) to avoid a single feature dominating the split of all trees and improve the generalization ability of the model. The maximum depth of a single tree is set to 8 layers, and the minimum sample size of a leaf node is 5 to prevent overfitting. The output layer integrates the prediction results of 100 trees through a mean ensemble strategy. Finally, the classifier takes the arithmetic mean of the prediction values ​​of all trees as the element antagonism coefficient as the output. For example, during the model training phase, the sample nutrient parameter set is used as input and the sample element antagonism coefficient set is used as label. By adjusting hyperparameters such as the number and depth of trees, the mean squared error (MSE) of the out-of-bag data is minimized, and finally an element antagonism classifier that can accurately map nutrient parameters and antagonism coefficients is formed.

[0055] Finally, the potassium and nitrogen contents within the average nutrient parameters are input into the element antagonism classifier, and the element antagonism coefficient is obtained from the classification output. For example, the potassium and nitrogen contents within the average nutrient parameters (e.g., 200 mg / kg, 160 mg / kg) are input into the element antagonism classifier, and the element antagonism coefficient is obtained from the classification output, such as 2.5%. The element antagonism coefficient can reflect the intensity of nitrogen inhibition of potassium absorption in the soil. The value ranges from 0 to 1. The larger the element antagonism coefficient, the more significant the inhibitory effect of nitrogen on potassium absorption is under the current soil nitrogen and potassium content conditions. That is, the process of watermelon roots absorbing potassium is more strongly interfered with by nitrogen, which may lead to a reduction in the actual amount of potassium obtained by the plant due to nitrogen antagonism, even if the soil potassium content is sufficient.

[0056] In summary, compared to existing technologies, this application analyzes the nutrient error range of the nutrient parameter array, processes and obtains element antagonism coefficients, compensates for the optimization step size to obtain an optimized step size set, adjusts the first adjustment scheme to obtain a second adjustment scheme set, and continues to optimize and obtain the optimal adjustment scheme to adjust the target soil region. Thus, by integrating soil spatial differences (sample error range) and element interactions (element antagonism coefficients), and by finely compensating for the optimization step size, multiple schemes are generated and iteratively screened. This solves the problems of fixed step sizes and easy omission of optimal solutions in traditional adjustments, enabling the adjustment scheme to adapt to actual soil differences while avoiding element interaction risks, ultimately achieving precision and efficiency in nutrient stabilization and regulation.

[0057] In summary, the embodiments of this application have at least the following technical effects: Compared to existing technologies, this application first collects soil nutrient parameters from multiple monitoring points within the target soil area for continuously cropped watermelons, obtaining a nutrient parameter array and calculating the average nutrient parameters. In this way, through scientific sampling and calculation, the random errors of single-point sampling are eliminated, and benchmark data representing the overall nutrient status of the target soil area for continuously cropped watermelons is obtained, providing a reliable basis for subsequent adjustment schemes.

[0058] Secondly, this application randomly generates a first adjustment scheme for soil conditioning, combines it with the average nutrient parameters to predict the adjustment, obtains the first conditioned nutrient parameters, conducts a stability assessment, and obtains the stability fitness. Thus, by randomly generating the first adjustment scheme, combining it with the current average soil nutrient parameters to predict the adjustment effect, and quantifying its proximity to the ideal state, a quantifiable starting point and evaluation standard are provided for subsequent optimization.

[0059] Furthermore, this application obtains the continuous cropping parameters of the target soil region and configures the optimization step size. This dynamically matches the optimization step size with the actual degree of soil degradation, balancing adjustment efficiency and accuracy, and providing a reasonable search basis for subsequent iterative optimization of the scheme.

[0060] Finally, this application analyzes the nutrient error range of the nutrient parameter array, processes and obtains the element antagonism coefficient, compensates for the optimization step size, obtains an optimized step size set, adjusts the first adjustment scheme to obtain a second adjustment scheme set, and continues to optimize and obtain the optimal adjustment scheme to adjust the target soil area. Thus, by integrating soil spatial differences (sample error range) and element interactions (element antagonism coefficients), and by finely compensating for the optimization step size, multiple schemes are generated and iteratively screened. This solves the problems of fixed step sizes and easy omission of optimal solutions in traditional adjustments, enabling the adjustment scheme to adapt to actual soil differences while avoiding element interaction risks, ultimately achieving precision and efficiency in nutrient stabilization and regulation.

[0061] Through the aforementioned technical solution, this application collects data on the spatial heterogeneity of nutrients in the target soil area for continuously cropped watermelons at multiple monitoring points, accurately predicts the regulation effect using machine learning models, dynamically configures and optimizes the step size based on the continuous cropping time, and compensates and calibrates the step size using sample error amplitude and element antagonism coefficient. Finally, the optimal solution is selected through iterative optimization. This effectively overcomes the interference of nitrogen and potassium antagonism on nutrient absorption, adapts to the nutrient patterns of continuously cropped soils changing with planting years, improves the adaptability of the regulation scheme to complex soil conditions, and reduces the risk of nutrient imbalance caused by monitoring errors or blind regulation through refined step size design and multi-dimensional parameter integration. In this way, an upgrade from empirical regulation to dynamic, precise, and intelligent regulation is achieved, ultimately realizing precise regulation of soil nutrients for continuously cropped watermelons.

[0062] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0063] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0064] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0065] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0066] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0067] Although preferred embodiments of the invention have been described, those skilled in the art, once they have learned the basic inventive concept, can make other changes and modifications to these embodiments.

[0068] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A dynamic regulation method for maintaining soil nutrient stability in watermelon cultivation, characterized in that, The method includes: Soil nutrient parameters were collected from multiple monitoring points within the target soil area of ​​continuously cropped watermelons to obtain a nutrient parameter array. The average nutrient parameters were calculated, where each nutrient parameter includes the content of potassium and nitrogen. A first adjustment scheme for soil adjustment is randomly generated, and the adjustment prediction is performed in combination with the average nutrient parameters to obtain the first adjustment nutrient parameters. A stability assessment is then performed to obtain the stability fitness. In this process, a nutrient adjustment predictor based on machine learning pre-trained is used for adjustment prediction. The calculation of stability fitness includes: Obtain standard nutrient parameters; Calculate the similarity between the first adjusted nutrient parameter and the standard nutrient parameter to obtain the first homeostatic fitness. Obtain the continuous cropping parameters of the target soil area and configure the optimization step size, wherein the optimization step size includes the percentage range of adjusting the adjustment scheme; The nutrient error range of the nutrient parameter array is analyzed and the element antagonism coefficient is obtained. The optimization step size is compensated to obtain an optimization step size set. The first adjustment scheme is adjusted to obtain a second adjustment scheme set. The adjustment and optimization are continued to obtain the optimal adjustment scheme. The target soil area is adjusted. Element antagonism analysis is performed based on the average nutrient parameters to obtain the element antagonism coefficient. The optimization continues until a preset number of optimizations is reached, at which point the adjustment scheme with the highest stability and adaptability is output as the optimal adjustment scheme.

2. The dynamic regulation method for maintaining soil nutrient stability in watermelon cultivation according to claim 1, characterized in that, Soil nutrient parameters were collected from multiple monitoring points within the target soil area for continuously cropped watermelons to obtain a nutrient parameter array. The average nutrient parameters were then calculated, including: Soil nutrient parameters were collected from multiple monitoring points within the target soil area of ​​continuously cropped watermelons to obtain a nutrient parameter array, where each nutrient parameter includes the content of potassium and nitrogen. Calculate the mean of the nutrient parameter array to obtain the average nutrient parameters.

3. The dynamic regulation method for maintaining soil nutrient stability in watermelon cultivation according to claim 1, characterized in that, A first regulation scheme for soil conditioning is randomly generated. This scheme is then combined with the average nutrient parameters to predict the first regulated nutrient parameters. A stability assessment is performed to obtain the stability fitness, including: Obtain a space for soil conditioning schemes, wherein the space for soil conditioning schemes includes a potassium replenishment space and a nitrogen replenishment space; A first adjustment scheme for soil adjustment is randomly generated within the adjustment scheme space; Based on the first adjustment scheme and the average nutrient parameters, an adjustment prediction is made to obtain the first adjustment nutrient parameters; Obtain standard nutrient parameters and calculate the first homeostatic fitness of the first adjusted nutrient parameters.

4. The dynamic regulation method for maintaining soil nutrient stability in watermelon cultivation according to claim 1, characterized in that, Based on the first adjustment scheme and average nutrient parameters, adjustment prediction is performed to obtain the first adjustment nutrient parameters, including: Based on historical watermelon soil conditioning data, a set of sample conditioning schemes and a set of sample nutrient parameters were collected. The conditioned nutrient parameters under different sample conditioning schemes and sample nutrient parameters were also collected to obtain a set of sample conditioned nutrient parameters. Construct a nutrient adjustment predictor based on machine learning; The nutrient adjustment predictor is trained under supervision using the set of sample adjustment schemes, the set of sample nutrient parameters, and the set of sample adjusted nutrient parameters until the training converges.

5. The dynamic regulation method for maintaining soil nutrient stability in watermelon cultivation according to claim 1, characterized in that, Obtain the continuous cropping parameters of the target soil area and configure the optimization step size, including: Obtain the continuous cropping parameters of the target soil area, wherein the continuous cropping parameters include the continuous cropping time; The average continuous cropping time of watermelons is obtained, the ratio of the continuous cropping time to the average continuous cropping time is calculated, and the preset step size is adjusted and configured as the optimized step size. The optimized step size includes the percentage range of the adjustment scheme.

6. The dynamic regulation method for maintaining soil nutrient stability in watermelon cultivation according to claim 1, characterized in that, The nutrient error amplitude of the nutrient parameter array is analyzed, and the element antagonism coefficient is obtained. The optimization step size is compensated to obtain an optimization step size set. The first adjustment scheme is adjusted to obtain a second adjustment scheme set, and further adjustment and optimization are performed to obtain the optimal adjustment scheme, including: Calculate the average error amplitude between the nutrient parameter array and the average sample parameters, and use it as the sample error amplitude; Based on the average nutrient parameters, elemental antagonism analysis was performed to obtain the elemental antagonism coefficient; The step size compensation coefficient is calculated based on the sample error amplitude and the element antagonism coefficient. The optimized step size is compensated using the step size compensation coefficient to obtain a compensation step size interval. All optimized step sizes within the compensation step size interval are selected to obtain an optimized step size set, wherein the minimum scale of the optimized step size is one percent. Using the optimized step size set, the first adjustment scheme is adjusted to obtain the second adjustment scheme set; Continue iterative optimization until the preset number of optimizations is reached, and output the adjustment scheme with the highest stability and fitness as the optimal adjustment scheme.

7. The dynamic regulation method for maintaining soil nutrient stability in watermelon cultivation according to claim 6, characterized in that, Based on the aforementioned average nutrient parameters, elemental antagonism analysis was performed to obtain elemental antagonism coefficients, including: Based on historical watermelon planting soil conditioning data, a set of sample nutrient parameters was collected, and the extent to which nitrogen inhibited potassium absorption under different sample nutrient parameters was collected. The set of sample element antagonism coefficients was then labeled. Based on the set of sample nutrient parameters and the set of sample element antagonism coefficients, an element antagonism classifier is constructed. The potassium and nitrogen content within the average nutrient parameters is input into the element antagonism classifier, and the element antagonism coefficient is obtained by classification output.

Images