Farmland ecological meteorological monitoring method and device

Abstract

The invention provides a farmland ecological meteorological monitoring method and device, and the method comprises the steps: constructing a rainfall risk time sequence model based on historical daily scale air temperature and rainfall data; the current spring effective accumulated temperature is obtained and compared with the historical same period, and a climate anomaly factor is obtained; on the basis of the abnormal factors and the current observation sequence, the most probable hidden state at the current moment is determined through the forward probability, rolling prediction is conducted through the corrected transfer matrix from the current moment to the prediction moment, and the rainfall risk probability from the current moment to the prediction moment is obtained; comparing the rainfall risk probability with the soil moisture bearing capacity of the farmland to be monitored, and obtaining the risk probability actually borne by the farmland ecosystem as a first risk probability; based on real-time soil conditions and land parcel characteristics, estimating the probability of ecological risks caused by farming adjustment behaviors; a risk driving value is obtained based on the first risk probability, an ecological cost value is obtained based on the ecological risk probability, and when the risk driving value is larger than the ecological cost value, a final farming decision is generated, and the farming decision behavior is dynamically adjusted based on the current real-time climate through deep processing and association of the farmland ecological meteorological data.

Description

Technical Field

[0001] This application relates to the field of farmland ecological meteorological monitoring technology, specifically a farmland ecological meteorological monitoring method and device. Background Technology

[0002] Farmland ecological meteorological monitoring is a process of real-time and continuous data collection, analysis and management of climate environment, soil environment, crop physiological state and biological factors in farmland ecosystems, using sensors, communication and information processing.

[0003] Chinese patent CN111327654A discloses a method and system for monitoring farmland information. The method includes: identifying a monitoring device associated with a target account, wherein the monitoring device corresponds to the farmland to be monitored; receiving event information from the farmland transmitted by the monitoring device through an information sharing network, wherein the target account and the monitoring device are bound together in the information sharing network; and displaying the event information from the farmland. This application addresses the technical problem that relying on on-site observation or consultation to obtain farmland information can easily lead to incomplete information and a poor user experience.

[0004] This patent only focuses on receiving and displaying data, lacking in-depth data processing and correlation. It cannot provide feedback for agricultural activities based on the results of farmland ecological meteorological monitoring. In the autumn of 2025, in some areas of Henan Province, continuous heavy rains during the autumn harvest season led to waterlogging in farmland, preventing timely harvesting of crops. If the weather conditions during the harvest season could be predicted in advance, and different types of seeds or earlier sowing times could be selected, this situation could be avoided. Therefore, monitoring farmland ecological meteorology to guide specific agricultural activities has become a key issue that needs to be addressed. Summary of the Invention

[0005] To overcome the aforementioned problems in the prior art, this application provides a method and apparatus, which adopts the following technical solution:

[0006] Firstly, this application provides a method for monitoring farmland eco-meteorology, including:

[0007] Based on historical daily temperature and rainfall data, a time-series model of rainfall risk is constructed. The current effective accumulated temperature for spring is obtained and compared with historical data for the same period to identify climate anomaly factors. Based on these anomaly factors and the current observation sequence, the most likely hidden state at the current moment is determined using forward probability. From the current moment to the predicted moment, a rolling prediction is performed using a corrected transition matrix to obtain the rainfall risk probability from the current moment to the predicted moment. The rainfall risk probability is compared with the soil moisture carrying capacity of the farmland to be monitored to obtain the actual risk probability borne by the farmland ecosystem, which serves as the first risk probability. Based on real-time soil conditions and plot characteristics, the ecological risk probability caused by agricultural adjustments is estimated. A risk driver value is obtained based on the first risk probability, and an ecological cost value is obtained based on the ecological risk probability. When the risk driver value is greater than the ecological cost value, the final agricultural adjustment decision is generated.

[0008] Furthermore, a time-series model of rainfall risk is constructed, including: discretizing historical rainfall data based on a predefined rule for classifying observable rainfall states to obtain long-term historical observation sequences for model training; training the three core parameters of the rainfall risk time-series model based on daily temperature over a predefined time span, where the core parameters include an initial probability vector, a transition probability matrix, and an emission probability matrix; iteratively calculating the expected frequency of the hidden state sequence based on the current core parameters and the historical observation sequences, re-estimating the three core parameters based on the expected frequency until the convergence condition is met, and outputting the optimized initial probability vector, transition probability matrix, and emission probability matrix to obtain the trained rainfall risk time-series model.

[0009] Furthermore, the current effective accumulated temperature in spring is obtained and compared with the same period in history to obtain climate anomaly factors, including: setting temperature thresholds according to the target crop type, including the daily average temperature above the temperature threshold in the accumulated temperature, establishing a statistical baseline for the spring window based on the daily temperature over a preset time span; starting from the spring window, the current total effective accumulated temperature is accumulated in real time every day, and the standardized deviation between the current total effective accumulated temperature and the historical average accumulated temperature is used as anomaly factors.

[0010] Secondly, this application also provides a farmland ecological meteorological monitoring device, comprising:

[0011] A module for constructing a time series model of rainfall risk is used to build a time series model of rainfall risk based on historical daily temperature and rainfall data;

[0012] The module for obtaining climate anomaly factors is used to obtain the current effective accumulated temperature in spring and compare it with the same period in history to obtain climate anomaly factors.

[0013] The rainfall risk probability acquisition module is used to determine the most likely hidden state at the current moment based on anomaly factors and the current observation sequence through forward probability. From the current moment to the prediction moment, rolling prediction is performed through the corrected transition matrix to obtain the rainfall risk probability from the current moment to the prediction moment.

[0014] The first risk probability acquisition module is used to compare the rainfall risk probability with the soil moisture carrying capacity of the farmland to be monitored, and obtain the actual risk probability that the farmland ecosystem can bear, which is used as the first risk probability.

[0015] The ecological risk probability acquisition module is used to estimate the probability of ecological risks caused by agricultural adjustment behaviors based on real-time soil conditions and plot characteristics.

[0016] The overall ecological benefit score acquisition module is used to obtain the risk driving value based on the first risk probability and the ecological cost value based on the ecological risk probability. When the risk driving value is greater than the ecological cost value, the final agricultural adjustment decision is generated.

[0017] Thirdly, this application provides an electronic device, comprising:

[0018] One or more processors; a memory; and one or more computer programs, wherein the one or more computer programs are stored in the memory, and the one or more computer programs include instructions that, when executed by the device, cause the device to perform the method as described in the first aspect.

[0019] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when run on a computer, causes the computer to perform the method described in the first aspect.

[0020] Fifthly, this application provides a computer program that, when executed by a computer, performs the method described in the first aspect.

[0021] In one possible design, the program in the fifth aspect can be stored wholly or partially on a storage medium packaged with the processor, or it can be stored wholly or partially on a memory not packaged with the processor.

[0022] This application has the following beneficial effects:

[0023] 1. This application constructs a time-series model of rainfall risk based on historical daily temperature and rainfall data; it obtains the current effective accumulated temperature in spring and compares it with historical data for the same period to identify climate anomaly factors. By introducing these anomaly factors, this application dynamically corrects the rainfall risk time-series model, ensuring that the prediction results are based on current climate change trends and exhibit dynamic adaptability.

[0024] 2. This application determines the most probable hidden state at the current moment based on anomaly factors and the current observation sequence using forward probability. From the current moment to the predicted moment, a rolling prediction is performed using a corrected transition matrix to obtain the rainfall risk probability from the current moment to the predicted moment. The rainfall risk probability is compared with the soil moisture carrying capacity of the farmland to be monitored to obtain the actual risk probability that the farmland ecosystem can bear, which is used as the first risk probability. The first risk probability in this application is screened by soil moisture carrying capacity, which enhances the reliability of subsequent agricultural decision-making.

[0025] 3. This application estimates the probability of ecological risks arising from agricultural adjustments based on real-time soil conditions and plot characteristics; it obtains a risk driver value based on the first risk probability and an ecological cost value based on the ecological risk probability. When the risk driver value is greater than the ecological cost value, a final agricultural adjustment decision is generated. Through the ecological sustainability balance constraint, the final agricultural decision prevents planting failures caused by unsuitable soil conditions.

[0026] 4. This application can accurately predict the risks of the autumn harvest several months later, enabling farmers to take adjustments to their strategies during spring sowing and avoid risks at the source. It breaks through the time limitations of traditional weather forecasts, realizes in-depth processing and correlation of farmland ecological meteorological data, and dynamically adjusts agricultural decision-making behavior based on the current real-time climate. Attached Figure Description

[0027] Figure 1 This is a flowchart of the farmland ecological meteorological monitoring method according to an embodiment of this application;

[0028] Figure 2 A flowchart illustrating the construction of rainfall risk probability in an embodiment of this application;

[0029] Figure 3 This is a flowchart illustrating the process of obtaining abnormal factors according to an embodiment of this application.

[0030] Figure 4 This is a flowchart illustrating the ecological risk assessment process according to an embodiment of this application.

[0031] Figure 5 This is a flowchart of the farmland ecological meteorological monitoring device according to an embodiment of this application. Detailed Implementation

[0032] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein in the specification of the application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application; the terms "comprising" and "having," and any variations thereof, in the specification, claims, and foregoing drawings of this application, are intended to cover non-exclusive inclusion. The terms "first," "second," etc., in the specification, claims, or foregoing drawings of this application are used to distinguish different objects, not to describe a particular order.

[0033] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0034] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0035] Example 1, please refer to Figure 1 The following is a flowchart of a farmland ecological meteorological monitoring method provided in an embodiment of this application. The specific implementation process of the method is as follows:

[0036] Step 101: Construct a time series model of rainfall risk based on historical daily temperature and rainfall data.

[0037] It should be noted that the preset time span can be meteorological data within 20 or 30 years. The preset time span can be set according to actual needs and can be set by those skilled in the art according to actual application scenarios. For example, the preset time span can be set to 5 years, 10 years, etc., and there is no limitation here.

[0038] Traditional weather forecasting can provide 7-15 day forecasts, while rainfall risk time series models, based on historical meteorological data, can provide probability forecasts for several months, transforming passive agricultural decision-making from passive 15-day adjustments into proactive long-term strategic adjustments.

[0039] In this embodiment of the application, based on the rainfall risk time series model, the soil structure and permeability data of the farmland to be monitored are periodically acquired, the soil moisture carrying capacity of the farmland in the monitoring area under continuous rainfall is calculated as an ecological risk limiting factor, and the frequency of historical continuous rainfall exceeding the threshold of the ecological risk limiting factor is calculated as the basic risk probability.

[0040] It should be noted that a time series model of rainfall risk can be built based on historical climate databases. This model can be used to realize the transition between rainfall and non-rainfall states and to predict the probability of continuous rainfall over multiple days.

[0041] Specifically, for constructing a time-series model of rainfall risk, please refer to... Figure 2 ,include:

[0042] Step 11: Based on the preset rules for dividing observable rainfall states, the historical rainfall data is discretized to obtain long-term historical observation sequences for model training. At the same time, the hidden states that cannot be directly observed by the rainfall risk time series model are defined.

[0043] Specifically, the latent states, which cannot be directly observed, are defined based on meteorological and agronomic experience. These latent states include cold-wet patterns, warm-dry patterns, normal patterns, and warm-wet patterns. Cold-wet patterns are characterized by low temperatures, which carry a risk of high rainfall; warm-dry patterns are characterized by high temperatures, which carry a risk of low rainfall; normal patterns are characterized by temperatures close to the historical average, which carry a risk of moderate rainfall; and warm-wet patterns are characterized by temperatures above the seasonal average, which carry a risk of high-intensity, short-duration rainfall and are prone to convective heavy rainfall or tropical cyclones. This application utilizes these latent states to learn the transition probabilities between them, enabling long-term climate change modeling.

[0044] Specifically, rainfall data is discretized into observation states such as no rainfall, light rain, and heavy rainfall. During discretization, the rainfall range for each discrete state is recorded, and a representative rainfall amount is associated with each observation state. For example, (25mm / d).

[0045] Step 12: Train the three core parameters of the precipitation risk time series model based on daily temperature data over a preset time span. The core parameters include the initial probability vector. Transition probability matrix and emission probability matrix .

[0046] Specifically, the initial probability vector Randomly assign the probability of each hidden state at the beginning of the sequence; initialize the transition probability matrix. Randomly assign the probabilities of transitions between hidden states, with each row summing to 1; initialize the emission probability matrix. : Randomly assign the probability of each hidden state emitting each observed state, with each row summing to 1. Specifically, calculate the probability of being in a certain hidden state at any time given the occurrence of the observed sequence, and the probability of transitioning from the first state to the second state at any time.

[0047] Step 13: Calculate the expected frequency of the hidden state sequence based on the current core parameters and historical observation sequences, and re-estimate the three core parameters based on the expected frequency.

[0048] Specifically, the transition probability in the updated core parameters is the expected number of times a sequence of observations transitions from the first state to the second state, divided by the total expected number of times a sequence of observations originates from the first state.

[0049] Step 14: Repeat step 13 until the convergence condition is met, and output the optimized initial probability vector. Transition probability matrix and emission probability matrix Obtain the trained time series model of rainfall risk.

[0050] Step 102: Obtain the current effective accumulated temperature for spring and compare it with historical data for the same period to identify climate anomaly factors. For details, please refer to... Figure 3 ,include:

[0051] Step 21: Set a temperature threshold according to the target crop type. The daily average temperature above the temperature threshold is included in the accumulated temperature. Based on the daily temperature over a preset time span, establish a statistical baseline for the spring window.

[0052] It should be noted that the daily effective accumulated temperature is the absolute value of the difference between the daily average temperature and the temperature threshold. Assuming the current spring season runs from March 1st to May 31st, this time window is set as the start and end dates of the current spring season. The daily effective accumulated temperature is defined as follows: ,in The average daily temperature This represents the temperature threshold.

[0053] Specifically, establishing a statistical baseline for the spring window includes: for each year within a preset time span, calculating the total effective accumulated temperature, historical average accumulated temperature, and historical standard deviation within the spring window, and using the historical average accumulated temperature and historical standard deviation as the statistical baseline. The total effective accumulated temperature can be expressed as... ,in For the total effective accumulated temperature, This refers to the daily effective accumulated temperature. Specifically, it refers to the historical average accumulated temperature. The average value of the total effective accumulated temperature series, and the historical standard deviation. This represents the standard deviation of the total effective accumulated temperature series. The calculation of the mean and standard deviation is a prior art technique and will not be elaborated here.

[0054] Step 22: Starting from the spring window, accumulate the current total effective accumulated temperature in real time every day, and use the standardized deviation between the current total effective accumulated temperature and the historical average accumulated temperature as the outlier factor.

[0055] Specifically, ,in This represents the current total effective accumulated temperature. This represents the effective accumulated temperature for the current spring. The anomaly factor can be expressed as: ,in The historical average accumulated temperature, This represents the historical standard deviation. The meaning of outliers: when... The time indicates that the current spring season is significantly warmer than usual. The time indicates that the current spring season is significantly colder than usual. This indicates that there are significant climate anomalies in the current spring season.

[0056] Step 103: Based on the anomaly factor and the current observation sequence, determine the most likely hidden state at the current moment through forward probability, and perform rolling prediction from the current moment to the prediction moment through the corrected transition matrix to obtain the rainfall risk probability from the current moment to the prediction moment.

[0057] It should be noted that the anomaly factor, as input to the rainfall risk time series model, corrects the transition probability matrix and initial probability vector in the rainfall risk time series model, obtaining the corrected core parameters. For example: when If it can be determined that the current latent state tends towards a warm and dry pattern, then the initial probability of heavy rainfall during the autumn harvest period will be reduced, as will the probability of transitioning from a warm and dry to a cold and humid pattern.

[0058] It should be noted that the current observation sequence is real-time meteorological data for the current spring season.

[0059] In this embodiment of the application, based on the anomaly factor and the current observation sequence, the most probable hidden state at the current moment is determined by forward probability, including:

[0060] The initial probability vector and transition probability matrix of the rainfall risk time series model are corrected by using the Sigmoid function to obtain the corrected initial probability vector and corrected transition probability matrix.

[0061] Based on all rainfall data from the start of the observation sequence to the current moment, the probability of each hidden state occurring at the current moment is used to determine the starting state, with the hidden state having the highest probability being selected.

[0062] It should be noted that the forward probability is in time. Observed partial sequences In time The hidden state is exactly in the state The joint probability, i.e. .

[0063] In this embodiment of the application, from the current time to the predicted time, rolling prediction is performed using a modified transition matrix to obtain the rainfall risk probability from the current time to the predicted time. This includes: from the initial state to the predicted time, multi-step rolling prediction is performed based on the modified transition probability matrix to obtain a specific hidden state sequence for each future time point. The probability of continuous rainfall during the critical agricultural period is calculated using the specific hidden state sequence and the emission probability matrix, which is then used as the rainfall risk probability.

[0064] Specifically, the calculation time The forward probability at time step, i.e. ,in The rainfall risk time series model is in a hidden state at the beginning. The initial probability, Hidden state The first observed value The probability of emission. From the interval to Perform a loop and calculate the time. Forward probability , ,right Perform summation to obtain the summation from time... All hidden states Transition to the current time step Hidden state The total probability when calculating up to the last time step of the sequence. At that time, the forward probabilities of all hidden states are summed to obtain the risk probability of the entire observation sequence. The risk probability here is the observed sequence. The probability, that is, the likelihood of a heavy rainfall risk event occurring.

[0065] It should be noted that the key agricultural period here is set for the autumn harvest, but this can be adjusted according to actual needs.

[0066] Step 104: Compare the probability of rainfall risk with the soil moisture carrying capacity of the farmland to be monitored to obtain the actual risk probability that the farmland ecosystem can withstand, which is used as the first risk probability.

[0067] It should be noted that real-time soil conditions include current soil temperature, humidity, and permeability, while plot characteristics include slope, soil texture, and organic matter content. Adjustment actions include advancing the sowing time.

[0068] It should be noted that the first risk probability here is the probability value after screening the soil moisture carrying capacity of the farmland to be monitored. When the rainfall does not exceed the soil moisture carrying capacity threshold, the rainfall risk probability obtained through step 103 does not pose a risk to the soil of the farmland to be monitored.

[0069] Specifically, based on the characteristics of the plots collected over a preset time span, the soil moisture carrying capacity threshold of farmland plots under continuous heavy rainfall is determined. The rainfall amount during the critical autumn harvest period is obtained based on the rainfall risk probability. When the rainfall amount exceeds the soil moisture carrying capacity threshold, the rainfall risk probability limited by the actual ecological carrying capacity is obtained.

[0070] Specifically, obtaining the rainfall amount during the critical autumn harvest period based on the probability of rainfall risk can be expressed as obtaining the expected daily rainfall: ,in Indicates in Daily observation status The probability, For state The represented rainfall. The rainfall for consecutive rainy periods is obtained by superimposing the expected daily rainfall. This can be expressed as: Indicates continuity Total rainfall intensity forecast for the day, This marks the beginning of the heavy rainfall.

[0071] The soil moisture carrying capacity threshold represents the maximum amount of rainfall that the soil can absorb.

[0072] Step 105: Based on real-time soil conditions and plot characteristics, estimate the probability of ecological risks arising from agricultural adjustments. Real-time soil conditions include real-time soil temperature and humidity, and plot characteristics include slope, slope length, soil texture, and organic matter. Please refer to [link / reference needed] for details. Figure 4 :

[0073] Step 41: Based on the slope, slope length, soil texture, and organic matter of the farmland plot, construct the inherent soil erodibility as a base multiplier. The higher the slope and erodibility, the larger the base multiplier. The inherent soil erodibility in this application can be calculated using existing techniques, i.e., using the soil texture, organic matter, and permeability of the plot, and according to the general soil loss equation, calculate the inherent soil erodibility of the plot, i.e., the K-factor. The K-factor is a static parameter of the plot, quantifying the soil's sensitivity to rainfall erosion, and is a key base multiplier, which will not be elaborated upon here.

[0074] Step 42 involves introducing real-time soil conditions for dynamic correction. This involves comparing real-time humidity with arable humidity. When real-time humidity is significantly higher than arable humidity, a humidity penalty is applied; when real-time temperature is too low, a temperature penalty is applied. A significantly higher real-time humidity indicates that tillage at this time will disrupt soil aggregates. A lower temperature indicates that tillage at this time will reduce seed germination rates.

[0075] Step 43: Multiply the base multiplier, humidity penalty, and temperature penalty to output a standardized value, which serves as the ecological risk assessment. When the standardized value is close to 1, it indicates that the ecological risk of current cultivation is extremely high. Specifically, this can be expressed as: ,in This is the ecological risk assessment value. Basic multiplier, As a soil erodibility factor, Slope length factor As a humidity penalty, Temperature penalties are applied. Using a general soil loss equation, combined with soil erodibility factors, slope and slope length factors, and temperature and humidity penalties, the ecological costs of agricultural adjustments are quantified.

[0076] Step 106: Obtain the risk driving value based on the first risk probability and the ecological cost value based on the ecological risk probability. When the risk driving value is greater than the ecological cost value, the final agricultural adjustment decision is generated.

[0077] Specifically, the risk driver value obtained based on the first risk probability can be expressed as: The ecological cost value obtained based on the ecological risk probability can be expressed as: ,in Risk-driven value, As ecological cost value, This represents the filtered probability of rainfall risk. Weights for user risk preferences The sensitivity coefficient for the growth period of crops. Weighting based on ecological benefits.

[0078] The risk driver value is obtained based on the first risk probability, and the ecological cost value is obtained based on the ecological risk probability. When the risk driver value is greater than the ecological cost value, the final agricultural adjustment decision is generated. When the first risk probability is significantly greater than the ecological risk probability, the final agricultural adjustment decision is generated.

[0079] It should be noted that generating the final agricultural adjustment decision includes advancing sowing, and selecting waterlogging-resistant seeds when advancing sowing is not possible.

[0080] Example 2: Please refer to Figure 5 The present application provides an embodiment of a farmland ecological meteorological monitoring device, the specific contents of which include:

[0081] A module for constructing a time series model of rainfall risk is used to build a time series model of rainfall risk based on historical daily temperature and rainfall data.

[0082] The module for obtaining climate anomaly factors is used to obtain the current effective accumulated temperature in spring and compare it with the same period in history to obtain climate anomaly factors.

[0083] The rainfall risk probability acquisition module is used to determine the most likely hidden state at the current moment based on anomaly factors and the current observation sequence through forward probability. From the current moment to the prediction moment, rolling prediction is performed through the corrected transition matrix to obtain the rainfall risk probability from the current moment to the prediction moment.

[0084] The first risk probability acquisition module is used to compare the rainfall risk probability with the soil moisture carrying capacity of the farmland to be monitored, and obtain the actual risk probability that the farmland ecosystem can bear, which is used as the first risk probability.

[0085] The ecological risk probability acquisition module is used to estimate the probability of ecological risks caused by agricultural adjustment behaviors based on real-time soil conditions and plot characteristics.

[0086] The overall ecological benefit score acquisition module is used to obtain the risk driving value based on the first risk probability and the ecological cost value based on the ecological risk probability. When the risk driving value is greater than the ecological cost value, the final agricultural adjustment decision is generated.

[0087] Example 3: This application provides an electronic device, including: one or more processors; a memory; and one or more computer programs, wherein the one or more computer programs are stored in the memory, and the one or more computer programs include instructions that, when executed by the device, cause the device to perform the method as described in Example 1.

[0088] Example 4: This application provides a computer-readable storage medium storing a computer program that, when run on a computer, causes the computer to perform the method described in Example 1.

[0089] Example 5: This application provides a computer program that, when executed by a computer, performs the method described in Example 1.

[0090] This application constructs a time-series model of rainfall risk based on historical daily temperature and rainfall data; obtains the current effective accumulated temperature in spring and compares it with the historical data for the same period to obtain climate anomaly factors; based on the anomaly factors and the current observation sequence, it determines the most likely hidden state at the current moment through forward probability, and performs rolling prediction from the current moment to the prediction moment through a corrected transition matrix to obtain the rainfall risk probability from the current moment to the prediction moment; it compares the rainfall risk probability with the soil moisture carrying capacity of the farmland to obtain the actual risk probability borne by the farmland ecosystem as the first risk probability; based on real-time soil conditions and plot characteristics, it estimates the ecological risk probability caused by agricultural adjustment behavior; it obtains the risk driving value based on the first risk probability and the ecological cost value based on the ecological risk probability; when the risk driving value is greater than the ecological cost value, the final agricultural adjustment decision is generated. This application deeply processes and correlates farmland ecological meteorological data to dynamically adjust agricultural decision-making behavior based on the current real-time climate.

[0091] One possible use case example: Consider a cornfield on a plain, with loam soil of moderate permeability. The target crop is corn, and the decision-making time is mid-April. The goal is to mitigate the risk of soil waterlogging caused by continuous heavy rainfall during the autumn harvest in October.

[0092] Step 1: Obtain daily temperature and rainfall data from the past 20 years, construct a rainfall risk time series model, and calculate the soil moisture carrying capacity threshold by analyzing the soil structure and permeability data of the plot: large-scale waterlogging will be caused when the rainfall exceeds 45 mm for 3 consecutive days.

[0093] Step 2: The current real-time soil temperature is 11°C and the humidity is 28%. The current effective accumulated temperature in spring up to mid-April is higher than the historical average for the same period. Therefore, the anomaly factor value is 1.8, indicating that the current spring is warmer than usual.

[0094] Step 3: Using the anomaly factor 1.8 as input to the rainfall risk time series model, the transition probability matrix of the rainfall risk time series model is corrected, increasing the probability of a warm-dry pattern and decreasing the probability of a cold-wet pattern. Rolling forecasts are then performed using the corrected rainfall risk time series model, yielding a 45% probability of rainfall exceeding 45mm for three consecutive days during the autumn harvest season.

[0095] Step 4: Based on the current soil temperature and humidity and plot structure information, the ecological risk probability is 0.2, which is considered low to medium risk.

[0096] Step 5: Since farmers are risk-averse, the user risk preference weight is set to 1.4; since the plot is located near a water source protection area, the ecological benefit weight is set to 1.3; and the crop growth period sensitivity coefficient is set to 0.9. Therefore, the risk-driven value is 45%. 1.4 = 0.63, the ecological cost is 0.2 1.3 0.9 = 0.234. Therefore, the risk-driven value minus the ecological cost value is 0.396, indicating that the risk-driven value is greater than the ecological cost value. Thus, agricultural adjustments are necessary, meaning that the current soil temperature and humidity conditions meet the safety threshold for early corn planting.

[0097] Obviously, the embodiments described above are only some embodiments of this application, not all embodiments. The accompanying drawings show preferred embodiments of this application, but do not limit the patent scope of this application. This application can be implemented in many different forms; rather, the purpose of providing these embodiments is to provide a more thorough and comprehensive understanding of the disclosure of this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or make equivalent substitutions for some of the technical features. Any equivalent structures made using the content of this application's specification and drawings, directly or indirectly applied to other related technical fields, are similarly within the scope of patent protection of this application.

Claims

1. A method for monitoring farmland ecological meteorology, characterized in that, include: A time series model of rainfall risk is constructed based on historical daily temperature and rainfall data; Obtain the current effective accumulated temperature for spring and compare it with the same period in history to identify climate anomaly factors; Based on the anomaly factor and the current observation sequence, the most likely hidden state at the current moment is determined by forward probability. From the current moment to the predicted moment, rolling prediction is performed by the corrected transition matrix to obtain the rainfall risk probability from the current moment to the predicted moment. By comparing the probability of rainfall risk with the soil moisture carrying capacity of the farmland to be monitored, the actual risk probability that the farmland ecosystem can withstand is obtained as the first risk probability; Based on real-time soil conditions and plot characteristics, the probability of ecological risks caused by agricultural adjustment behaviors is estimated. The risk driver value is obtained based on the first risk probability, and the ecological cost value is obtained based on the ecological risk probability. When the risk driver value is greater than the ecological cost value, the final agricultural adjustment decision is generated.

2. The method for monitoring farmland ecological meteorology according to claim 1, characterized in that, Constructing a time-series model of rainfall risk includes: Based on the pre-defined rules for classifying observable rainfall states, historical rainfall data is discretized to obtain long-term historical observation sequences for model training. The three core parameters of the time series model of rainfall risk are trained based on daily temperature over a preset time span. The core parameters include the initial probability vector, the transition probability matrix, and the emission probability matrix. Based on the current core parameters and historical observation sequences, the expected frequency of the hidden state sequence is calculated iteratively. The three core parameters are re-estimated based on the expected frequency until the convergence condition is met. The optimized initial probability vector, transition probability matrix and emission probability matrix are output to obtain the trained rainfall risk time series model.

3. The method for monitoring farmland ecological meteorology according to claim 2, characterized in that, When obtaining long-term historical observation sequences for model training, it is also necessary to define the hidden states that cannot be directly observed in the rainfall risk time series model.

4. The method for monitoring farmland ecological meteorology according to claim 1, characterized in that, Obtain the current effective accumulated temperature for spring and compare it with historical data for the same period to identify climate anomaly factors, including: Temperature thresholds are set according to the target crop type. Daily average temperatures above the temperature thresholds are counted as accumulated temperature. A statistical baseline is established for the spring window based on daily temperature over a preset time span. Starting from the spring window, the current total effective accumulated temperature is accumulated daily in real time, and the standardized deviation between the current total effective accumulated temperature and the historical average accumulated temperature is used as an outlier.

5. The method for monitoring farmland ecological meteorology according to claim 4, characterized in that, Establishing a statistical baseline for the spring window includes: For each year within a preset time span, calculate the total effective accumulated temperature, historical average accumulated temperature, and historical standard deviation within the spring window, and use the historical average accumulated temperature and historical standard deviation as the statistical baseline.

6. The method for monitoring farmland ecological meteorology according to claim 1, characterized in that, Based on the anomaly factor and the current observation sequence, the most probable hidden state at the current moment is determined through forward probability, including: The initial probability vector and transition probability matrix of the rainfall risk time series model are corrected by using the Sigmoid function to obtain the corrected initial probability vector and corrected transition probability matrix. Based on all rainfall data from the start of the observation sequence to the current moment, the probability of each hidden state occurring at the current moment is used to determine the starting state, with the hidden state having the highest probability being selected.

7. The method for monitoring farmland ecological meteorology according to claim 1, characterized in that, From the current time to the forecast time, a rolling forecast is performed using the corrected transition matrix to obtain the probability of rainfall risk from the current time to the forecast time, including: Starting from the initial state to the prediction time, multi-step rolling prediction is performed based on the modified transition probability matrix to obtain a specific hidden state sequence at each future time point. By using the specific hidden state sequence and the emission probability matrix, the probability of continuous rainfall during the critical agricultural period is calculated as the rainfall risk probability.

8. A farmland ecological meteorological monitoring device, used to implement the methods of claims 1-7, characterized in that, include: A module for constructing a time series model of rainfall risk is used to build a time series model of rainfall risk based on historical daily temperature and rainfall data; The module for obtaining climate anomaly factors is used to obtain the current effective accumulated temperature in spring and compare it with the same period in history to obtain climate anomaly factors. The rainfall risk probability acquisition module is used to determine the most likely hidden state at the current moment based on anomaly factors and the current observation sequence through forward probability. From the current moment to the prediction moment, rolling prediction is performed through the corrected transition matrix to obtain the rainfall risk probability from the current moment to the prediction moment. The first risk probability acquisition module is used to compare the rainfall risk probability with the soil moisture carrying capacity of the farmland to be monitored, and obtain the actual risk probability that the farmland ecosystem can bear, which is used as the first risk probability. The ecological risk probability acquisition module is used to estimate the probability of ecological risks caused by agricultural adjustment behaviors based on real-time soil conditions and plot characteristics. The overall ecological benefit score acquisition module is used to obtain the risk driving value based on the first risk probability and the ecological cost value based on the ecological risk probability. When the risk driving value is greater than the ecological cost value, the final agricultural adjustment decision is generated.

9. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-7.

10. A computer-readable storage medium storing computer instructions thereon, characterized in that, When executed by the processor, this instruction implements the steps of the method as described in any one of claims 1-7.

Images