A method for optimizing straw returning and fertilization by fusing soil carbon-nitrogen ratio

Abstract

This invention relates to the field of agricultural fertilization, and more particularly to an optimized method for straw return fertilization that integrates soil carbon-nitrogen ratio. The method includes: standardizing farmland static feature data to obtain standardized feature vectors and feature standardization thresholds; performing coupling analysis on the straw return amount and soil carbon-nitrogen ratio in the standardized feature vector pool to obtain risk factors; integrating the risk potential energy field function using path gradients to obtain symmetric path cost gain factors; correcting path cost gain using Euclidean distance to obtain an adaptive distance metric; and performing fuzzy C-means clustering on the standardized farmland feature vectors using the adaptive distance metric to obtain differentiated fertilization management zones. This solves the problem that existing Euclidean distance-based management zones cannot identify key feature nonlinear coupling risks, leading to inaccurate fertilization decisions.

Description

Technical Field

[0001] This invention relates to the field of agricultural fertilization technology, and in particular to an optimized method for straw return fertilization that incorporates the soil carbon-nitrogen ratio. Background Technology

[0002] In modern precision agriculture, pre-planning management zoning is typically required to achieve differentiated and refined management of large areas of farmland. The purpose of management zoning is to group plots with similar soil properties and productivity potential into the same management unit, facilitating the application of standardized field practices. Currently, fuzzy C-means (FCM) clustering is a commonly used technique for farmland management zoning. This method collects and analyzes soil physicochemical property data from multiple sampling points within the farmland, such as soil organic matter content, total nitrogen content, available phosphorus content, available potassium content, and pH value, representing each sampling point as a feature vector composed of multidimensional static attributes. Subsequently, the FCM algorithm calculates similarity based on the Euclidean distance of each feature vector in multidimensional space, and through an iterative optimization process, groups plots with high similarity into the same category, thus forming management zoning. This type of zoning method based on static soil property data provides a fundamental basis for guiding fertilization, irrigation, and other agricultural activities under conventional conditions.

[0003] However, with the widespread application of straw return farming, the management zoning method based on the standard FCM algorithm has gradually revealed its limitations. After straw return, the nitrogen balance and crop nitrogen availability of a field depend not only on its inherent static soil properties but also on the dynamic biochemical processes dominated by straw characteristics, especially the nitrogen-scavenging effect caused by microorganisms decomposing straw. The core similarity measure of existing FCM algorithms still relies on Euclidean distance, an isotropic linear spatial measurement method that can only independently measure the numerical differences of each feature dimension and perform linear accumulation, making it difficult to capture the nonlinear coupling effects that may exist between different feature dimensions. In the actual scenario of straw return, there is a significant interaction between the soil carbon-nitrogen ratio and the amount of straw returned. When a high carbon-nitrogen ratio and a high amount of straw returned occur simultaneously in the same field, the resulting risk of microbial nitrogen scavenging shows a rapidly increasing nonlinear trend, rather than a simple superposition of two independent risks. Traditional FCM algorithms, by neglecting this type of nonlinear coupling effect, may group fields with multiple high-risk characteristics into the same partition as fields with only a single high-risk characteristic or no risk characteristic, leading to a deviation between the partitioning results and actual agronomic needs. This limitation of distance metric makes the management partitioning results of existing methods in straw return scenarios inaccurate, making it difficult to provide accurate and reliable decision-making basis for subsequent fertilization optimization. Summary of the Invention

[0004] In view of this, the present invention aims to propose an optimized method for straw return fertilization that integrates soil carbon-nitrogen ratio, in order to solve the problem that existing management zoning based on Euclidean distance cannot identify the risk of nonlinear coupling of key features, resulting in inaccurate fertilization decisions.

[0005] To achieve the above objectives, the technical solution of the present invention is implemented as follows:

[0006] An optimized method for straw return fertilization that incorporates the soil carbon-nitrogen ratio includes the following steps:

[0007] Step S1: Obtain the standardized feature vector and feature standardization threshold of farmland by standardizing the static feature data of farmland;

[0008] Step S2: Risk factors are obtained by coupling analysis of straw return amount and soil carbon-nitrogen ratio in the standardized feature vector of farmland;

[0009] Step S3: Obtain the symmetric path cost gain factor by integrating the path gradient of the risk factor;

[0010] Step S4: Obtain an adaptive distance metric between grid sampling points in the farmland by correcting the path cost gain of the Euclidean distance between grid sampling points in the farmland;

[0011] Step S5: Perform fuzzy C-means clustering on the standardized feature vector of farmland using an adaptive distance metric between grid sampling points in the farmland to obtain differentiated fertilization management zones.

[0012] Furthermore, the step of standardizing farmland static feature data to obtain standardized farmland feature vectors and feature standardization thresholds includes:

[0013] The target farmland area was divided into grids and sampling points were set in each grid. Soil samples were collected at each sampling point and the static soil characteristic data corresponding to the sampling location were recorded, including: soil organic matter content, soil total nitrogen content, straw return amount and straw type.

[0014] Soil total carbon content is obtained by approximate analysis of soil organic matter content; soil carbon-nitrogen ratio is obtained by evaluating the ratio of soil total carbon content to soil total nitrogen content.

[0015] For any target grid sampling point in the target farmland area, the straw return amount, soil carbon-nitrogen ratio, soil organic matter content, soil total nitrogen content and soil total carbon content of the target grid sampling point are Z-score standardized in each dimension in all grid sampling points, and the corresponding multidimensional feature vectors are used as the multidimensional feature vectors of the target grid sampling point.

[0016] Set thresholds for straw return amount and soil carbon-nitrogen ratio, and then perform Z-score standardization on the straw return amount feature dimension and soil carbon-nitrogen ratio feature dimension, respectively, to obtain the standardized thresholds for straw return amount and soil carbon-nitrogen ratio.

[0017] Furthermore, the risk factors are obtained by coupling analysis of the amount of straw returned to the field and the soil carbon-nitrogen ratio in the standardized feature vector pool of farmland, including:

[0018] By performing threshold integration on standardized characteristic data of straw return to the field and soil basic carbon-nitrogen ratio, a nonlinear risk response assessment is obtained; by performing feature coupling analysis on the nonlinear risk response assessment, risk factors are obtained.

[0019] Furthermore, the process of obtaining a nonlinear risk response assessment by performing threshold integration on standardized characteristic data of straw return to the field and the soil's baseline carbon-nitrogen ratio includes:

[0020] For any target grid sampling point in the target farmland area, the standardized threshold of straw return to the field is divided by the square of the calculated straw return to the field at the target grid sampling point as the straw return to the field threshold difference assessment of the target grid sampling point; the reciprocal of the sum of the constant 1 and the straw return to the field threshold difference assessment of the target grid sampling point is used as the state transformation function of the straw return to the field at the target grid sampling point; the integral of the state transformation function of the straw return to the field at the target grid sampling point over the range from the constant 0 to the straw return to the field at the target grid sampling point is used as the straw return to the field risk response assessment of the target grid sampling point.

[0021] The standard threshold of soil carbon-nitrogen ratio is divided by the square of the calculated soil carbon-nitrogen ratio of the target grid sampling point as the threshold difference assessment of soil carbon-nitrogen ratio of the target grid sampling point; the reciprocal of the sum of constant 1 and the threshold difference assessment of soil carbon-nitrogen ratio of the target grid sampling point is used as the state transformation function of soil carbon-nitrogen ratio of the target grid sampling point; the integral of the state transformation function of soil carbon-nitrogen ratio of the target grid sampling point as the integrand over the range from constant 0 to the numerical range of soil carbon-nitrogen ratio of the target grid sampling point is used as the risk response assessment of soil carbon-nitrogen ratio of the target grid sampling point.

[0022] The result of multiplying the risk response assessment of straw return to the field at the target grid sampling point with the risk response assessment of soil carbon-nitrogen ratio at the target grid sampling point is used as the nonlinear risk response assessment of the target grid sampling point.

[0023] Furthermore, the process of obtaining risk factors through feature coupling analysis of nonlinear risk response assessment includes:

[0024] Set a risk gain coefficient; multiply the risk gain coefficient by the nonlinear risk response assessment of the target grid sampling point and add the result to a constant 1 as the risk factor of the target grid sampling point.

[0025] Furthermore, the step of obtaining the symmetric path cost gain factor by integrating the path gradient of the risk potential energy field function includes:

[0026] By performing gradient calculation on the risk factors, the risk gradient vector of each sampling point on the path is obtained; by integrating the magnitude of the risk gradient vector on the path, the cumulative risk gradient value of the path is obtained; by performing path scale normalization on the cumulative risk gradient value, the symmetric path cost gain factor is obtained.

[0027] Furthermore, the step of obtaining the risk gradient vector of each sampling point on the path by performing gradient calculation on the risk factors includes:

[0028] For any target grid sampling point in the target farmland area, the partial derivatives of the risk factor with respect to the characteristic dimensions of straw return amount and soil carbon-nitrogen ratio are calculated as gradient components on the characteristic dimensions. The gradient components in the characteristic dimensions of soil organic matter content, soil total nitrogen content and soil total carbon content are set to 0. The gradient components of each characteristic dimension are combined according to the characteristic dimensions to form a vector, which is used as the risk gradient vector of the target grid sampling point.

[0029] Furthermore, the step of obtaining the cumulative risk gradient value of the path by integrating the magnitude of the risk gradient vector along the path includes:

[0030] Set the integration step size; during the clustering of grid sampling points, obtain all cluster centers during the clustering iteration process; on the connection path between the target grid sampling point and any target cluster center, divide the path into multiple continuous sampling positions according to the set integration step size; calculate the magnitude of the risk gradient vector at each sampling position, and use the sum of the magnitudes of the risk gradient vectors at each sampling position as the cumulative risk gradient value of the path.

[0031] Furthermore, the step of obtaining the symmetric path cost gain factor by performing path scale normalization on the cumulative risk gradient value includes:

[0032] Set the path cost gain coefficient; set the baseline distance scale parameter by the Euclidean distance between non-repeating point pairs in all grid sampling points; use the result of adding the square of the Euclidean distance between the target grid sampling point and the target cluster center to the baseline distance scale parameter as the denominator of the path scale normalization; multiply the cumulative risk gradient value by the Euclidean distance between the target grid sampling point and the target cluster center as the numerator of the path scale normalization; multiply the corresponding fraction by the path cost gain coefficient and add it to the constant 1 as the symmetric path cost gain factor.

[0033] Furthermore, the step of obtaining an adaptive distance metric by correcting the path cost gain of the Euclidean distance includes:

[0034] The Euclidean distance between the target grid sampling point and the target cluster center is multiplied by the symmetric path cost gain factor between the target grid sampling point and the target cluster center, and used as the adaptive distance metric between the target grid sampling point and the target cluster center.

[0035] Compared with the prior art, the present invention has the following advantages:

[0036] This invention presents an optimized straw return fertilization method that integrates soil carbon-nitrogen ratio. By introducing a risk potential energy field and path cost gain factor, it enables the identification of the nonlinear coupling effect between straw return amount and soil carbon-nitrogen ratio during farmland zoning. In actual straw return scenarios, when a field simultaneously has high straw load and a high carbon-nitrogen ratio, the microbial decomposition process significantly intensifies nitrogen competition, making it difficult for traditional Euclidean distance-based clustering methods to accurately identify this risk. The technical approach proposed in this invention can form a sensitive response to risk thresholds in the feature space and, through dynamic correction of the risk gradient, more accurately characterize the differences between different fields according to agronomic laws, thereby avoiding the misclassification of high-risk and low-risk areas. At the application level, this invention not only improves the scientificity and accuracy of farmland management zoning but also directly outputs differentiated fertilization recommendations. High-coupling-risk areas can be identified and receive nitrogen fertilizer compensation strategies, high straw load areas can receive nitrogen protection during the initial stage, and areas with limited soil nitrogen supply can receive long-term improvement schemes. By linking zoning and fertilization strategies, this invention provides farmers with a comprehensive fertilization optimization method that takes into account both short-term crop needs and long-term soil health, thereby improving crop yield and quality and promoting the sustainable use of soil resources. Attached Figure Description

[0037] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0038] Figure 1 This is a flowchart illustrating an optimized method for straw return fertilization based on soil carbon-nitrogen ratio, as described in an embodiment of the present invention. Detailed Implementation

[0039] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0040] See Figure 1 This is a flowchart of a method for optimizing straw return fertilization based on soil carbon-nitrogen ratio, as provided in Embodiment 1 of the present invention. Figure 1 As shown, an optimized method for straw return fertilization that incorporates the soil carbon-nitrogen ratio may include:

[0041] Step S1: Standardize the farmland static feature data to obtain the farmland standardized feature vector and feature standardization threshold.

[0042] This step aims to provide farmland characteristic data at a uniform scale for subsequent risk assessment and cluster analysis. First, the target farmland area is divided into grids, and sampling points are set in each grid. Soil samples are collected at each sampling point, and the corresponding static soil characteristic data are recorded, including: soil organic matter content, total nitrogen content, straw return amount, and straw type. Approximate analysis is performed using soil organic matter content to obtain the total carbon content. The ratio of total carbon content to total nitrogen content is evaluated to obtain the soil carbon-nitrogen ratio. For any target grid sampling point in the target farmland area, the straw return amount, soil carbon-nitrogen ratio, soil organic matter content, total nitrogen content, and total carbon content of the target grid sampling point are Z-score standardized across all grid sampling points, and the resulting multidimensional feature vectors are used as the multidimensional feature vectors of the target grid sampling point.

[0043] In this embodiment of the invention, the threshold for straw return to the field and the threshold for soil carbon-nitrogen ratio are set as follows: the threshold for straw return to the field is set to 600 kg per mu, and the threshold for soil carbon-nitrogen ratio is set to 12. The thresholds for straw return to the field and soil carbon-nitrogen ratio can be adjusted according to the differences in actual farmland scenarios, and are not required. The thresholds for straw return to the field and soil carbon-nitrogen ratio are Z-score standardized in the straw return to the field feature dimension and the soil carbon-nitrogen ratio feature dimension, respectively, to obtain the standardized thresholds for straw return to the field and soil carbon-nitrogen ratio.

[0044] This completes the standardization process of farmland static feature data, resulting in the acquisition of standardized feature vectors and feature standardization thresholds.

[0045] Step S2: Risk factors are obtained by coupling analysis of the amount of straw returned to the field and the soil carbon-nitrogen ratio in the standardized feature vector of farmland.

[0046] In clustering calculations for farmland management zones, existing technologies typically use the original static feature values ​​of each plot for distance measurement, treating the multidimensional feature space as uniform. However, in the specific application scenario of straw return to the field, the potential for nitrogen loss due to decomposition in a plot is determined by the coupling relationship between the amount of straw returned and the soil's basic carbon-nitrogen ratio. Existing distance measurement technologies, because they treat the contributions of each feature dimension as independent, cannot effectively identify the coupling effects produced by specific combinations of feature values. Therefore, to construct a similarity measure that better reflects agronomical characteristics, the first optimization of this invention is to construct a risk potential energy field in the entire feature space. The construction logic of this risk potential energy field is that the potential energy value at any point in the space should reflect the specific combination state formed when the straw return amount and the soil's basic carbon-nitrogen ratio are both at a high level. By constructing such a potential energy field function, a basic reference is provided for subsequent distance calculations of risky terrain.

[0047] In summary, firstly, a nonlinear risk response assessment is obtained by performing threshold integration on the standardized characteristic data of straw return amount and soil baseline carbon-nitrogen ratio. Specifically, for any target grid sampling point in the target farmland area, the standardized threshold of straw return amount is divided by the square of the calculated straw return amount of the target grid sampling point as the straw return amount threshold difference assessment for the target grid sampling point; the reciprocal of the sum of constant 1 and the straw return amount threshold difference assessment for the target grid sampling point is used as the state transformation function of straw return amount at the target grid sampling point; and the integral of the state transformation function of straw return amount at the target grid sampling point as the integrand over the range from constant 0 to the straw return amount value of the target grid sampling point is used as the straw return amount risk of the target grid sampling point. Response assessment: The standard threshold of soil carbon-nitrogen ratio is divided by the square of the calculated soil carbon-nitrogen ratio of the target grid sampling point as the threshold difference assessment of soil carbon-nitrogen ratio of the target grid sampling point; the reciprocal of the sum of constant 1 and the threshold difference assessment of soil carbon-nitrogen ratio of the target grid sampling point is used as the state transformation function of soil carbon-nitrogen ratio of the target grid sampling point; the integral of the state transformation function of soil carbon-nitrogen ratio of the target grid sampling point as the integrand over the range of constant 0 to the numerical range of soil carbon-nitrogen ratio of the target grid sampling point is used as the risk response assessment of soil carbon-nitrogen ratio of the target grid sampling point; the result of multiplying the risk response assessment of straw return to the field of the target grid sampling point with the risk response assessment of soil carbon-nitrogen ratio of the target grid sampling point is used as the nonlinear risk response assessment of the target grid sampling point.

[0048] In one implementation, assume the first The standardized value of straw return to the field at each grid sampling point is: ;No. The standardized values ​​of soil carbon-nitrogen ratio at each grid sampling point are: The standardized threshold for straw return to the field is: The standardized threshold for soil carbon-nitrogen ratio is: Then the first The calculation expression for the nonlinear risk response assessment of a grid sampling point is as follows:

[0049]

[0050] in, Indicates the first Nonlinear risk response assessment of individual grid sampling points; Indicates the first Standardized values ​​of straw return to the field at each grid sampling point; Indicates the first Standardized values ​​of soil carbon-nitrogen ratio at each grid sampling point; This represents the standardized threshold for the amount of straw returned to the field. This represents the standardized threshold for the soil carbon-nitrogen ratio; This represents the standardized value of the amount of straw returned to the field in the integrand; This represents the standardized value of the soil carbon-nitrogen ratio in the integrand.

[0051] After obtaining the nonlinear risk response assessment of the target grid sampling point, the risk factor is obtained by performing feature coupling analysis on the nonlinear risk response assessment. Specifically, a risk gain coefficient is set. In this embodiment of the invention, the risk gain coefficient is set to 1, which can be adjusted according to the actual scenario and is not required. The result of multiplying the risk gain coefficient by the nonlinear risk response assessment of the target grid sampling point and adding it to the constant 1 is used as the risk factor of the target grid sampling point.

[0052] In one implementation, it is assumed that the risk gain coefficient is Then the first The expression for calculating the risk factor for each grid sampling point is:

[0053]

[0054] in, Indicates the first Risk factors for each grid sampling point; Indicates the risk gain coefficient; Indicates the first Nonlinear risk response assessment of individual grid sampling points.

[0055] It should be noted that in the actual scenario of straw returning to the field, existing technologies directly use the original feature values ​​for distance calculation, which leads to two specific problems: First, it incorrectly assumes that risk increases linearly with the increase of feature values; second, it fails to identify the coupling effect when multiple risk factors coexist. To address the nonlinearity problem in risk assessment, this invention does not directly use feature values ​​but instead constructs a feature state transformation term, namely... and In real-world scenarios, when the amount of straw returned to the field or the soil carbon-nitrogen ratio is below the safe threshold, its contribution to the risk of nitrogen loss due to decomposition is negligible; however, once this threshold is exceeded, the risk increases dramatically. The construction of the integral term in the above formula is precisely to capture this nonlinear situation. The integrand in the integral term ensures that when the eigenvalue is much smaller than the threshold, its contribution to the integral area is minimal, while when the eigenvalue exceeds the threshold, its contribution to the integral area increases rapidly. Therefore, by calculating the cumulative area of ​​this integral, a linear original eigenvalue is transformed into an indicator that can truly reflect the degree of risk accumulation, thereby solving the problem of inaccurate assessment of a single risk factor. Secondly, to address the coupling problem in risk assessment—specifically, identifying the exacerbated risk arising from the simultaneous occurrence of high straw content and high soil carbon-nitrogen ratio—existing technologies treat the risks of different characteristics as independent and additive, which is inconsistent with reality. This invention employs a feature coupling term, multiplying the results of the above-mentioned integral transformations of two key features. This ensures that in nitrogen-rich soil (low carbon-nitrogen ratio), even with the addition of large amounts of straw, the risk of nitrogen depletion remains controllable; conversely, in nitrogen-poor soil (high carbon-nitrogen ratio), the addition of large amounts of straw will trigger catastrophic nitrogen competition. Mathematically, this multiplicative structure ensures that only when the values ​​of both the straw return amount and the soil's baseline carbon-nitrogen ratio corresponding to a grid sampling point simultaneously exceed their respective risk thresholds will the risk be mitigated. and This design ensures that only when both integral terms are large will their product yield a value significantly higher than in other cases. If only a single feature is at a high-risk level, its corresponding integral term, though large, will still result in a low-risk level after multiplication with a smaller integral term. Through this design, the risk factor of the final calculated grid sampling points can effectively identify and quantify the coupled risk formed by the combined effect of two key risk factors.

[0056] Thus, the risk factors were obtained by coupling analysis of the amount of straw returned to the field and the soil carbon-nitrogen ratio in the standardized feature vector of farmland.

[0057] Step S3: Obtain the symmetric path cost gain factor by performing path gradient integration on the risk factor.

[0058] In step S2, this invention has defined a non-uniform risk potential energy field in the multidimensional feature space by constructing risk factors, which can reflect the degree of inherent coupling risk of each field. However, if the clustering is simply modified based on the risk factors of each point, there are still two problems: First, since the risk factors map multidimensional features to a one-dimensional scalar, there may be fields with different feature combinations whose calculated risk factors are exactly the same, causing potential information confusion; Second, the standard distance metric itself is spatially uniform, and it fails to reflect that the similarity metric between fields should have different sensitivities in different regions of the risk potential energy field.

[0059] Therefore, the second optimization step of this invention aims to solve the above-mentioned problems. Its core logic is that the true agronomical similarity between any target grid sampling point and any target cluster center should not be determined solely by their respective risk factors, but rather by the complexity of the terrain traversed by the path connecting these two points in the risk potential field. A path traversing a region of drastic risk change (i.e., a large risk gradient) implies greater differences in agronomic management between the two points, and their similarity should be correspondingly reduced (i.e., the distance should be amplified). To achieve this, this invention constructs a symmetric path gain factor. This factor quantitatively measures the ruggedness of the path by integrating the gradient magnitude of the risk factors at each point on the straight path between any target grid sampling point and any target cluster center. This ruggedness is an inherent property of the path, independent of direction, thus fundamentally ensuring the symmetry of the final distance measurement. Through this design, even if two fields with different origins happen to have the same risk factor, because their positions in the feature space are different, the connecting paths to the same cluster center are also different, and the risk terrain traversed is naturally different, resulting in different calculated path ruggedness. This not only solves the problem of information confusion between fields with the same score but different qualities, but more importantly, it enables the final distance measurement to truly reflect the real differences between fields caused by crossing different risk areas.

[0060] In summary, the risk gradient vectors of each sampling point along the path are first obtained by performing gradient calculation on the risk factors. Specifically, for any target grid sampling point in the target farmland area, the partial derivatives of the risk factors with respect to the straw return amount feature dimension and the soil carbon-nitrogen ratio feature dimension are calculated as the gradient components of that feature dimension. The gradient components in the feature dimensions of soil organic matter content, soil total nitrogen content, and soil total carbon content are set to 0. The gradient components of each feature dimension are combined according to the feature dimension to form a vector, which is used as the risk gradient vector of the target grid sampling point.

[0061] Then, the cumulative risk gradient value of the path is obtained by integrating the magnitude of the risk gradient vector on the path. Specifically, an integration step size is set. In this embodiment of the invention, the integration step size is set to the number of grid sampling points between the target grid sampling point and the target cluster center point. Thus, an approximate evaluation is performed by accumulating the magnitudes of the risk gradient vectors of the grid sampling points between the target grid sampling point and the target cluster center point. During the clustering process of the grid sampling points, all cluster centers in the clustering iteration process are obtained. On the connection path between the target grid sampling point and any target cluster center, the path is divided into multiple continuous sampling positions according to the set integration step size. The magnitude of the risk gradient vector is calculated at each sampling position, and the result of summing the magnitudes of the risk gradient vectors at each sampling position is used as the cumulative risk gradient value of the path.

[0062] Finally, the cumulative risk gradient value is normalized by path scale to obtain the symmetric path cost gain factor. Specifically, the path cost gain coefficient is set to 1 in this embodiment, but it can be adjusted according to the actual scenario and is not required. The baseline distance scale parameter is set by the Euclidean distance between non-repeating point pairs in all grid sampling points. In this embodiment, the squares of the Euclidean distances between all grid sampling point pairs are accumulated, and the arithmetic mean of the accumulated calculation results is used as the baseline distance scale parameter. The result of adding the square of the Euclidean distance between the target grid sampling point and the target cluster center to the baseline distance scale parameter is used as the denominator of the path scale normalization. The cumulative risk gradient value is multiplied by the Euclidean distance between the target grid sampling point and the target cluster center, which is used as the numerator of the path scale normalization. The result of multiplying the corresponding fraction by the path cost gain coefficient and adding it to the constant 1 is used as the symmetric path cost gain factor.

[0063] In one implementation, the path cost gain coefficient is assumed to be... ;No. The risk gradient vector of each grid sampling point is: ;No. The grid sampling point and the first The first cluster center point Each grid sampling point is The reference distance scale parameter is: Then the first The grid sampling point and the first The expression for calculating the symmetric path cost gain factor between cluster centroids is:

[0064]

[0065] in, Indicates the first The grid sampling point and the first Symmetric path cost gain factor between cluster centroids; This represents the path cost gain coefficient; Indicates the first The grid sampling point and the first The first cluster center point Risk gradient vector of each grid sampling point; Indicates the first Multidimensional feature vectors of each grid sampling point; Indicates the first Multidimensional feature vectors of cluster center points; Indicates the first The cluster center and the first Euclidean distance between grid sampling points; Indicates the first The grid sampling point and the first The first cluster center point The magnitude of the risk gradient vector of each grid sampling point.

[0066] It should be noted that after step S2 addresses the quantification of the inherent risk of a single field, this step aims to solve the problem of how to correct the similarity measure between two points based on the risk terrain they occupy. Existing Euclidean distance methods, like measurements on a flat map, assume a uniform feature space and fail to reflect the inherent agronomical differences between different regions. The symmetric path cost gain factor constructed in this invention introduces a perception of risk terrain into this similarity measure through the path integral structure. Firstly, the core of the formula is the path gradient modular integral term, i.e. This item quantifies the connection of the first The grid sampling point and the first The sum of the risk-changing regions traversed by the straight-line paths of the cluster centers is calculated using the path gradient modulo integral in a discrete summation form in this embodiment of the invention. Gradient The magnitude of the gradient magnitude characterizes the steepness of the risk potential field. A path traversing a region of gentle risk (small gradient magnitude) will have a small integral value, indicating that the two points are agronomically located within a relatively homogeneous risk plateau, and their similarity is high. Conversely, a path traversing a region of drastic risk change (large gradient magnitude) will have a large integral value, indicating that classifying the two points into the same category requires crossing a significant risk gradient, thus their similarity should be reduced (i.e., the distance should be increased). Next, the cumulative risk gradient value is multiplied by the Euclidean distance between the target grid sampling points and the target cluster centers to obtain the total path risk exposure. This ensures a measure of similarity while considering the path's ruggedness and length. Finally, a scale adjustment term is introduced. The gain effect is normalized. In summary, the symmetric path cost gain factor makes the final distance assessment no longer just a reflection of geometric distance, but a comprehensive measure that incorporates risk gradient information along the path, and better reflects the actual agronomic differences. When comparing two points, if the path connecting them is flat, the symmetric path cost gain factor is close to 1, and the distance is mainly determined by Euclidean distance; if the path is rugged, the symmetric path cost gain factor will be significantly greater than 1, thus amplifying the distance between them and reducing their similarity.

[0067] This completes the process of obtaining the symmetric path cost gain factor by integrating the path gradient of the risk factor.

[0068] Step S4: By correcting the path cost gain of the Euclidean distance between grid sampling points in the farmland, an adaptive distance metric between grid sampling points in the farmland is obtained.

[0069] After obtaining the symmetric path cost gain factor between the target grid sampling point and the target cluster center, the path cost gain is further corrected on the Euclidean distance using the symmetric path cost gain factor to obtain an adaptive distance metric. Specifically, the Euclidean distance between the target grid sampling point and the target cluster center is multiplied by the symmetric path cost gain factor between the target grid sampling point and the target cluster center to obtain the adaptive distance metric between the target grid sampling point and the target cluster center.

[0070] Thus, the adaptive distance metric between grid sampling points in farmland is obtained by correcting the path cost gain of the Euclidean distance between the grid sampling points in the farmland.

[0071] Step S5: Use the adaptive distance metric between grid sampling points in the farmland to perform fuzzy C-means clustering on the standardized feature vector of the farmland to obtain differentiated fertilization management zones.

[0072] In this embodiment of the invention, the target number of clusters is set to 6, that is, the farmland grid sampling points are divided into 6 clusters through the clustering process. The membership matrix is ​​updated in the fuzzy C-means clustering process by the adaptive distance metric between the target grid sampling points and the target cluster centers, and the clustering process is completed to obtain the optimized partitioning results of the grid sampling points in the farmland.

[0073] After obtaining the optimized zoning results of the grid sampling points in the farmland, statistical analysis is first performed on all plots within each zoning to calculate the representative characteristics of that zoning. The average straw return amount and average soil basic carbon-nitrogen ratio are calculated for all plots within that zoning. Subsequently, based on the representative characteristics of each zoning, targeted fertilization management recommendations are generated. For a zoning where members generally exhibit a risk of coupled high straw return and high soil basic carbon-nitrogen ratio, fertilization should focus on applying a certain amount of fast-acting nitrogen fertilizer (such as urea) in the early stages of crop sowing to compensate for the temporary fixation of soil nitrogen by microorganisms during the decomposition of high carbon-nitrogen ratio organic matter, thereby avoiding nitrogen starvation in the seedling stage. For zoning identified as having insignificant risk characteristics and balanced soil fertility, a conventional balanced fertilization scheme based on target yield is adopted. Furthermore, the method of this invention can also identify two other important intermediate zoning types:

[0074] Firstly, there is the high straw load area, characterized by a significantly higher amount of straw returned to the field than the average, but a lower basic carbon-nitrogen ratio in the soil. This indicates that the soil's basic nitrogen supply capacity is still acceptable, but due to excessive straw input, there is still strong short-term nitrogen depletion pressure. For this area, the amount of readily available nitrogen used as seed fertilizer or base fertilizer should be moderately increased to ensure that the crop start-up period is not affected, but there is no need to significantly increase the total nitrogen amount as in high coupling risk areas.

[0075] Secondly, there are areas with limited soil nitrogen supply, characterized by a significantly higher-than-average basic carbon-nitrogen ratio, but with low to medium levels of straw return to the field. This indicates that the main problem in this area is not the straw returned in the current season, but rather the low nitrogen availability of the soil's organic matter, resulting in poor basic nitrogen supply capacity. For this area, fertilization should not only involve adjusting the amount of nitrogen fertilizer applied in the current season, but also include the combined application of low-carbon-nitrogen ratio organic fertilizers or slow-release nitrogen fertilizers to improve soil properties and enhance basic nitrogen supply capacity in the long term.

[0076] Thus, the fuzzy C-means clustering of standardized feature vectors of farmland using an adaptive distance metric between grid sampling points in the farmland was completed, resulting in differentiated fertilization management zones.

[0077] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An optimized method for straw return fertilization that incorporates soil carbon-nitrogen ratio, characterized in that: Step S1: Obtain the standardized feature vector and feature standardization threshold of farmland by standardizing the static feature data of farmland; Step S2: Risk factors are obtained by coupling analysis of straw return amount and soil carbon-nitrogen ratio in the standardized feature vector of farmland; Step S3: Obtain the symmetric path cost gain factor by integrating the path gradient of the risk factor; Step S4: Obtain an adaptive distance metric between grid sampling points in the farmland by correcting the path cost gain of the Euclidean distance between grid sampling points in the farmland; Step S5: Perform fuzzy C-means clustering on the standardized feature vector of farmland using an adaptive distance metric between grid sampling points in the farmland to obtain differentiated fertilization management zones; The step of obtaining the symmetric path cost gain factor by integrating the path gradient of the risk factor includes: obtaining the risk gradient vector of each sampling point on the path by performing gradient calculation on the risk factor. The cumulative risk gradient value of the path is obtained by integrating the magnitude of the risk gradient vector on the path; the symmetric path cost gain factor is obtained by normalizing the cumulative risk gradient value by the path scale.

2. The method for optimizing straw return fertilization based on soil carbon-nitrogen ratio according to claim 1, characterized in that, The process of standardizing farmland static feature data to obtain standardized farmland feature vectors and feature standardization thresholds includes: The target farmland area was divided into grids and sampling points were set in each grid. Soil samples were collected at each sampling point and the static soil characteristic data corresponding to the sampling location were recorded, including: soil organic matter content, soil total nitrogen content, straw return amount and straw type. Soil total carbon content is obtained by approximate analysis of soil organic matter content; soil carbon-nitrogen ratio is obtained by evaluating the ratio of soil total carbon content to soil total nitrogen content. For any target grid sampling point in the target farmland area, the straw return amount, soil carbon-nitrogen ratio, soil organic matter content, soil total nitrogen content and soil total carbon content of the target grid sampling point are Z-score standardized in each dimension in all grid sampling points, and the corresponding multidimensional feature vectors are used as the multidimensional feature vectors of the target grid sampling point. Set thresholds for straw return amount and soil carbon-nitrogen ratio, and then perform Z-score standardization on the straw return amount feature dimension and soil carbon-nitrogen ratio feature dimension, respectively, to obtain the standardized thresholds for straw return amount and soil carbon-nitrogen ratio.

3. The method for optimizing straw return fertilization based on soil carbon-nitrogen ratio according to claim 1, characterized in that, The risk factors are obtained by coupling analysis of the amount of straw returned to the field and the soil carbon-nitrogen ratio in the standardized feature vector of farmland, including: By performing threshold integration on standardized characteristic data of straw return to the field and soil basic carbon-nitrogen ratio, a nonlinear risk response assessment is obtained; by performing feature coupling analysis on the nonlinear risk response assessment, risk factors are obtained.

4. The method for optimizing straw return fertilization based on soil carbon-nitrogen ratio according to claim 3, characterized in that, The process of obtaining a nonlinear risk response assessment by threshold integration of standardized characteristic data on straw return to the field and the soil's baseline carbon-nitrogen ratio includes: For any target grid sampling point in the target farmland area, the standardized threshold of straw return to the field is divided by the square of the calculated straw return to the field at the target grid sampling point as the straw return to the field threshold difference assessment of the target grid sampling point; the reciprocal of the sum of the constant 1 and the straw return to the field threshold difference assessment of the target grid sampling point is used as the state transformation function of the straw return to the field at the target grid sampling point; the integral of the state transformation function of the straw return to the field at the target grid sampling point over the range from the constant 0 to the straw return to the field at the target grid sampling point is used as the straw return to the field risk response assessment of the target grid sampling point. The standard threshold of soil carbon-nitrogen ratio is divided by the square of the calculated soil carbon-nitrogen ratio of the target grid sampling point as the threshold difference assessment of soil carbon-nitrogen ratio of the target grid sampling point; the reciprocal of the sum of constant 1 and the threshold difference assessment of soil carbon-nitrogen ratio of the target grid sampling point is used as the state transformation function of soil carbon-nitrogen ratio of the target grid sampling point; the integral of the state transformation function of soil carbon-nitrogen ratio of the target grid sampling point as the integrand over the range from constant 0 to the numerical range of soil carbon-nitrogen ratio of the target grid sampling point is used as the risk response assessment of soil carbon-nitrogen ratio of the target grid sampling point. The result of multiplying the risk response assessment of straw return to the field at the target grid sampling point with the risk response assessment of soil carbon-nitrogen ratio at the target grid sampling point is used as the nonlinear risk response assessment of the target grid sampling point.

5. The method for optimizing straw return fertilization based on soil carbon-nitrogen ratio according to claim 3, characterized in that, The method of obtaining risk factors through feature coupling analysis of nonlinear risk response assessment includes: Set a risk gain coefficient; multiply the risk gain coefficient by the nonlinear risk response assessment of the target grid sampling point and add the result to a constant 1 as the risk factor of the target grid sampling point.

6. The method for optimizing straw return fertilization based on soil carbon-nitrogen ratio according to claim 1, characterized in that, The step of obtaining the risk gradient vector of each sampling point on the path by performing gradient calculation on the risk factors includes: For any target grid sampling point in the target farmland area, the partial derivatives of the risk factor with respect to the characteristic dimensions of straw return amount and soil carbon-nitrogen ratio are calculated as gradient components on the characteristic dimensions. The gradient components in the characteristic dimensions of soil organic matter content, soil total nitrogen content and soil total carbon content are set to 0. The gradient components of each characteristic dimension are combined according to the characteristic dimensions to form a vector, which is used as the risk gradient vector of the target grid sampling point.

7. The method for optimizing straw return fertilization based on soil carbon-nitrogen ratio according to claim 1, characterized in that, The step of obtaining the cumulative risk gradient value of the path by integrating the magnitude of the risk gradient vector on the path includes: Set the integration step size; during the clustering of grid sampling points, obtain all cluster centers during the clustering iteration process; on the connection path between the target grid sampling point and any target cluster center, divide the path into multiple continuous sampling positions according to the set integration step size; calculate the magnitude of the risk gradient vector at each sampling position, and use the sum of the magnitudes of the risk gradient vectors at each sampling position as the cumulative risk gradient value of the path.

8. The method for optimizing straw return fertilization based on soil carbon-nitrogen ratio according to claim 1, characterized in that, The step of obtaining a symmetric path cost gain factor by normalizing the cumulative risk gradient value through path scaling includes: Set the path cost gain coefficient; set the baseline distance scale parameter by the Euclidean distance between non-repeating point pairs in all grid sampling points; use the result of adding the square of the Euclidean distance between the target grid sampling point and the target cluster center to the baseline distance scale parameter as the denominator of the path scale normalization; multiply the cumulative risk gradient value by the Euclidean distance between the target grid sampling point and the target cluster center as the numerator of the path scale normalization; multiply the corresponding fraction by the path cost gain coefficient and add it to the constant 1 as the symmetric path cost gain factor.

9. The method for optimizing straw return fertilization based on soil carbon-nitrogen ratio according to claim 1, characterized in that, The process of obtaining an adaptive distance metric between grid sampling points in farmland by correcting the path cost gain of the Euclidean distance between them includes: The Euclidean distance between the target grid sampling point and the target cluster center is multiplied by the symmetric path cost gain factor between the target grid sampling point and the target cluster center, and used as the adaptive distance metric between the target grid sampling point and the target cluster center.

Images