Agricultural producing area environment monitoring big data analysis method and system

Abstract

The invention relates to the technical field of agricultural production area analysis, and relates to an agricultural production area environment monitoring big data analysis method and system, and the method comprises the steps: collecting agricultural production area multi-dimensional space-time environment data and crop growth state data, and carrying out the preprocessing of the collected data, so as to form space-time feature sequence data containing a timestamp; constructing, training and optimizing a soil organic matter dynamic evolution prediction model, and inputting the farming management measure dynamic regulation and control set quantity and spatial-temporal characteristic sequence data input in real time into the soil organic matter dynamic evolution prediction model for simulation analysis, so as to predict the dynamic evolution prediction of the soil organic matter after the farming management measure of the dynamic regulation and control set quantity is implemented. According to the expected change track and the final achievement state of the soil organic matter content of the agricultural producing area in the whole management period, an operation guidance scheme of the agricultural producing area is generated according to a simulation analysis result. According to the invention, through big data analysis and deep learning intelligent prediction, the utilization efficiency of agricultural resources and the sustainability of production management are significantly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural production area analysis, and more particularly, to a big data analysis method and system for monitoring the environment of agricultural production areas. Background Art

[0002] Soil is the cornerstone of agricultural production, and the content of soil organic matter is a key core indicator for measuring soil fertility, health status, and sustainable utilization ability. It not only directly affects the growth, development, yield, and quality of crops, but also plays an irreplaceable role in carbon sequestration, soil and water conservation, and ecological environment protection. With the rapid development of modern agriculture towards precision and intelligence, real-time, dynamic, and precise monitoring and management of the agricultural production area environment, especially effective prediction and regulation of key elements such as soil organic matter that change dynamically, have become an urgent need to improve agricultural production efficiency, ensure the quality and safety of agricultural products, and promote the sustainable development of agriculture. Currently, in the management of soil organic matter in agricultural production areas and the support of farming decision-making, existing methods have some inherent limitations. On the one hand, the monitoring of soil organic matter often relies on periodic laboratory sampling and analysis, which is not only costly and time-consuming, but also difficult to capture the high spatial heterogeneity and temporal dynamic change characteristics of soil organic matter, resulting in insufficient representativeness and timeliness of the monitoring results. On the other hand, existing farming management suggestions mostly rely on empirical judgments or static and general fertilization guidelines, lacking refined consideration of the dynamic environmental conditions of specific plots, the growth status of crops, and the combined effects of different farming measures. Moreover, regarding how to quantitatively evaluate the impact of different farming management measures on the future evolution trajectory of soil organic matter and make forward-looking optimization decisions based on this, existing farming management suggestions often fail to provide effective support, resulting in the adjustment of farming management measures being often lagged and lacking sufficient scientific predictability and precision, and it is difficult to meet the strict requirements of modern precision agriculture for efficient resource utilization and environmental protection. Summary of the Invention

[0003] The purpose of the present invention is to provide a big data analysis method and system for monitoring the environment of agricultural production areas, which realizes dynamic and precise prediction of soil organic matter in agricultural production areas and intelligent optimization decision-making of farming management through big data analysis and deep learning intelligent prediction, significantly improving the efficiency of agricultural resource utilization and the sustainability of production management.

[0004] The present invention is achieved through the following technical solutions: A big data analysis method for monitoring the environment of agricultural production areas, the steps of the method include: Collect multi-dimensional spatio-temporal environmental data and crop growth status data covering the agricultural production area, and preprocess the multi-dimensional spatio-temporal environmental data and crop growth status data to form spatio-temporal feature sequence data containing timestamps; Construct a prediction model for the dynamic evolution of soil organic matter, where the prediction model for the dynamic evolution of soil organic matter is configured to: receive the spatio-temporal feature sequence data and the corresponding agricultural management measures as inputs, and output the predicted value of the soil organic matter content in the agricultural production area; For a preset target change rate of the soil organic matter content in the agricultural production area, drive the prediction model for the dynamic evolution of soil organic matter to perform iterative optimization to determine the dynamic regulation setting amount of the agricultural management measures that enables the predicted value of the soil organic matter content to reach the target change rate; Input the dynamic regulation setting amount of the agricultural management measures, combined with the spatio-temporal feature sequence data input in real time, into the prediction model for the dynamic evolution of soil organic matter for simulation analysis, so as to predict the expected change trajectory and the final achieved state of the soil organic matter content in the agricultural production area during the entire management cycle after implementing the agricultural management measures with the dynamic regulation setting amount, and generate an operation guidance plan for the agricultural production area based on the simulation analysis results.

[0005] Optionally, the preprocessing of the multi-dimensional spatio-temporal environment data and the crop growth status data is specifically as follows: Perform spatial interpolation processing on the multi-dimensional spatio-temporal environment data to generate a continuous data layer reflecting the spatial distribution characteristics of the agricultural production area; Perform preprocessing and feature extraction on the crop growth status data to obtain time series features characterizing the crop growth process and its impact on soil organic matter; Align, grid, and fuse the continuous data layer and the time series features in the time and space dimensions, and perform missing value filling and feature transformation to form the spatio-temporal feature sequence data.

[0006] Optionally, the construction of the prediction model for the dynamic evolution of soil organic matter is specifically as follows: Construct a prediction model for the dynamic evolution of soil organic matter based on a deep spatio-temporal encoder-decoder neural network architecture; The prediction model for the dynamic evolution of soil organic matter includes an encoder and a decoder, where: The encoder is configured to use convolutional long short-term memory network units to process the spatio-temporal feature sequence data and extract its deep spatio-temporal dynamic features; The decoder is configured to combine the attention mechanism and generate the predicted value of the soil organic matter content based on the deep spatio-temporal dynamic features extracted by the encoder and the corresponding agricultural management measures; The prediction model for the dynamic evolution of soil organic matter is constructed as: input the spatio-temporal feature sequence data into the encoder for processing, and input the agricultural management measures as conditional information, together with the deep spatio-temporal dynamic features output by the encoder, into the decoder for processing; The decoder outputs the predicted values of soil organic matter content for each predetermined spatial grid unit within the agricultural production area.

[0007] Optionally, the training process of the soil organic matter dynamic evolution prediction model is as follows: Obtain and divide the training set and test set that include historical soil organic matter content, corresponding spatio-temporal feature sequence data, and agricultural management measures; Input the training set into the soil organic matter dynamic evolution prediction model in batches, obtain the predicted values of the soil organic matter content for this batch of data through forward propagation, and calculate the loss between the predicted values and the actual soil organic matter content corresponding to this batch of training data according to the preset optimization objective function; Calculate the gradient of the soil organic matter dynamic evolution prediction model through backpropagation, and iteratively update the parameters of the soil organic matter dynamic evolution prediction model through an optimization algorithm; Until after iterating the preset maximum number of times, evaluate the soil organic matter dynamic evolution prediction model through the test set to complete the training of the soil organic matter dynamic evolution prediction model.

[0008] Optionally, calculating the loss between the predicted value and the actual soil organic matter content corresponding to this batch of training data according to the preset optimization objective function is specifically as follows: For the predicted values of soil organic matter content output by the soil organic matter dynamic evolution prediction model and the corresponding true soil organic matter content target values, calculate the prediction deviations at each corresponding data point; Quantify each prediction deviation based on the predefined mean square error loss function to obtain the individual error values at each data point; Perform an aggregation process on the individual error values to jointly form the optimization objective function value for guiding the update of the parameters of the soil organic matter dynamic evolution prediction model.

[0009] Optionally, driving the soil organic matter dynamic evolution prediction model to perform iterative optimization to determine the dynamic regulation setting amount of agricultural management measures that enables the predicted value of soil organic matter content to achieve the target change rate is specifically as follows: Generate candidate dynamic regulation setting amounts of agricultural management measures according to the allowable adjustment range and constraint conditions of the preset dynamic regulation setting amounts of agricultural management measures; For each group of candidate dynamic regulation setting amounts of agricultural management measures processed in the current iteration cycle, perform the following respectively: Input the candidate dynamic regulation setting amount of agricultural management measures into the soil organic matter dynamic evolution prediction model to obtain the expected change trajectory of the soil organic matter content for each predetermined spatial grid unit within the agricultural production area under this group of candidate agricultural management measures; Based on the optimization objective function with the achievement degree of the target change rate of soil organic matter content in the agricultural production area as the index, combined with the predefined evaluation function, calculate the comprehensive efficacy score of the dynamic regulation setting amount of this group of candidate farm management measures based on the expected change trajectory; According to the comprehensive efficacy scores of the dynamic regulation setting amounts of each group of candidate farm management measures, and within the allowable adjustment range and constraint conditions, generate new dynamic regulation setting amounts of candidate farm management measures, and return them to the input soil organic matter dynamic evolution prediction model to perform iterative updates until the preset maximum number of iterations is reached; Select the one with the optimal comprehensive efficacy score as the dynamic regulation setting amount of the farm management measures for output.

[0010] Optionally, the optimization objective function with the achievement degree of the target change rate of soil organic matter content in the agricultural production area as the index has the following specific calculation formula:

[0011] Among them, is the optimization objective function, is the total number of predefined spatial grid units in the agricultural production area, is the predefined normal constant adjustment factor, j is the index of the spatial grid unit, is the predicted soil organic matter content of the j-th spatial grid unit, is the end time of the prediction period, is the initial soil organic matter content of the j-th predefined spatial grid unit, is the target change rate of soil organic matter content of the j-th predefined spatial grid unit.

[0012] Optionally, the calculation formula of the evaluation function is:

[0013] Among them, is the evaluation value, 、 are the weight coefficients respectively, is the economic cost, is the environmental impact assessment value, and m is the current dynamic regulation setting amount of the candidate farm management measures.

[0014] Optionally, the calculation formula of the comprehensive efficacy score is:

[0015] Among them, is the comprehensive efficacy score of the dynamic regulation setting amount of the candidate farm management measures, 、 are the weight coefficients respectively.

[0016] The big data analysis system for agricultural production area environment monitoring includes: A data acquisition unit that collects multi-dimensional spatio-temporal environment data and crop growth status data covering the agricultural production area, and pre-processes the multi-dimensional spatio-temporal environment data and crop growth status data to form spatio-temporal feature sequence data including timestamps; A model construction unit that constructs a dynamic evolution prediction model for soil organic matter. The dynamic evolution prediction model for soil organic matter is configured to: receive the spatio-temporal feature sequence data and the corresponding agricultural management measures as inputs, and output the predicted value of the soil organic matter content in the agricultural production area; An optimization unit that drives the dynamic evolution prediction model for soil organic matter to perform iterative optimization according to the preset target change rate of the soil organic matter content in the agricultural production area, so as to determine the dynamic regulation setting amount of the agricultural management measures that enables the predicted value of the soil organic matter content to reach the target change rate; An analysis unit that inputs the dynamic regulation setting amount of the agricultural management measures and the spatio-temporal feature sequence data input in real time into the dynamic evolution prediction model for soil organic matter for simulation analysis, so as to predict the expected change trajectory and the final achieved state of the soil organic matter content in the agricultural production area during the entire management cycle after implementing the agricultural management measures with the dynamic regulation setting amount, and generate an operation guidance plan for the agricultural production area based on the simulation analysis results.

[0017] The technical solution of the present invention has at least the following advantages and beneficial effects: On the one hand, the present invention constructs a spatio-temporal feature sequence including timestamps by collecting and integrating multi-dimensional spatio-temporal environment data and crop growth status data, which can more comprehensively and dynamically reflect the actual situation of the agricultural production area and lay a solid data foundation for subsequent accurate prediction. On the other hand, the present invention constructs a dynamic evolution prediction model for soil organic matter based on the deep spatio-temporal encoder-decoder neural network architecture, which can not only effectively process complex spatio-temporal feature sequence data, deeply understand its internal relationship and combine agricultural management measures as conditional inputs, but also significantly improve the accuracy and spatial fineness of soil organic matter content prediction through advanced technologies such as the attention mechanism. Moreover, by driving the prediction model to perform iterative optimization, the present invention can not only actively explore and determine the optimal dynamic regulation setting amount of agricultural management measures to achieve the preset target change rate of soil organic matter, but also perform forward-looking simulation analysis in combination with real-time data to generate specific and operable operation guidance plans, realizing a fundamental leap from traditional passive monitoring and empirical decision-making to active prediction and intelligent optimization regulation. Description of the Drawings

[0018] Figure 1 It is a schematic flow chart of the big data analysis method for agricultural production area environment monitoring provided by the present invention; Figure 2It is a schematic diagram of the principle of the big data analysis system for agricultural production area environment monitoring provided by the present invention. Detailed implementation manners

[0019] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Obviously, what is described is a part rather than all of the present invention. Components of the present invention usually described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations.

[0020] As Figure 1 shown, Figure 1 It is a schematic diagram of the process of the big data analysis method for agricultural production area environment monitoring provided by the present invention.

[0021] In one embodiment, the present invention provides a big data analysis method for agricultural production area environment monitoring. The steps of the method include: Collect multi-dimensional spatio-temporal environment data and crop growth state data covering the agricultural production area, and preprocess the multi-dimensional spatio-temporal environment data and crop growth state data to form spatio-temporal feature sequence data containing time stamps; Construct a dynamic evolution prediction model of soil organic matter. The dynamic evolution prediction model of soil organic matter is configured to: receive the spatio-temporal feature sequence data and the corresponding agricultural management measures as inputs, and output the predicted value of the soil organic matter content in the agricultural production area; For a preset target change rate of the soil organic matter content in the agricultural production area, drive the dynamic evolution prediction model of soil organic matter to perform iterative optimization to determine the dynamic regulation setting amount of the agricultural management measures that enables the predicted value of the soil organic matter content to reach the target change rate; Input the dynamic regulation setting amount of the agricultural management measures, combined with the spatio-temporal feature sequence data input in real time, into the dynamic evolution prediction model of soil organic matter for simulation analysis, so as to predict the expected change trajectory and the final achieved state of the soil organic matter content in the agricultural production area during the entire management cycle after implementing the agricultural management measures with the dynamic regulation setting amount, and generate an operation guidance plan for the agricultural production area based on the simulation analysis results.

[0022] Specifically, in this embodiment, the soil temperature and humidity, and conductivity data are collected in real time through a soil sensor array deployed in the agricultural production area. The air temperature and humidity, light, and rainfall data are collected through a small automatic weather station. The normalized difference vegetation index is extracted from the satellite remote sensing images obtained regularly as the crop growth status data, and the collected data is preprocessed to form spatio-temporal feature sequence data with accurate timestamps. A dynamic evolution prediction model for soil organic matter is constructed. The dynamic evolution prediction model for soil organic matter is based on a deep neural network architecture and receives the above spatio-temporal feature sequence data and the corresponding historical farming management measures of the plot. It can be understood that the farming management measures include fertilization type coding, fertilization amount, irrigation amount, etc., as inputs, and outputs the predicted values of the soil organic matter content at future specific time points for each grid in the plot. This embodiment mainly aims at the target change rate of an average 5% increase in soil organic matter content in the next year set by the user, and uses a genetic algorithm to drive the constructed dynamic evolution prediction model for soil organic matter to perform iterative optimization. In the iterative process, the candidate setting amounts of farming management measures are continuously adjusted, and the goal is to find a set of farming management measures that can make the soil organic matter content output by the prediction model closest to the target change rate. Finally, the dynamically regulated setting amounts of the farming management measures obtained by optimizing the genetic algorithm, combined with the latest spatio-temporal feature sequence data, are input into the dynamic evolution prediction model for soil organic matter again for a one-year simulation analysis. This analysis predicts the monthly change trajectory of the soil organic matter content and the expected achievement status at the end of the year under these farming management measures. Based on the results of this simulation analysis, the system automatically generates an agricultural production area operation guidance plan, which clearly lists the recommended fertilization types, quantities, and irrigation suggestions for each month.

[0023] In specific implementation, the preprocessing of the multi-dimensional spatio-temporal environmental data and crop growth status data is specifically as follows: Perform spatial interpolation processing on the multi-dimensional spatio-temporal environmental data to generate a continuous data layer reflecting the spatial distribution characteristics of the agricultural production area; Preprocess and extract features from the crop growth status data to obtain time series features characterizing the crop growth process and its impact on soil organic matter; Align, grid, and fuse the continuous data layer and the time series features in the time and space dimensions, and perform missing value filling and feature transformation to form the spatio-temporal feature sequence data.

[0024] In specific implementation, for multi-dimensional spatio-temporal environmental data in this embodiment, including discrete point-like soil environmental data obtained from a soil sensor array, ordinary Kriging interpolation method is used for spatial interpolation processing to convert it into a continuous soil attribute data layer covering the entire agricultural production area with a resolution of 10 meters × 10 meters. For crop growth status data in this embodiment, including NDVI time series data obtained by satellite remote sensing, it is smoothed and denoised, and key phenological period nodes such as the start date, peak date, and end date of the growing season and the cumulative NDVI value of the growing season are extracted based on the smoothed NDVI time series as time series features characterizing the crop growth process and its potential impact on soil organic matter. The continuous soil data layer generated by the above processing and the crop growth phenological time series features extracted are unified in the time and space dimensions. Spatially, all data are mapped onto a unified 10 meters × 10 meters geographic grid unit. After alignment and fusion, if there are still missing values in the grid data due to reasons such as sensor failures or cloud cover, linear interpolation method based on adjacent time point data is used for filling, and minimum-maximum normalization processing is performed on all fused numerical features to form the final spatio-temporal feature sequence data.

[0025] In the specific implementation of this embodiment, the construction of the soil organic matter dynamic evolution prediction model is specifically as follows: Construct a soil organic matter dynamic evolution prediction model based on a deep spatio-temporal encoder-decoder neural network architecture; The soil organic matter dynamic evolution prediction model includes an encoder and a decoder, where: The encoder is configured to use convolutional long short-term memory network units to process the spatio-temporal feature sequence data and extract its deep spatio-temporal dynamic features; The decoder is configured to combine an attention mechanism to generate the predicted value of the soil organic matter content based on the deep spatio-temporal dynamic features extracted by the encoder and the corresponding farm management measures; The soil organic matter dynamic evolution prediction model is constructed as follows: input the spatio-temporal feature sequence data into the encoder for processing, and use the farm management measures as conditional information and input them together with the deep spatio-temporal dynamic features output by the encoder into the decoder for processing; The decoder outputs the predicted value of the soil organic matter content for each predetermined spatial grid unit within the agricultural production area.

[0026] In specific implementation, the soil organic matter dynamic evolution prediction model of this embodiment adopts a specific architecture based on a deep spatio-temporal encoder-decoder neural network, specifically including an encoder and a decoder: The encoder is stacked by three layers of convolutional long short-term memory network (ConvLSTM) units, which is used to receive spatio-temporal feature sequence data as input. Each layer of ConvLSTM units processes the spatial structure of the input feature map through its convolutional kernel and captures the dynamic changes of the time series through its LSTM gating mechanism, so as to layer by layer extract and encode the deep spatio-temporal dynamic features in the data, and finally output a fixed-length context vector as a compact representation of the historical spatio-temporal information. The decoder also adopts three layers of ConvLSTM units and combines an attention mechanism at its input end. In this embodiment, the Bahdanau attention mechanism is specifically adopted. The decoder receives the context vector output by the encoder and, at each decoding time step, selectively focuses on the output features of the encoder in different historical time series states according to the weights calculated by the attention mechanism. At the same time, the decoder also receives the corresponding agricultural management measures as additional inputs. It can be understood that the corresponding agricultural management measures in this embodiment are specifically feature vectors containing parameters such as fertilizer type, fertilizer application rate, and irrigation amount, and this vector is used as a conditional input at each prediction time step.

[0027] In a further implementation of this embodiment, the spatio-temporal feature sequence data obtained by preprocessing is input into the encoder. After the encoder finishes processing, its final hidden state and the output sequence of all time steps are transmitted. At each prediction time step, the decoder uses its current hidden state, the predicted output of the previous time step, the weighted encoder output calculated by the attention mechanism, and the current agricultural management measure condition information, and through its ConvLSTM unit and a subsequent convolutional layer, the convolutional layer is used to adjust the output channels and restore the spatial dimension, and generate the predicted value of the soil organic matter content of each 10-meter × 10-meter predetermined spatial grid unit in the agricultural production area at the next time point.

[0028] In the specific implementation of this embodiment, the training process of the soil organic matter dynamic evolution prediction model is as follows: Obtain and divide the training set and test set containing historical soil organic matter content, corresponding spatio-temporal feature sequence data and agricultural management measures; Input the training set into the soil organic matter dynamic evolution prediction model in batches, obtain the predicted values of the soil organic matter content of this batch of data through forward propagation, and calculate the loss between the predicted values and the actual soil organic matter content corresponding to this batch of training data according to the preset optimization objective function; Calculate the gradient of the soil organic matter dynamic evolution prediction model through backpropagation, and iteratively update the parameters of the soil organic matter dynamic evolution prediction model through an optimization algorithm; After iterating until the preset maximum number of times, the soil organic matter dynamic evolution prediction model is evaluated using the test set to complete the training of the soil organic matter dynamic evolution prediction model.

[0029] Specifically, the training process of the soil organic matter dynamic evolution prediction model is as follows: Obtain from the historical database the historical soil organic matter content of each plot in the past five years, the corresponding spatio-temporal feature sequence data, and the corresponding records of farming management measures. Divide this dataset into a training set and a test set in a ratio of 80%. At the start of training, input the training set data into the initialized soil organic matter dynamic evolution prediction model in batches. For each batch of data, the soil organic matter dynamic evolution prediction model performs forward propagation to obtain the predicted sequence of soil organic matter content corresponding to this batch of data. Calculate the loss value between the predicted value output by the soil organic matter dynamic evolution prediction model and the actual soil organic matter content corresponding to this batch of training data according to the preset optimization objective function. And calculate the gradient of this loss value with respect to the soil organic matter dynamic evolution prediction model through the backpropagation algorithm. Use the Adam optimization algorithm to iteratively update the model's parameters according to the calculated gradient and the preset learning rate of 0.001 until the preset maximum number of iterations is reached to complete the training. Then evaluate the generalization ability and prediction accuracy of the finally trained soil organic matter dynamic evolution prediction model using the test set to complete the training of the soil organic matter dynamic evolution prediction model.

[0030] Further, calculating the loss between the predicted value and the actual soil organic matter content corresponding to this batch of training data according to the preset optimization objective function is specifically: For the predicted value of soil organic matter content output by the soil organic matter dynamic evolution prediction model and the corresponding true soil organic matter content target value, calculate the prediction deviation at each corresponding data point; Based on the predefined mean square error loss function, quantify each prediction deviation to obtain the individual error value of each data point; Aggregate the individual error values to jointly form the optimization objective function value for guiding the update of the parameters of the soil organic matter dynamic evolution prediction model.

[0031] In implementation, after the soil organic matter dynamic evolution prediction model performs forward propagation on a batch of input spatio-temporal feature sequence data and farming management measures, it outputs the predicted values of soil organic matter content for each sample in this batch at each prediction time point and each spatial grid cell. For each predicted value output by the model and its corresponding actual soil organic matter content target value in the training data, the prediction deviation between the two is calculated, and based on the predefined mean square error (MSE) loss function, each prediction deviation is quantified. Specifically, the square of the deviation is calculated to obtain the individual square error value of this data point, and the individual square error values of all data points in this batch are aggregated and processed to solve their average value, obtaining the mean square error loss value of this batch, which is used to represent the optimization objective function value for guiding the update of the parameters of the soil organic matter dynamic evolution prediction model.

[0032] In the specific application of this embodiment, the soil organic matter dynamic evolution prediction model is driven to perform iterative optimization to determine the dynamic regulation setting amount of farming management measures that enables the predicted value of soil organic matter content to achieve the target change rate, specifically: Generate candidate dynamic regulation setting amounts of farming management measures according to the preset allowable adjustment range and constraint conditions of the dynamic regulation setting amounts of farming management measures; For each group of candidate dynamic regulation setting amounts of farming management measures processed in the current iteration cycle, respectively execute: Input the candidate dynamic regulation setting amount of farming management measures into the soil organic matter dynamic evolution prediction model to obtain the expected change trajectory of soil organic matter content in each predetermined spatial grid cell within the agricultural production area under this group of candidate farming management measures; Based on the optimization objective function with the achievement degree of the target change rate of soil organic matter content in the agricultural production area as the index, and combined with the predefined evaluation function, calculate the comprehensive effectiveness score of this group of candidate dynamic regulation setting amounts of farming management measures based on the expected change trajectory; According to the comprehensive effectiveness scores of each group of candidate dynamic regulation setting amounts of farming management measures, and within the allowable adjustment range and constraint conditions, generate new candidate dynamic regulation setting amounts of farming management measures and return them to input into the soil organic matter dynamic evolution prediction model to perform iterative update until the preset maximum number of iterations is reached; Select the one with the optimal comprehensive effectiveness score as the dynamic regulation setting amount of farming management measures for output.

[0033] In specific implementation, the allowable adjustment range and constraint conditions of the preset dynamic regulation setting amount of agricultural management measures are specifically defined based on agricultural expert knowledge. It can be understood that the allowable adjustment range in this embodiment includes a nitrogen fertilizer application amount range of 0 - 200 kg / ha, an organic fertilizer application amount of 0 - 3000 kg / ha, and an irrigation frequency of 0 - 5 times / growing season. The constraint condition is that the total nitrogen input does not exceed a specific environmental threshold. In this embodiment, the genetic algorithm is used for optimization. Initialize the GA population, where each individual represents a group of candidate dynamic regulation setting amounts of agricultural management measures. The initial population is randomly generated within the allowable range, and the population size is set to 50. Enter the iterative optimization loop, with a total of 100 iterations. In each iteration cycle, for each group of candidate dynamic regulation setting amounts of agricultural management measures in the population, perform the following operations: Input this group of candidate setting amounts, the corresponding initial soil data, and the meteorological data during the prediction period into the trained prediction model for the dynamic evolution of soil organic matter. After the model runs, it outputs the expected value of the soil organic matter content in each predetermined spatial grid unit within the agricultural production area at the end of the prediction period under this measure.

[0034] According to the optimization objective function with the achievement degree of the target change rate of soil organic matter content in the agricultural production area as the index, calculate the value of this individual.

[0035]

[0036] Among them, is the optimization objective function, is the total number of predetermined spatial grid units within the agricultural production area, is the preset normal constant adjustment factor, j is the index of the spatial grid unit, is the predicted soil organic matter content of the jth spatial grid unit, is the end time of the prediction period, is the initial soil organic matter content of the jth predetermined spatial grid unit, is the target change rate of the soil organic matter content of the jth predetermined spatial grid unit. When calculating, for each grid unit j, first calculate its actual predicted relative change rate of soil organic matter , and then calculate the absolute difference between this actual change rate and the target change rate . The difference is multiplied by the adjustment factor and then added by 1 to be used as the denominator. After taking its reciprocal, the results of all grid units are summed to obtain the final value. The larger the

[0037] Calculate the value of this individual according to the evaluation function.

[0038]

[0039] Among them, is the evaluation value, , are the weight coefficients respectively, is the economic cost, is the environmental impact assessment value, and m is the dynamic regulation setting quantity of the current candidate agricultural management measure. The calculation steps of are as follows: According to the current candidate agricultural management measure m, clarify all types of fertilizers planned to be used and their respective planned application total amounts. Obtain the unit price of each planned fertilizer. For each fertilizer, multiply its planned application total amount by its unit price to get the cost of this fertilizer. Add up the costs of all types of fertilizers to get the total fertilizer cost. According to the irrigation plan in measure m, combined with the unit irrigation cost, estimate the total irrigation cost. According to other possible agricultural operations in measure m, estimate the cost of agricultural operations. Add up the total fertilizer cost, the total irrigation cost and the costs of other relevant operations to get the total economic cost of the candidate agricultural management measure m The calculation steps of are as follows: According to the current candidate agricultural management measure m, calculate the total nitrogen amount planned to be applied to the farmland through various fertilizers, that is, the total nitrogen input amount. According to the crop type planted in this farmland and the expected yield level, estimate the nitrogen amount that the crop is expected to absorb and utilize during the entire growth cycle, that is, the estimated nitrogen absorption amount of the crop. Subtract the estimated nitrogen absorption amount of the crop in the second step from the total nitrogen input amount calculated in the first step to get the potential nitrogen surplus amount. Judge this nitrogen surplus amount: If the total nitrogen input amount is less than or equal to the estimated nitrogen absorption amount of the crop, it is considered that the nitrogen leaching risk under this measure is extremely low, and the environmental impact assessment value is recorded as zero or a very small preset positive value. If the total nitrogen input amount is greater than the estimated nitrogen absorption amount of the crop, multiply this nitrogen surplus amount by an empirical leaching coefficient preset according to the local soil type and rainfall conditions, and the result of the product is the environmental impact assessment value

[0040] Calculate the comprehensive efficiency score of this individual according to the comprehensive efficiency scoring formula.

[0041] For the comprehensive efficiency score described above, its calculation formula is:

[0042] Among them, is the comprehensive efficiency score of the dynamic regulation setting quantity of the candidate agricultural management measure, , They are weight coefficients respectively. By calculating and ratio, the comprehensive efficiency score is obtained. During the optimization process, the goal is to find the dynamic regulation setting amount of the agricultural management measures that maximizes this value. A high score indicates that the measure can not only effectively achieve the expected goal of soil organic matter, but also has relatively low economic and environmental costs.

[0043] After evaluating all individuals in the current generation, a new generation of population is generated through operations such as crossover and mutation according to the comprehensive efficiency scores of each group of individuals. Enter the next generation iteration.

[0044] After reaching the preset maximum number of iterations, select the dynamic regulation setting amount of the agricultural management measures with the highest comprehensive efficiency score from all individuals as the final optimization result for output.

[0045] As Figure 2 shown, the present invention also provides another embodiment: an agricultural production area environmental monitoring big data analysis system, including: A data acquisition unit, which collects multi-dimensional spatio-temporal environmental data and crop growth status data covering the agricultural production area, and preprocesses the multi-dimensional spatio-temporal environmental data and crop growth status data to form spatio-temporal feature sequence data containing time stamps; A model construction unit, which constructs a dynamic evolution prediction model of soil organic matter. The dynamic evolution prediction model of soil organic matter is configured to: receive the spatio-temporal feature sequence data and the corresponding agricultural management measures as inputs, and output the predicted value of the soil organic matter content in the agricultural production area; An optimization unit, which drives the dynamic evolution prediction model of soil organic matter to perform iterative optimization for a preset target change rate of the soil organic matter content in the agricultural production area, so as to determine the dynamic regulation setting amount of the agricultural management measures that enables the predicted value of the soil organic matter content to reach the target change rate; An analysis unit, which inputs the dynamic regulation setting amount of the agricultural management measures and the spatio-temporal feature sequence data input in real time into the dynamic evolution prediction model of soil organic matter for simulation analysis, so as to predict the expected change trajectory and the final achieved state of the soil organic matter content in the agricultural production area during the entire management cycle after implementing the agricultural management measures with the dynamic regulation setting amount, and generate an operation guidance plan for the agricultural production area based on the simulation analysis results.

[0046] The above is only a preference of the present invention and is not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A big data analysis method for agricultural production area environmental monitoring, characterized in that The steps of the method include: Collect multi-dimensional spatio-temporal environmental data and crop growth status data covering agricultural production areas, and preprocess the multi-dimensional spatio-temporal environmental data and crop growth status data to form spatio-temporal feature sequence data including timestamps; Construct a dynamic evolution prediction model of soil organic matter, and the dynamic evolution prediction model of soil organic matter is configured to: receive the spatio-temporal feature sequence data and the corresponding agricultural management measures as inputs, and output the predicted value of the soil organic matter content in the agricultural production area; For a preset target change rate of soil organic matter content in the agricultural production area, drive the dynamic evolution prediction model of soil organic matter to perform iterative optimization to determine the dynamic regulation setting amount of agricultural management measures that enables the predicted value of soil organic matter content to reach the target change rate; Input the dynamic regulation setting amount of the agricultural management measures, combined with the spatio-temporal feature sequence data input in real time, into the dynamic evolution prediction model of soil organic matter for simulation analysis, so as to predict the expected change trajectory and the final achieved state of the soil organic matter content in the agricultural production area during the entire management cycle after implementing the agricultural management measures with the dynamic regulation setting amount, and generate an operation guidance plan for the agricultural production area based on the simulation analysis results.

2. The agricultural production area environmental monitoring big data analysis method according to claim 1, characterized in that The preprocessing of the multi-dimensional spatio-temporal environmental data and crop growth status data is specifically: Perform spatial interpolation processing on the multi-dimensional spatio-temporal environmental data to generate a continuous data layer reflecting the spatial distribution characteristics of the agricultural production area; Preprocess and extract features from the crop growth status data to obtain time series features characterizing the crop growth process and its impact on soil organic matter; Align, grid, and fuse the continuous data layer and the time series features in the time and space dimensions, and perform missing value filling and feature transformation to form the spatio-temporal feature sequence data.

3. The agricultural production area environment monitoring big data analysis method according to claim 2, characterized in that The construction of the dynamic evolution prediction model of soil organic matter is specifically: Construct a dynamic evolution prediction model of soil organic matter based on a deep spatio-temporal encoder-decoder neural network architecture; The dynamic evolution prediction model of soil organic matter includes an encoder and a decoder, where: The encoder is configured to use convolutional long short-term memory network units to process the spatio-temporal feature sequence data and extract its deep spatio-temporal dynamic features; The decoder is configured to combine an attention mechanism and generate the predicted value of the soil organic matter content based on the deep spatio-temporal dynamic features extracted by the encoder and the corresponding agricultural management measures; The dynamic evolution prediction model of soil organic matter is constructed as: input the spatio-temporal feature sequence data into the encoder for processing, and input the agricultural management measures as conditional information, together with the deep spatio-temporal dynamic features output by the encoder, into the decoder for processing; The decoder outputs the predicted value of the soil organic matter content of each predetermined spatial grid unit in the agricultural production area.

4. The agricultural production area environment monitoring big data analysis method according to claim 3, characterized in that The training process of the dynamic evolution prediction model of soil organic matter is: Obtain and divide a training set and a test set including historical soil organic matter content, corresponding spatio-temporal feature sequence data, and agricultural management measures; The training set is input into the soil organic matter dynamic evolution prediction model in batches, and the predicted value of the soil organic matter content of this batch of data is obtained through forward propagation. The loss between the predicted value and the actual soil organic matter content corresponding to this batch of training data is calculated according to a preset optimization objective function; Through backpropagation, the gradient of the soil organic matter dynamic evolution prediction model is calculated, and the parameters of the soil organic matter dynamic evolution prediction model are iteratively updated through an optimization algorithm; After iterating the preset maximum number of times, the soil organic matter dynamic evolution prediction model is evaluated through the test set to complete the training of the soil organic matter dynamic evolution prediction model.

5. The agricultural production area environment monitoring big data analysis method according to claim 4, characterized in that The loss between the predicted value and the actual soil organic matter content corresponding to this batch of training data calculated according to the preset optimization objective function is specifically: For the predicted value of the soil organic matter content output by the soil organic matter dynamic evolution prediction model and the corresponding true soil organic matter content target value, the prediction deviation at each corresponding data point is calculated; Based on a predefined mean square error loss function, each prediction deviation is quantified to obtain the individual error value of each data point; The individual error values are aggregated to jointly form an optimization objective function value for guiding the update of the parameters of the soil organic matter dynamic evolution prediction model.

6. The agricultural production area environmental monitoring big data analysis method according to claim 5, characterized in that The soil organic matter dynamic evolution prediction model is driven to perform iterative optimization to determine the dynamic regulation setting amount of the agricultural management measures that enables the predicted value of the soil organic matter content to achieve the target change rate, which is specifically: According to the allowable adjustment range and constraint conditions of the preset dynamic regulation setting amount of the agricultural management measures, a candidate dynamic regulation setting amount of the agricultural management measures is generated; For each group of candidate dynamic regulation setting amounts of the agricultural management measures processed in the current iteration cycle, the following are respectively executed: The candidate dynamic regulation setting amount of the agricultural management measures is input into the soil organic matter dynamic evolution prediction model to obtain the expected change trajectory of the soil organic matter content of each predetermined spatial grid unit in the agricultural production area under this group of candidate agricultural management measures; According to the optimization objective function with the achievement degree of the target change rate of the soil organic matter content in the agricultural production area as an index, and in combination with a predefined evaluation function, the comprehensive efficacy score of this group of candidate dynamic regulation setting amounts of the agricultural management measures is calculated based on the expected change trajectory; According to the comprehensive efficacy scores of each group of candidate dynamic regulation setting amounts of the agricultural management measures, and within the allowable adjustment range and constraint conditions, a new candidate dynamic regulation setting amount of the agricultural management measures is generated and returned to the input soil organic matter dynamic evolution prediction model to perform iterative update until the preset maximum number of iterations is reached; Select the candidate with the optimal comprehensive efficacy score as the dynamic regulation setting amount of the agricultural management measures for output.

7. The agricultural production area environment monitoring big data analysis method according to claim 6, characterized in that The specific calculation formula of the optimization objective function with the achievement degree of the target change rate of the soil organic matter content in the agricultural production area as an index is: Among them, is the optimization objective function, is the total number of predetermined spatial grid cells within the agricultural production area, is a preset normal constant adjustment factor, and j is the index of the spatial grid cell, is the predicted soil organic matter content of the j-th spatial grid cell, is the end time of the prediction period, is the initial soil organic matter content of the j-th predetermined spatial grid cell, is the target change rate of the soil organic matter content of the j-th predetermined spatial grid cell.

8. The agricultural production area environmental monitoring big data analysis method according to claim 7, characterized in that The calculation formula of the evaluation function is: Among them, is the evaluation value, , are the weight coefficients respectively, is the economic cost, is the environmental impact assessment value, and m is the dynamic regulation setting quantity of the current candidate agricultural management measure.

9. The agricultural production area environmental monitoring big data analysis method according to claim 8, wherein The calculation formula of the comprehensive efficacy score is: Among them, is the comprehensive effectiveness score set for the dynamic regulation of candidate farm management measures, , are the weight coefficients respectively.

10. The agricultural production area environmental monitoring big data analysis system is characterized in that, including: A data acquisition unit collects multi-dimensional spatio-temporal environmental data and crop growth status data covering agricultural production areas, and preprocesses the multi-dimensional spatio-temporal environmental data and crop growth status data to form spatio-temporal feature sequence data containing timestamps; A model construction unit constructs a dynamic evolution prediction model of soil organic matter. The dynamic evolution prediction model of soil organic matter is configured to: receive the spatio-temporal feature sequence data and corresponding farm management measures as inputs, and output a predicted value of the soil organic matter content in the agricultural production area; An optimization unit drives the dynamic evolution prediction model of soil organic matter to perform iterative optimization for a preset target change rate of the soil organic matter content in the agricultural production area, so as to determine a dynamic regulation setting amount of the farm management measures that enables the predicted value of the soil organic matter content to achieve the target change rate; An analysis unit inputs the dynamic regulation setting amount of the farm management measures, combined with the spatio-temporal feature sequence data input in real time, into the dynamic evolution prediction model of soil organic matter for simulation analysis, so as to predict the expected change trajectory and final achieved state of the soil organic matter content in the agricultural production area during the entire management cycle after implementing the farm management measures with the dynamic regulation setting amount, and generate an operation guidance plan for the agricultural production area based on the simulation analysis results.

Images