Rapid analysis method of fertilizer application amount based on grass yield

Abstract

The present invention discloses a method for rapid analysis of fertilizer application based on grass yield, and specifically relates to the technical field of increasing plant yield. The processing method comprises the following steps: monitoring and collecting information related to soil properties in a target area, and using an artificial neural network algorithm to construct a model using crop yield data and historical fertilizer application data. The present invention is based on a method for rapid analysis of fertilizer application based on grass yield, evaluates soil properties in a target area, and constructs a model for soil suitable for plants required by users through an artificial neural network algorithm based on the evaluation results and using crop yield data and historical fertilizer application data as factors, outputs fertilizer application data with the highest grass yield, and optimizes and adjusts the obtained model to make the acquired data more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of increasing plant yield, and more particularly to a method for quickly analyzing fertilizer application amount based on grass yield. Background Art

[0002] In Northwest my country and some special provinces, due to the existence of animal husbandry, herders need to increase the production of specific herbs in order to obtain more profits from cattle and sheep breeding. For local governments, increasing forage production also has the advantages of protecting the ecological environment, improving animal husbandry production efficiency, promoting sustainable agricultural development, improving soil health and responding to climate change.

[0003] On the premise that the environment is basically normal and stable, achieving an increase in the yield of specific herbaceous plants is usually determined by a variety of factors, such as the tolerance of the herbaceous plants, soil conditions, and the amount of fertilizer applied. Some soils are not suitable for the growth of herbaceous plants required by users due to reasons such as pH, nutrient content, and soil texture. At this time, it is difficult or even impossible to increase grass yields by simply changing the amount of fertilizer applied. For soils suitable for the production of herbaceous plants that meet user needs, excessive fertilization can easily cause seedling burn and lead to plant death, while too little fertilization can lead to plant malnutrition and affect the ecological development of animal husbandry in the entire region. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a rapid analysis method of fertilizer application amount based on grass yield, evaluates the soil properties of the target area, and builds a model of the soil of plants suitable for user needs through an artificial neural network algorithm based on the evaluation results and crop yield data and historical fertilizer application amounts as factors, outputs fertilizer application amount data with the highest grass yield, and optimizes and adjusts the obtained model to make the acquired data more accurate.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] The rapid analysis method of fertilizer application amount based on grass yield includes the following steps:

[0007] Step S1, monitoring and collecting information related to soil properties in the target area, as well as crop yield data and historical fertilizer application data when building a model using an artificial neural network algorithm;

[0008] Step S2, constructing a soil test evaluation model to evaluate whether the soil in the target area can achieve increased yield by changing the amount of fertilizer applied when planting the herbaceous plants required by the user;

[0009] Step S3, for soil in the target area where yield increase can be achieved by changing the amount of fertilizer applied, an artificial neural network model is constructed using crop yield data, soil test evaluation index and historical fertilizer application amount as factors, and the fertilizer application amount data with the highest grass yield is output;

[0010] Step s4, use the optimization algorithm to adjust the weights to adjust the parameters of the trained model to improve the accuracy of the model.

[0011] Specifically, in step S1, the artificial neural network can use a large number of input variables to improve the accuracy of prediction, and the TensorFlow, Keras and PyTorch libraries in Python are commonly used to build and train the model. In actual use, due to factors such as equipment cost and complexity of real-time monitoring, the present invention is based on the premise that the soil and required plant samples in the target area are sufficient and the growth environment is relatively stable, and the soil and required plant data are continuously monitored, which may lead to the phenomenon of rationality change at a certain monitoring moment.

[0012] In step S2, a soil test evaluation model is constructed. The specific process is to collect nutrient availability information, plant absorption capacity information and soil property information in the target area, wherein the nutrient availability information includes the nutrient availability coefficient, which is calibrated as Y yx , the plant absorption capacity information includes the plant absorption capacity coefficient, calibrated as Z xs , soil property information includes soil property coefficient, calibrated as T xz ;

[0013] Nutrient effectiveness coefficient Y yx By measuring the content of elements (such as nitrogen, phosphorus, and potassium) that are effective for plant growth in the soil of the target area, randomly collected soils are mixed together to obtain the weight of the effective element content, which is calibrated as H ys , the soil mass of the statistical mixture is calibrated as R z1 , then the nutrient effectiveness coefficient Y yx =H ys / R zl ;

[0014] Plant absorption capacity coefficient Z xs Collect the same herbaceous plants growing in the target soil, measure the number and length of plant roots, and mark them as S x and C x , where x is the root number of a single plant, x == 1, 2, 3, 4, ..., n, n is a positive integer, and the total height of the plant above the soil is counted and calibrated as G ts , then the plant absorption capacity coefficient

[0015] Soil property coefficient T xzCollect the most suitable pH value of the herbaceous plants that users need to produce, calibrated as J sy , determine the target soil pH, calibrated as M sj , then the soil property coefficient T xz =|M si -J sy |;

[0016] The soil test evaluation model is constructed by weighting the nutrient availability information, plant absorption capacity information and soil property information in the target area to generate the soil test evaluation index TR pg , the corresponding coefficients are nutrient effectiveness coefficient Y yx , Plant absorption capacity coefficient Z xs , soil property coefficient T xz , the formula is TR pg =α1×Y yx +α2×Z xs -α3×T xz ;

[0017] At the same time, α1, α2, and α3 are all greater than 0 and are set according to actual conditions. For example, the expert weighting method is adopted, that is, experts in related fields are invited to determine the weights of various indicators through professional opinion surveys and comprehensive evaluations to ensure that the weight coefficients can accurately reflect the importance of various indicators in soil testing and evaluation. In addition, the use of various methods such as hierarchical analysis method and fuzzy comprehensive evaluation method to determine the weight coefficients can also be considered to ensure the objectivity and scientificity of the weight coefficients. I will not go into details here.

[0018] In step S2, the soil test evaluation index obtained from the soil test evaluation model is used to reflect the ability of the soil in the target area to promote the growth of the herbaceous plants that the user needs to produce. The larger its value is, the more existing elements effective for plant growth are contained in the soil in the target area, the stronger the target herbaceous plants' ability to absorb nutrients in the soil, and the more suitable it is for the pH value in the soil in the target area, the higher the ability of the soil in the target area to promote the growth of the herbaceous plants that the user needs to produce.

[0019] In step S2, when the soil test evaluation index TR pg When it is greater than or equal to the set planting threshold, it means that the soil in the target area is suitable for planting the herbaceous plants required by the user. Only adjusting the amount of fertilizer can effectively increase the grass yield. The evaluation results are output and the optimal amount of fertilizer is subsequently evaluated.

[0020] When soil tests are performed to assess the TR index pgWhen it is less than the set planting threshold, it means that the soil conditions in the target area are not suitable for planting the herbaceous plants required by the user. Simply adjusting the amount of fertilizer is unlikely to effectively increase grass yields. The evaluation results are output and it is recommended to replace the plant species that need to be planted.

[0021] In step S3, an artificial neural network (ANN) algorithm is used to construct a model using soil test evaluation index, crop yield data, and historical fertilizer application rates as known parameters, and the optimal fertilizer application rate under the maximum yield of the target crop is finally output;

[0022] Furthermore, the method for constructing the artificial neural network model of the present invention is as follows:

[0023] Step S3.1, data collection and preprocessing, collecting the crop yield data and historical fertilizer application amount of the grass plants required by the target soil planting user as the model input, and after removing outliers and processing missing values, normalizing or standardizing the data, specifically converting the categorical data (such as soil type) into numerical form, such as adjusting it to between 0 and 1;

[0024] Step S3.2, determine the model structure, determine the input and output layers: the number of nodes in the input layer is equal to the number of features taken. In this case, the features include soil test evaluation index (nutrient availability information, plant absorption capacity information and soil property information), historical fertilization data (application amount of nitrogen, phosphorus, potassium and other elements), and the predicted target number is used as the node of the output layer. For example, if the amount of N (nitrogen), P (phosphorus) and K (potassium) is predicted respectively, there should be three output nodes;

[0025] Designing hidden layers: Number of hidden layers: In the beginning, you can try using only one or two hidden layers. Adding more hidden layers can help the network learn more complex patterns, but it may also lead to overfitting.

[0026] Number of neurons in each layer: The number of neurons in the hidden layer is some intermediate value between the number of neurons in the input layer and the output layer. There is no fixed formula, but you can start with a relatively small number, such as twice the number of input features in each hidden layer, and then adjust it through experiments;

[0027] Activation function: Common activation functions for hidden layers include ReLU (Rectified Linear Unit) because it performs better in preventing the gradient vanishing problem. Other options include Sigmoid or Tanh. The specific choice depends on the nature of the problem and the performance of the network.

[0028] Step S3.4, training the model, divide the data set into training set, validation set and test set according to the ratio of 70% training, 15% validation and 15% test, use the training set data to train the model, monitor the performance on the validation set to adjust the model parameters to avoid overfitting, and adjust the network structure and training parameters such as learning rate, batch size, etc. based on the performance on the validation set.

[0029] In step S4, according to the results of model training, the weights are adjusted using an optimization algorithm to adjust the parameters of the trained model. The following are the detailed steps for adjusting the weights using the optimization algorithm:

[0030] Step S4.1, initialize the network weights. At the beginning of training, assign small random values ​​to the weights in the network to break the symmetry and allow the algorithm to learn. Common initialization methods include He initialization and Xavier initialization.

[0031] Step S4.2, define the function:

[0032] Forward propagation: The training data (soil test evaluation index, historical fertilization data, and crop yield data) are input into the network. When the data propagates forward in the network, it is processed by the weighted and activation functions of each layer and finally generates an output, that is, a prediction of the target fertilization amount. For the i-th neuron, the specific formula is: i =∑ j (w ij x j )+b i , where x is the input, w is the weight, b is the bias term, and j is the input number. Substituting the formula into the activation formula f, we can get the output a i =f(z i );

[0033] Loss function: Use a loss function (such as mean square error MSE) to calculate the difference between the predicted fertilizer application amount and the actual fertilizer application amount. Specifically, Where n is the number of samples, y i is the actual value of sample i, and is the predicted value of sample i;

[0034] Back propagation: The gradient of the loss function with respect to each weight is calculated through the back propagation algorithm. This process starts from the output layer and propagates backward through the network, using the chain rule to update the gradient of each weight. For weight w ij , and its gradient calculation formula is: In the formula, is the derivative of the loss with respect to the output of the activation function, is the derivative of the activation function, is the weighted input z i For weight w ijThe derivative of , which is usually equal to the input x j ;

[0035] Step S4.3, use an optimization algorithm (such as SGD, Adam, etc.) to update the weights in the network according to the calculated gradient, and adjust the weights to a certain extent to reduce the loss. The weight update is based on the gradient descent algorithm, and its basic formula is: Among them, w ij new and w ij old are the weights before and after the update, μ is the learning rate, is the loss function with respect to weight w ij The gradient of

[0036] Step S4.4, iterate steps S4.2 and S4.3 until the performance of the model on the validation set is no longer significantly improved, or a predetermined number of iterations is reached, thereby completing the optimization of the artificial neural network model.

[0037] Technical effects and advantages of the present invention:

[0038] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a rapid analysis method of fertilizer application amount based on grass yield, evaluates the soil properties of the target area, and builds a model of the soil of plants suitable for user needs through an artificial neural network algorithm based on the evaluation results and crop yield data and historical fertilizer application amounts as factors, outputs fertilizer application amount data with the highest grass yield, and optimizes and adjusts the obtained model to make the acquired data more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 The figure is a flow chart of the method for quickly analyzing the amount of fertilizer to be applied based on grass yield according to the present invention. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] The invention is a rapid analysis method for fertilizer application amount based on grass yield, which evaluates the soil properties of the target area, and builds a model of the soil of plants suitable for user needs through an artificial neural network algorithm based on the evaluation results and crop yield data and historical fertilizer application amounts as factors, outputs fertilizer application amount data with the highest grass yield, and optimizes and adjusts the obtained model to make the acquired data more accurate.

[0042] Example 1

[0043] like Figure 1 As shown, the steps of the rapid analysis method of fertilizer application amount based on grass yield are as follows:

[0044] Step S1, monitoring and collecting information related to soil properties in the target area, as well as crop yield data and historical fertilizer application data when building a model using an artificial neural network algorithm;

[0045] Step S2, constructing a soil test evaluation model to evaluate whether the soil in the target area can achieve increased yield by changing the amount of fertilizer applied when planting the herbaceous plants required by the user;

[0046] Step S3, for soil in the target area where yield increase can be achieved by changing the amount of fertilizer applied, an artificial neural network model is constructed using crop yield data, soil test evaluation index and historical fertilizer application amount as factors, and the fertilizer application amount data with the highest grass yield is output;

[0047] Step S4, use the optimization algorithm to adjust the weights to adjust the parameters of the trained model to improve the accuracy of the model.

[0048] Specifically, in step S1, the artificial neural network can use a large number of input variables to improve the accuracy of prediction, and the TensorFlow, Keras and PyTorch libraries in Python are commonly used to build and train the model. In actual use, due to factors such as equipment cost and complexity of real-time monitoring, the present invention is based on the premise that the soil and required plant samples in the target area are sufficient and the growth environment is relatively stable, and the soil and required plant data are continuously monitored, which may lead to the phenomenon of rationality change at a certain monitoring moment.

[0049] In step S2, a soil test evaluation model is constructed. The specific process is to collect nutrient availability information, plant absorption capacity information and soil property information in the target area, wherein the nutrient availability information includes the nutrient availability coefficient, which is calibrated as Y yx , the plant absorption capacity information includes the plant absorption capacity coefficient, calibrated as Z xs , soil property information includes soil property coefficient, calibrated as T xz ;

[0050] Nutrient effectiveness coefficient Y yx By measuring the content of elements (such as nitrogen, phosphorus, and potassium) that are effective for plant growth in the soil of the target area, randomly collected soils are mixed together to obtain the weight of the effective element content, which is calibrated as H ys , the soil mass of the statistical mixture is calibrated as R zl , then the nutrient effectiveness coefficient Y yx =H ys / Rzl ;

[0051] Plant absorption capacity coefficient Z xs Collect the same herbaceous plants growing in the target soil, measure the number and length of plant roots, and mark them as S x and C x , where x is the root number of a single plant, x == 1, 2, 3, 4, ..., n, n is a positive integer, and the total height of the plant above the soil is counted and calibrated as G ts , then the plant absorption capacity coefficient

[0052] Soil property coefficient T xz Collect the most suitable pH value of the herbaceous plants that users need to produce, calibrated as J sy , determine the target soil pH, calibrated as M sj , then the soil property coefficient T xz =|M sj -J sy |;

[0053] The soil test evaluation model is constructed by weighting the nutrient availability information, plant absorption capacity information and soil property information in the target area to generate the soil test evaluation index TR pg , the corresponding coefficients are nutrient effectiveness coefficient Y yx , Plant absorption capacity coefficient Z xs , soil property coefficient T xz , the formula is TR pg =α1×Y yx +α2×Z xs -α3×T xz ;

[0054] At the same time, α1, α2, and α3 are all greater than 0 and are set according to actual conditions. For example, the expert weighting method is adopted, that is, experts in related fields are invited to determine the weights of various indicators through professional opinion surveys and comprehensive evaluations to ensure that the weight coefficients can accurately reflect the importance of various indicators in soil testing and evaluation. In addition, the use of various methods such as hierarchical analysis method and fuzzy comprehensive evaluation method to determine the weight coefficients can also be considered to ensure the objectivity and scientificity of the weight coefficients. I will not go into details here.

[0055] In step S2, the soil test evaluation index obtained from the soil test evaluation model is used to reflect the ability of the soil in the target area to promote the growth of the herbaceous plants that the user needs to produce. The larger its value is, the more existing elements effective for plant growth are contained in the soil in the target area, the stronger the target herbaceous plants' ability to absorb nutrients in the soil, and the more suitable it is for the pH value in the soil in the target area, the higher the ability of the soil in the target area to promote the growth of the herbaceous plants that the user needs to produce.

[0056] In step S2, when the soil test evaluation index TR pg When it is greater than or equal to the set planting threshold, it means that the soil in the target area is suitable for planting the herbaceous plants required by the user. Only adjusting the amount of fertilizer can effectively increase the grass yield. The evaluation results are output and the optimal amount of fertilizer is subsequently evaluated.

[0057] When soil tests are performed to assess the TR index pg When it is less than the set planting threshold, it means that the soil conditions in the target area are not suitable for planting the herbaceous plants required by the user. Simply adjusting the amount of fertilizer is unlikely to effectively increase grass yields. The evaluation results are output and it is recommended to replace the plant species that need to be planted.

[0058] In step S3, an artificial neural network (ANN) algorithm is used to construct a model using soil test evaluation index, crop yield data, and historical fertilizer application rates as known parameters, and the optimal fertilizer application rate under the maximum yield of the target crop is finally output;

[0059] Furthermore, the method for constructing the artificial neural network model of the present invention is as follows:

[0060] Step S3.1, data collection and preprocessing, collecting the crop yield data and historical fertilizer application amount of the grass plants required by the target soil planting user as the model input, and after removing outliers and processing missing values, normalizing or standardizing the data, specifically converting the categorical data (such as soil type) into numerical form, such as adjusting it to between 0 and 1;

[0061] Step S3.2, determine the model structure, determine the input and output layers: the number of nodes in the input layer is equal to the number of features taken. In this case, the features include soil test evaluation index (nutrient availability information, plant absorption capacity information and soil property information), historical fertilization data (application amount of nitrogen, phosphorus, potassium and other elements), and the predicted target number is used as the node of the output layer. For example, if the amount of N (nitrogen), P (phosphorus) and K (potassium) is predicted respectively, there should be three output nodes;

[0062] Designing hidden layers: Number of hidden layers: In the beginning, you can try using only one or two hidden layers. Adding more hidden layers can help the network learn more complex patterns, but it may also lead to overfitting.

[0063] Number of neurons in each layer: The number of neurons in the hidden layer is some intermediate value between the number of neurons in the input layer and the output layer. There is no fixed formula, but you can start with a relatively small number, such as twice the number of input features in each hidden layer, and then adjust it through experiments;

[0064] Activation function: Common activation functions for hidden layers include ReLU (Rectified Linear Unit) because it performs better in preventing the gradient vanishing problem. Other options include Sigmoid or Tanh. The specific choice depends on the nature of the problem and the performance of the network.

[0065] Step S3.4, training the model, divide the data set into training set, validation set and test set according to the ratio of 70% training, 15% validation and 15% test, use the training set data to train the model, monitor the performance on the validation set to adjust the model parameters to avoid overfitting, and adjust the network structure and training parameters such as learning rate, batch size, etc. based on the performance on the validation set.

[0066] In step S4, according to the results of model training, the weights are adjusted using an optimization algorithm to adjust the parameters of the trained model. The following are the detailed steps for adjusting the weights using the optimization algorithm:

[0067] Step S4.1, initialize the network weights. At the beginning of training, assign small random values ​​to the weights in the network to break the symmetry and allow the algorithm to learn. Common initialization methods include He initialization and Xavier initialization.

[0068] Step S4.2, define the function:

[0069] Forward propagation: The training data (soil test evaluation index, historical fertilization data, and crop yield data) are input into the network. When the data propagates forward in the network, it is processed by the weighted and activation functions of each layer and finally generates an output, that is, a prediction of the target fertilization amount. For the i-th neuron, the specific formula is: i =∑ j (w ij x j )+b i , where x is the input, w is the weight, b is the bias term, and j is the input number. Substituting the formula into the activation formula f, we can get the output a i =f(z i );

[0070] Loss function: Use a loss function (such as mean square error MSE) to calculate the difference between the predicted fertilizer application amount and the actual fertilizer application amount. Specifically, Where n is the number of samples, y i is the actual value of sample i, and is the predicted value of sample i;

[0071] Back propagation: The gradient of the loss function with respect to each weight is calculated through the back propagation algorithm. This process starts from the output layer and propagates backward through the network, using the chain rule to update the gradient of each weight. For weight w ij , and its gradient calculation formula is: In the formula, is the derivative of the loss with respect to the output of the activation function, is the derivative of the activation function, is the weighted input z i For weight w ij The derivative of , which is usually equal to the input x j ;

[0072] Step S4.3, use an optimization algorithm (such as SGD, Adam, etc.) to update the weights in the network according to the calculated gradient, and adjust the weights to a certain extent to reduce the loss. The weight update is based on the gradient descent algorithm, and its basic formula is: Among them, w ij new and w ij old are the weights before and after the update, μ is the learning rate, is the loss function with respect to weight w ij The gradient of

[0073] Step S4.4, iterate steps S4.2 and S4.3 until the performance of the model on the validation set is no longer significantly improved, or a predetermined number of iterations is reached, thereby completing the optimization of the artificial neural network model.

[0074] The above formulas are all dimensionless and calculated numerically. Specific dimension removal can be achieved by various means such as standardization, which will not be elaborated here. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0075] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, an ATA hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state ATA hard disk.

[0076] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0077] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0078] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0079] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0080] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0081] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: U disk, mobile ATA hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, and other media that can store program codes.

[0082] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. Rapid analysis method of fertilizer application amount based on grass yield, It is characterized in that The steps include: Step S1, monitoring and collecting information related to soil properties in the target area, as well as crop yield data and historical fertilizer application data when building a model using an artificial neural network algorithm; Step S2, constructing a soil test evaluation model to evaluate whether the soil in the target area can achieve increased yield by changing the amount of fertilizer applied when planting the herbaceous plants required by the user; Step S3, for soil in the target area where yield increase can be achieved by changing the amount of fertilizer applied, an artificial neural network model is constructed using crop yield data, soil test evaluation index and historical fertilizer application amount as factors, and the fertilizer application amount data with the highest grass yield is output; Step S4, using an optimization algorithm to adjust the weights to adjust the parameters of the trained model to improve the accuracy of the model; In step S2, a soil test evaluation model is constructed. The specific process is to collect nutrient availability information, plant absorption capacity information and soil property information in the target area. The nutrient availability information includes the nutrient availability coefficient, which is calibrated as , the plant absorption capacity information includes the plant absorption capacity coefficient, calibrated as , soil property information includes soil property coefficient, calibrated as ; Nutrient availability coefficient By measuring the content of elements effective for plant growth in the soil of the target area, randomly collected soils are mixed together to obtain the weight of the effective element content and calibrate it as , the soil mass of the statistical mixture is calibrated as , then the nutrient effectiveness coefficient ; Plant absorption capacity coefficient Collect the same herbaceous plants growing in the target soil, measure the number and length of plant roots, and mark them as and , where x is the root number of a single plant, x=1, 2, 3, 4, ..., n, n is a positive integer, and the total height of the plant above the soil is counted and calibrated as , then the plant absorption capacity coefficient ; Soil property coefficient Collect the most suitable pH value of the herbaceous plants that users need to produce and calibrate it , determine the target soil pH, calibrated as , then the soil property coefficient ; The soil test evaluation model is constructed by weighting the nutrient availability information, plant absorption capacity information and soil property information in the target area to generate a soil test evaluation index. , and the corresponding coefficients are nutrient effectiveness coefficients , Plant absorption capacity coefficient , soil property coefficient , the formula is , are the weight coefficients of the corresponding indicators, all greater than 0; Through artificial neural network algorithm, soil test evaluation index, crop yield data and historical fertilizer application amount are used as known parameters to build a model, and finally output the optimal fertilizer application amount under the maximum yield of the target crop; Furthermore, the construction method of the artificial neural network model is as follows: Step S3.1, data collection and preprocessing, collecting the crop yield data and historical fertilizer application amount of the grass plants required by the user when planted in the target soil as model input, and normalizing or standardizing the data after removing outliers and processing missing values; Step S3.2, determine the model structure, determine the input and output layers: the number of nodes in the input layer is equal to the number of features taken, the number of predicted targets is used as the nodes of the output layer, and the hidden layer is designed at the same time; Step S3.4, training the model, dividing the data set into a training set, a validation set, and a test set according to the ratio of 70% training, 15% validation, and 15% testing, using the training set data to train the model, while monitoring the performance on the validation set to adjust the model parameters to avoid overfitting, and adjusting the network structure and training parameters based on the performance on the validation set; In step S4, according to the results of model training, the weights are adjusted using an optimization algorithm to adjust the parameters of the trained model: Step S4.1, initializing network weights; Step S4.2, defining a function; Step S4.3, use optimization algorithm to adjust weights and reduce losses; Step S4.4, iterate steps S4.2 and S4.3 until the performance of the model on the validation set is no longer significantly improved, or a predetermined number of iterations is reached, thereby completing the optimization of the artificial neural network model.

2. The method for rapid analysis of fertilizer application amount based on grass yield according to claim 1, Features: In step S2, when the soil test evaluation index When it is greater than or equal to the set planting threshold, the evaluation result of the optimal fertilizer amount for subsequent evaluation is output; When soil tests are performed to assess the index When it is less than the set planting threshold, the evaluation result is output to recommend changing the plant species to be planted.

Images