Irrigation area farmland irrigation method based on optimization algorithm

Abstract

The invention relates to the technical field of agricultural production, in particular to an irrigation district farmland irrigation method based on an optimization algorithm, and the method comprises the steps: obtaining related historical data of a farmland irrigation district, and carrying out the preprocessing of the related historical data; correlation degree analysis is carried out on related historical data through grey correlation analysis, and a BP optimization crop model is constructed based on correlation degree; calculating a crop yield prediction error through the BP optimization crop model, and constructing a BP-Markov crop model based on the prediction error; and constructing an irrigation area farmland irrigation model based on the related historical data, and obtaining a farmland irrigation scheme based on the BP-Markov crop model and the irrigation area farmland irrigation model. According to the method, the BP-Markov crop model and the farmland irrigation model are constructed, and a proper irrigation area farmland irrigation scheme is provided, so that water resources are saved to the greatest extent under the condition of meeting crop irrigation requirements.

Description

Technical Field

[0001] The present application relates to the field of agricultural irrigation technology, and in particular to a method for irrigating farmland in an irrigation area based on an optimization algorithm. Background Art

[0002] At present, my country's agricultural irrigation model is relatively backward, and the traditional irrigation model based on experience is still widely used. Under this irrigation model, not only is the utilization efficiency of water resources low, resulting in water waste and aggravating the problem of water shortage in agricultural irrigation, but on the other hand, too much irrigation water will cause deep leakage and nitrogen and phosphorus loss, causing secondary salinization of the soil and other problems, which not only affects the normal growth of crops and does not reduce crop yields, but also pollutes groundwater and causes environmental and ecological problems. At the same time, crop growth requires a certain amount of water, and insufficient water will seriously affect crop yields; the demand for water at different stages of the crop growth cycle is also different. Therefore, how to irrigate in a timely and appropriate manner to improve water resource utilization while ensuring crop yields and reduce agricultural chemical pollution is an important topic.

[0003] In view of this, the present application proposes a method for irrigating farmland in an irrigated area based on an optimization algorithm. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In view of the above-mentioned shortcomings and deficiencies of the prior art, the present application provides an irrigation method for farmland in an irrigated area based on an optimization algorithm.

[0006] (II) Technical solution

[0007] In order to achieve the above objectives, this application adopts the following technical solutions:

[0008] In a first aspect, the present application provides an irrigation method for farmland in an irrigation area based on an optimization algorithm, comprising:

[0009] S1. Obtain relevant historical data of the farmland irrigation area and pre-process the relevant historical data;

[0010] The relevant historical data include: historical multi-source data, historical crop yield data and historical farmland irrigation data;

[0011] S2. Obtaining the correlation value between the historical multi-source data and the historical crop yield data through grey correlation analysis, and constructing a BP optimization crop model through a BP neural network based on the correlation value;

[0012] S3. Based on historical multi-source data and historical crop yield data, obtain crop yield prediction error by optimizing the crop model through BP; based on the crop yield prediction error, construct a BP-Markov crop model through Markov method;

[0013] S4, constructing an irrigation model for the irrigation area based on the historical crop yield data and the historical farmland irrigation data;

[0014] S5. For the farmland irrigation area, the crop growth data of the future time period is obtained through the BP-Markov crop model, and based on the crop growth data of the future time period, the farmland irrigation plan is obtained through the farmland irrigation model of the irrigation area.

[0015] Optionally, the correlation value between the historical multi-source data and the historical crop yield data obtained by grey correlation analysis in S2 includes:

[0016] The calculation process of the correlation value between the historical multi-source data and the historical crop yield data is as follows:

[0017] Δ i (k)=|x0(k)-x i (k)|

[0018]

[0019] Among them, Δ i (k) is the absolute difference corresponding to the i-th type of multivariate data in the k-th year, x0(k) is the crop yield in the k-th year, x i (k) is the i-th type of multi-source data in the k-th year, ε i (k) is the correlation coefficient, ρ is the resolution coefficient, Δ min is the minimum absolute difference, Δ max is the maximum absolute difference, r i is the correlation degree, and N is the number of multi-source data types.

[0020] Optionally, constructing a BP optimization crop model based on the correlation value through a BP neural network in S2 includes:

[0021] Based on the historical multi-source data obtained, according to the calculated correlation r i Determine the factors affecting crop yield, and construct a BP optimization crop model through a BP neural network based on the factors;

[0022] The construction method of the BP optimization crop model is:

[0023] S2-1. Determine the number of input layers, output layers and hidden layers of the BP neural network;

[0024] S2-2, optimize the connection weights and thresholds in the BP neural network based on genetic algorithm;

[0025] The connection weights include: connection weights from the input layer to the hidden layer, and connection weights from the hidden layer to the output layer.

[0026] Optionally, the step of obtaining the crop yield prediction error by optimizing the crop model through BP based on the historical multi-source data and the historical crop yield data in S3 includes:

[0027] S3-1, using historical multi-source data as input variables of the BP optimization crop model to obtain predicted crop yields;

[0028] S3-2, calculating the crop yield prediction error based on the predicted crop yield and the actual crop yield;

[0029] The calculation method of the crop yield prediction error ω(k) is:

[0030] ω(k)=|x0(k)-x0'(k)|

[0031] Among them, x0(k) is the crop yield in the kth year, x0'(k) is the predicted crop yield in the kth year, and ω(k) is the prediction error of the crop yield in the kth year.

[0032] Optionally, the step of constructing a BP-Markov crop model based on the crop yield prediction error by using the Markov method in S3 includes:

[0033] A crop yield prediction error state transfer matrix is ​​obtained based on the crop yield prediction error, and based on the crop yield prediction error state transfer matrix, the predicted crop yield obtained by using the BP optimization crop model is corrected.

[0034] Optionally, a crop yield prediction error state transfer matrix is ​​obtained based on the crop yield prediction error, including:

[0035] Divide the obtained crop yield prediction error into state intervals, and count the states and state transitions corresponding to the predicted crop yield each year, and obtain the crop yield prediction error state transition matrix based on the state transitions;

[0036] The crop yield prediction error state transfer matrix P is expressed as:

[0037]

[0038] Among them, r is the total number of states corresponding to the crop yield prediction error, P ab is the probability of transitioning from state a to state b.

[0039] Optionally, the step of constructing an irrigation model for an irrigation area based on the historical crop yield data and the historical farmland irrigation data in S4 includes:

[0040]

[0041] Among them, Y a is the actual crop yield, Y m is the maximum yield of crops, ET a The actual evaporation of crops, ET m is the potential evapotranspiration of crops, λ i is the crop sensitivity index at stage i, and i is the crop growth stage number.

[0042] Optionally, the step of obtaining crop growth data of a future time period for the farmland irrigation area through the BP-Markov crop model, and obtaining a farmland irrigation plan through the farmland irrigation model of the irrigation area based on the crop growth data of the future time period, includes:

[0043] The historical multi-source data screened by grey correlation analysis are used as input variables of the BP-Markov crop model to obtain crop growth data for future time periods;

[0044] The obtained crop growth data for the future time period is used as the corresponding actual crop yield in the irrigation model of the irrigation area farmland, and the irrigation water use plan for the irrigation area farmland in the future time period is obtained.

[0045] In a second aspect, the present application provides a computing device and a computer storage medium, including:

[0046] A computer program is stored on the computing device, and when the program is executed by the processor, the steps of the method described in the first aspect are implemented.

[0047] The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the steps of the method described in the first aspect are implemented.

[0048] (III) Beneficial effects

[0049] In this application, a suitable irrigation scheme for farmland in the irrigation area is proposed by combining the constructed BP-Markov crop model and farmland irrigation model, thereby achieving maximum water resource conservation while meeting the crop irrigation needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The present application is described with the aid of the following drawings:

[0051] Figure 1This is a schematic flow chart of a method for irrigation of farmland in an irrigation area based on an optimization algorithm in one embodiment of the present application;

[0052] Figure 2 A schematic diagram of the Tukey box plot method in one embodiment of the present application. DETAILED DESCRIPTION

[0053] In order to better explain the present invention and facilitate understanding, the present invention is described in detail below in conjunction with the accompanying drawings through specific implementation methods. It is understood that the specific embodiments described below are only used to explain the relevant inventions, rather than to limit the invention. It should also be noted that the embodiments and features in the embodiments of this application can be combined with each other in the absence of conflict; for ease of description, only the parts related to the invention are shown in the accompanying drawings.

[0054] Embodiment 1 provides a method for irrigating farmland in an irrigation area based on an optimization algorithm, such as Figure 1 As shown, the details are as follows:

[0055] S1. Obtain relevant historical data of the farmland irrigation area, and pre-process the relevant historical data.

[0056] In this embodiment, the relevant historical data includes: historical multi-source data, historical crop yield data, and historical farmland irrigation data.

[0057] In this embodiment, the historical multi-source data includes: geographic data, soil data and meteorological data of the farmland irrigation area.

[0058] In this embodiment, the historical farmland irrigation data includes the annual, quarterly, and monthly rainfall and artificial irrigation water volume corresponding to different crops.

[0059] In this embodiment, preprocessing includes data collection, data cleaning and data fitting.

[0060] S2. Obtain the correlation value between the historical multi-source data and the historical crop yield data through grey correlation analysis, and construct a BP optimization crop model through a BP neural network based on the correlation value.

[0061] S3. Based on historical multi-source data and historical crop yield data, obtain crop yield prediction error through BP optimization crop model; based on the crop yield prediction error, construct BP-Markov crop model through Markov method.

[0062] S4. Constructing an irrigation model for the irrigation area based on the historical crop yield data and the historical farmland irrigation data.

[0063] S5. For the farmland irrigation area, the crop growth data of the future time period is obtained through the BP-Markov crop model, and based on the crop growth data of the future time period, the farmland irrigation plan is obtained through the farmland irrigation model of the irrigation area.

[0064] In this embodiment, the farmland irrigation plan specifically includes the total irrigation amount required for crops at different growth stages; based on the calculated total irrigation amount and combined with the precipitation in the weather forecast, the amount of water required for artificial irrigation is obtained by calculation.

[0065] The first embodiment discloses that a suitable irrigation scheme for the irrigation area is proposed by combining the constructed BP-Markov crop model and the farmland irrigation model, thereby achieving maximum water resource conservation while meeting the crop irrigation demand.

[0066] Embodiment 2 provides a method for irrigating farmland in an irrigated area based on an optimization algorithm, which is as follows:

[0067] S1. Obtain relevant historical data of the farmland irrigation area, and pre-process the relevant historical data.

[0068] Based on the aforementioned step S1, in this embodiment, the crops corresponding to the obtained farmland irrigation area may be three crops: wheat, corn and cotton, but are not limited to the above three crops.

[0069] Based on the aforementioned step S1, regarding the acquisition of relevant historical data of farmland irrigation areas, it should be noted that an optional method for acquiring relevant historical data includes:

[0070] To sort out the land moisture, meteorological and other data of the irrigation district's high-efficiency water-saving demonstration area in the past three years, historical remote sensing satellite images, historical climate data, engineering data and other data can be collected through irrigation district information system collection, field research, literature query and other methods. The specific data types can be divided into:

[0071] 1) Geographic information data; specifically, obtaining historical remote sensing satellite images, and obtaining geographic information data of the irrigation area based on the obtained historical remote sensing satellite images;

[0072] 2) Meteorological data; specifically, meteorological data include daily precipitation, solar radiation, daily minimum temperature and daily maximum temperature;

[0073] 3) Soil data; specifically, soil data includes soil type, evaporation coefficient, water potential gradient and soil moisture content;

[0074] 4) Crop data; specifically, the crop data include accumulated temperature, photoperiod, vernalization coefficient, and potential grain filling rate at each growth stage of three types of crops: wheat, corn, and cotton;

[0075] 5) Crop irrigation data; specifically, the crop irrigation data includes irrigation time and irrigation water volume;

[0076] 6) Irrigation field data; specifically, the irrigation field data includes the irrigation period, irrigation water volume and per mu yield of various crops in the Weishan irrigation area.

[0077] Based on the aforementioned step S1, for the preprocessing of the relevant historical data of the farmland irrigation area, it should be noted that the optional relevant historical data preprocessing methods include data collection, data cleaning and data fitting. The specific preprocessing of the relevant historical data is as follows:

[0078] First, fitting of missing data. For this embodiment, the main missing data include the area of ​​each plot in the irrigation area, the crops planted in each season and their planting area, as well as climate and meteorological data and soil moisture data;

[0079] In this embodiment, for the missing data fitting of the area of ​​each plot in the irrigation area, the crops planted in each season and their planting area, it should be noted that:

[0080] The irrigation area mainly grows cotton, corn and wheat. First, historical satellite remote sensing images of the area from different periods are collected, and then spectral separation technology is used to analyze the planting conditions and planting areas of crops in the soil based on the color difference of different crops.

[0081] In this embodiment, for the missing data fitting of climate, meteorology and soil moisture data, it should be noted that:

[0082] Due to the late construction of irrigation district informatization, there is a lack of historical data support for the climate, meteorology, soil moisture and other data of previous years. Therefore, the historical data can be simulated and fitted based on the recently collected data and the statistical characteristics of the data.

[0083] Second, the integrated data is cleaned. For this embodiment, due to sensor failure, manual input errors or abnormal events, the collected data contains "dirty data" such as duplicate data and abnormal data. Then, the Tukey box plot method in statistics can be used to monitor the abnormal data, and the mean or other normal values ​​can be used to correct the "dirty data" according to the specific situation. The above-mentioned Tukey box plot method is as follows: Figure 2 shown.

[0084] S2. Obtain the correlation value between the historical multi-source data and the historical crop yield data through grey correlation analysis, and construct a BP optimization crop model through a BP neural network based on the correlation value.

[0085] Based on the aforementioned step S2, for the correlation analysis between the historical multi-source data and the historical crop yield data, it should be noted that:

[0086] The calculation formulas for the correlation values ​​between historical multi-source data and historical crop yield data are shown in formulas (1) to (3):

[0087] Δ i (k)=|x0(k)-x i (k)| (1)

[0088]

[0089] Among them, Δ i (k) is the absolute difference corresponding to the i-th type of multivariate data in the k-th year, x0(k) is the crop yield in the k-th year, x i (k) is the i-th type of multi-source data in the k-th year, ε i (k) is the correlation coefficient, ρ is the resolution coefficient, Δ min is the minimum absolute difference, Δ max is the maximum absolute difference, r i is the correlation degree, and N is the number of multi-source data types.

[0090] In this embodiment, according to the calculated correlation value r i , sort the influencing factors corresponding to the historical multi-source data, and set the correlation value r i Large historical multi-source data as an influencing factor for BP optimization crop model.

[0091] In addition, based on the aforementioned step S2, it should be noted that:

[0092] In this embodiment, each type of historical multi-source data corresponds to an influencing factor of the BP optimization crop model, and the correlation degree r is analyzed by the aforementioned grey correlation degree. i Multi-source data with larger values ​​are used as the final input variables of the BP optimization crop model.

[0093] In this embodiment, in order to further improve the prediction accuracy of crop yield, it is necessary to use a genetic algorithm to optimize the constructed BP neural network. The specific optimization process includes:

[0094] S2-1. Determine the number of input layers, output layers and hidden layers of the BP neural network;

[0095] S2-2, optimize the connection weights and thresholds in the BP neural network based on genetic algorithm;

[0096] In this embodiment, the connection weights include: connection weights from the input layer to the hidden layer, and connection weights from the hidden layer to the output layer.

[0097] Based on the above step S2-2, the connection weights and thresholds in the BP neural network are optimized by a genetic algorithm. The specific optimization steps include:

[0098] S2-2-1, randomly initialize the population;

[0099] S2-2-2, train and test each individual in the population through the BP neural network, obtain the prediction error corresponding to the crop yield in the irrigation area, and construct the objective function based on the obtained prediction error;

[0100] S2-2-3, constructing a population fitness value based on the constructed objective function, and constructing a subpopulation based on the constructed fitness value;

[0101] S2-2-4, randomly select two individuals from the sub-population, and obtain a new sub-population by exchanging the variables to be optimized corresponding to the two selected individuals;

[0102] S2-2-5. Based on the new sub-population, randomly select one individual; for the selected individual, randomly select a corresponding variable to be optimized based on the mutation probability, and randomly mutate the corresponding variable to be optimized to obtain a final sub-population;

[0103] S2-2-6. Based on each individual in the final sub-population, the BP neural network is trained and tested to obtain the prediction error corresponding to the crop yield in the irrigation area and reconstruct the objective function;

[0104] S2-2-7. Based on the final subpopulation and the reconstructed objective function, obtain the first-generation optimal solution, optimal weight and optimal threshold for the crop yield in the irrigation area;

[0105] S2-2-8. Based on the final sub-population, repeat the above steps S2-2-3 to S2-2-6 until the final optimal solution, optimal weight and optimal threshold for the crops in the irrigation area are obtained.

[0106] S3. Based on historical multi-source data and historical crop yield data, obtain crop yield prediction error through BP optimization crop model; based on the crop yield prediction error, construct BP-Markov crop model through Markov method.

[0107] Based on the above step S3, the prediction error of crop yield is obtained. It should be noted that:

[0108] S3-1, using historical multi-source data as input variables of the BP optimization crop model to obtain predicted crop yields;

[0109] S3-2. Calculate the crop yield prediction error based on the predicted crop yield and the actual crop yield.

[0110] In this embodiment, the calculation formula of the crop yield prediction error ω(k) is shown in formula (4):

[0111] ω(k)=|x0(k)-x0'(k)| (4)

[0112] Among them, x0(k) is the crop yield in the kth year, x0'(k) is the predicted crop yield in the kth year, and ω(k) is the prediction error of the crop yield in the kth year.

[0113] Based on the above step S3, it should be noted that:

[0114] A crop yield prediction error state transfer matrix is ​​obtained based on the crop yield prediction error, and based on the crop yield prediction error state transfer matrix, the predicted crop yield obtained by using the BP optimization crop model is corrected.

[0115] In this embodiment, a crop yield prediction error state transfer matrix is ​​obtained based on the crop yield prediction error. Specifically, the obtained crop yield prediction error is divided into state intervals, and the states and state transitions corresponding to the annual predicted crop yield are counted, and the crop yield prediction error state transfer matrix is ​​obtained based on the state transitions.

[0116] In this embodiment, based on the yield prediction error subset ω={ω(1), ω(2), ...ω(k)}, the state interval is divided according to the numerical value corresponding to each individual in the yield prediction error subset. The specific interval division method is: interval 1, interval 2, ..., interval r. Based on the divided intervals, the yield prediction error state subset P'={P1, P2, ...P k )}, based on the obtained yield prediction error state subset P', the yield prediction error transfer situation is obtained through mathematical statistics, and the crop yield prediction error state transfer matrix P is obtained according to the yield prediction error transfer situation.

[0117] In this embodiment, the aforementioned crop yield prediction error state transfer matrix P is expressed by formula (5), specifically:

[0118] Among them, r is the total number of states corresponding to the crop yield prediction error, P ab is the probability of transitioning from state a to state b, a=1,2…,r, and b=1,2…,r.

[0119] S4. Constructing an irrigation model for the irrigation area based on the historical crop yield data and the historical farmland irrigation data.

[0120] In this embodiment, in order to better formulate the best irrigation plan for the irrigation area, it is necessary to obtain a mapping relationship based on the historical data of crop yield and farmland irrigation data, so as to achieve: obtaining farmland irrigation data by obtaining crop yield.

[0121] Based on the above step S4, for constructing the irrigation model of the irrigation area farmland, it should be noted that:

[0122] The irrigation model of farmland in the irrigation area is constructed based on the Jensen model. The expression corresponding to the Jensen model is shown in formula (6):

[0123]

[0124] In the above formula (6), Y a is the actual crop yield, Y m is the maximum yield of crops, ET a The actual evaporation of crops, ET m is the potential evapotranspiration of crops, λ i is the crop sensitivity index at stage i, and i is the crop growth stage number.

[0125] Based on the above formula (6), the actual evapotranspiration ET of crops is a The calculation of can be performed by the following formula (7):

[0126] (ET a ) i +W i +S i+1 =S i +d i +P i +K i (7)

[0127] In the above formula (7), W i is the penetration amount in the i-th stage, S i+1 is the planned soil water content in the wet layer in the i+1th stage, S i is the planned soil water content in the wet layer in stage i, d i is the irrigation volume in the i-th stage, P i is the effective rainfall in stage i, K i is the groundwater recharge in stage i.

[0128] Based on the above formula (6), the potential evapotranspiration of crops ET m On the one hand, the actual evapotranspiration of crops in the fully irrigated crop irrigation experiment can be approximated as the potential evapotranspiration of crops ET. m On the other hand, it can also be calculated by the following formula (8):

[0129]

[0130] In the above formula (8), R n is the net radiation of the crop canopy, G is the soil flux, γ is the thermometer constant, u2 is the wind speed at 2m, e a is the saturated water vapor pressure, e d is the actual water vapor pressure, and Δ is the slope of the saturated water vapor pressure-temperature curve.

[0131] In this embodiment, when the crops are fully irrigated, the theoretical value of the Jensen model expression corresponding to formula (6) is equal to 1; when the crops are not fully irrigated, the theoretical value of the Jensen model expression is maximized based on formula (6).

[0132] In this embodiment, regarding the solution process of the expression corresponding to the Jensen model, it should be noted that: since the formula (6) corresponding to the Jensen model is in the form of an exponential function, it belongs to a nonlinear multiplication model, and dynamic programming can be used to solve the model, and the Jensen model is decomposed into a multi-stage decision process. The mathematical model after dynamic programming is as follows:

[0133] (1) Stage variables, specifically including: the growth stage n corresponding to each crop, n = 1, 2, …, N;

[0134] (2) Decision variables, including: irrigation volume at each stage d i and actual evaporation ET a ;

[0135] (3) State variables, including: water supply q at each stage i and the planned soil water content S in the wet layer i ;

[0136] (4) Transfer equation. Since the transfer equation corresponds to the state variable, there are two transfer equations, namely the soil water balance equation and the water distribution equation. The specific expressions are shown in formula (7) and formula (9), which are as follows:

[0137] Soil water balance equation:

[0138] (ET a ) i +W i +S i+1 =S i +d i +P i +K i (7)

[0139] Water distribution equation:

[0140] q i+1 =q i -d i (9)

[0141] (5) Objective function: Under the condition of a certain amount of water available for distribution, the Jensen model seeks to maximize the yield per unit area, that is: Y a / Y m →1.0, which is specifically expressed by formula (10):

[0142]

[0143] (6) Recursive equation. In this embodiment, when solving the Jensen model, it is necessary to perform reverse recursion and sequential decision making. The corresponding recursive equations are shown in formulas (11) and (12):

[0144]

[0145]

[0146] (7) Constraints. In this embodiment, the constraints corresponding to the Jensen model include decision constraints and soil moisture constraints. The specific constraints are shown in formula (13):

[0147]

[0148] (8) Initial conditions. In this embodiment, the expressions of the initial soil moisture content and initial water supply corresponding to the Jensen model are shown in formula (14):

[0149]

[0150] In this embodiment, it is necessary to explain that the Jensen model has two decision variables and state variables, and thus is a two-dimensional dynamic programming problem. In the decision-making process, the change of one variable (irrigation volume) will inevitably cause the change of another variable (actual evapotranspiration), and thus is difficult to solve. Therefore, in this embodiment, the Jensen model is solved by the successive approximation method of dimensionality reduction. The specific processing method is as follows: first, an initial estimate of variable one (i.e., actual evapotranspiration) is given, and then variable two (i.e., irrigation volume) is estimated; then variable two (i.e., irrigation volume) is fixed, and variable one (i.e., actual evapotranspiration) is estimated; variable one (i.e., actual evapotranspiration) obtained by estimation is compared with its corresponding initial estimate, and the initial estimate corresponding to variable one (actual evapotranspiration) is modified and recalculated until the initial estimate corresponding to variable one (actual evapotranspiration) is the same as its estimated value.

[0151] In this embodiment, based on the aforementioned successive approximation method, the solution process of the Jensen model is divided into two parts, specifically: decomposing the two-dimensional dynamic programming into two one-dimensional dynamic programming.

[0152] S5. For the farmland irrigation area, the crop growth data of the future time period is obtained through the BP-Markov crop model, and based on the crop growth data of the future time period, the farmland irrigation plan is obtained through the farmland irrigation model of the irrigation area.

[0153] Based on the aforementioned step S5, for obtaining crop growth data for a future time period through the BP-Markov crop model, it should be noted that:

[0154] The historical multi-source data screened by grey correlation analysis are used as input variables of the BP-Markov crop model to obtain crop growth data for future time periods.

[0155] Based on the aforementioned step S5, for obtaining the farmland irrigation plan through the irrigation model of the irrigation area farmland, it should be noted that:

[0156] The obtained crop growth data in the future time period is used as the input variable of the irrigation area farmland management model to obtain the irrigation water use plan for the irrigation area farmland in the future time period. Specifically, the obtained crop growth data in the future time period is used as the parameter Y in formula (6) of the Jensen model. a , and then solve the Jensen model through the above-mentioned successive approximation method to obtain the water allocation optimization plan.

[0157] The second embodiment discloses that, based on the constructed BP-Markov crop model and farmland irrigation model, a suitable irrigation plan for the farmland in the irrigation area is proposed, thereby achieving maximum water resource conservation while meeting the crop irrigation needs.

[0158] It should be noted that in the claims, any figure marks between brackets should not be understood as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claims. The word "one" or "an" preceding a component does not exclude the presence of multiple such components. In addition, it should be noted that in the description of this specification, the description of the terms "one embodiment", "some embodiments", "embodiment", "example", "specific example" or "some examples" and the like refers to the specific features, structures, materials or characteristics described in conjunction with the embodiment or example included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0159] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments after knowing the basic creative concept. Therefore, the claims should be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present invention.

[0160] Obviously, those skilled in the art can make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention should also include these modifications and variations.

Claims

1. A method for irrigation of farmland in an irrigated area based on an optimization algorithm, characterized in that: The farmland irrigation method comprises: S1. Obtain relevant historical data of the farmland irrigation area and pre-process the relevant historical data; The relevant historical data include: historical multi-source data, historical crop yield data and historical farmland irrigation data; S2. Obtaining the correlation value between the historical multi-source data and the historical crop yield data through grey correlation analysis, and constructing a BP optimization crop model through a BP neural network based on the correlation value; S3. Based on historical multi-source data and historical crop yield data, obtain crop yield prediction error by optimizing the crop model through BP; based on the crop yield prediction error, construct a BP-Markov crop model through Markov method; S4, constructing an irrigation model for the irrigation area based on the historical crop yield data and the historical farmland irrigation data; S5. For the farmland irrigation area, the crop growth data of the future time period is obtained through the BP-Markov crop model, and based on the crop growth data of the future time period, the farmland irrigation plan is obtained through the farmland irrigation model of the irrigation area.

2. The method for irrigating farmland in an irrigated area according to claim 1, characterized in that: The correlation values ​​between the historical multi-source data and the historical crop yield data obtained by grey correlation analysis in S2 include: The calculation process of the correlation value between the historical multi-source data and the historical crop yield data is as follows: Δ i (k)=|x0(k)-x i (k)| Among them, Δ i (k) is the absolute difference corresponding to the i-th type of multivariate data in the k-th year, x0(k) is the crop yield in the k-th year, x i (k) is the i-th type of multi-source data in the k-th year, ε i (k) is the correlation coefficient, ρ is the resolution coefficient, Δ min is the minimum absolute difference, Δ max is the maximum absolute difference, r i is the correlation degree, and N is the number of multi-source data types.

3. The method for irrigation of farmland in an irrigated area according to claim 2, characterized in that: The step S2 of constructing a BP optimized crop model based on the correlation value through a BP neural network includes: Based on the historical multi-source data obtained, according to the calculated correlation r i Determine the factors affecting crop yield, and construct a BP optimization crop model through a BP neural network based on the factors; The construction method of the BP optimization crop model is: S2-1. Determine the number of input layers, output layers and hidden layers of the BP neural network; S2-2, optimize the connection weights and thresholds in the BP neural network based on genetic algorithm; The connection weights include: connection weights from the input layer to the hidden layer, and connection weights from the hidden layer to the output layer.

4. The method for irrigating farmland in an irrigated area according to claim 3, characterized in that: The method of S3, based on historical multi-source data and historical crop yield data, obtains crop yield prediction error by optimizing crop model through BP, including: S3-1, using historical multi-source data as input variables of the BP optimization crop model to obtain predicted crop yields; S3-2, calculating the crop yield prediction error based on the predicted crop yield and the actual crop yield; The calculation method of the crop yield prediction error ω(k) is: ω(k)=|x0(k)-x0'(k)| Among them, x0(k) is the crop yield in the kth year, x0'(k) is the predicted crop yield in the kth year, and ω(k) is the prediction error of the crop yield in the kth year.

5. The method for irrigation of farmland in an irrigated area according to claim 4, characterized in that: The step S3 constructs a BP-Markov crop model based on the crop yield prediction error by using the Markov method, including: A crop yield prediction error state transfer matrix is ​​obtained based on the crop yield prediction error, and based on the crop yield prediction error state transfer matrix, the predicted crop yield obtained by using the BP optimization crop model is corrected.

6. The method for irrigation of farmland in an irrigated area according to claim 5, characterized in that: Based on the crop yield prediction error, the crop yield prediction error state transfer matrix is ​​obtained, including: Divide the obtained crop yield prediction error into state intervals, and count the states and state transitions corresponding to the predicted crop yield each year, and obtain the crop yield prediction error state transition matrix based on the state transitions; The crop yield prediction error state transfer matrix P is expressed as: Among them, r is the total number of states corresponding to the crop yield prediction error, P ab is the probability of transitioning from state a to state b.

7. The method for irrigating farmland in an irrigated area according to claim 1, characterized in that: The step S4 of constructing the irrigation model of the irrigation area based on the historical crop yield data and the historical farmland irrigation data includes: Among them, Y a is the actual crop yield, Y m is the maximum yield of crops, ET a The actual evaporation of crops, ET m is the potential evapotranspiration of crops, λ i is the crop sensitivity index at stage i, and i is the crop growth stage number.

8. The method for irrigating farmland in an irrigated area according to claim 1, characterized in that: The step S5 of obtaining crop growth data of a future time period for the farmland irrigation area through the BP-Markov crop model, and obtaining a farmland irrigation plan through the farmland irrigation model of the irrigation area based on the crop growth data of the future time period, includes: The historical multi-source data screened by grey correlation analysis are used as input variables of the BP-Markov crop model to obtain crop growth data for future time periods; The obtained crop growth data for the future time period is used as the corresponding actual crop yield in the irrigation model of the irrigation area farmland, and the irrigation water use plan for the irrigation area farmland in the future time period is obtained.

9. A computing device having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

10. A computer storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images