Intelligent irrigation management and control system

Abstract

The invention relates to the technical field of agricultural management, in particular to an intelligent irrigation management and control system which comprises the following steps: acquiring crop sample data, preprocessing the crop sample data, and extracting crop growth trend data from the processed crop sample data; constructing a self-attention mechanism model comprising three layers of Transform architectures, and training the self-attention mechanism model of the Transform architectures by using the crop growth trend data to obtain a basic model for detecting the crop growth condition; the method comprises the following steps: collecting sample data of a target area and an irrigation pipe network distribution map of the target area, segmenting the target area by using an irrigation pipe network, obtaining sample data of small areas, and preprocessing the sample data of the small areas. The plant growth condition of each region is accurately predicted, irrigation and fertilization are conveniently and specifically performed on crops, and the farmland fertilization accuracy in the region is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of agricultural management, in particular to an intelligent irrigation management system. BACKGROUND

[0002] The Internet of Things is a new network communication technology emerging in recent years. The Internet of Things uses local network or Internet communication technology to connect sensors, controllers, machines, personnel and objects together in a new way, forming a network of people, things and things, realizing informatization, remote management and control and intelligentization. With the development of smart agriculture, Internet of Things sensing technology, data transmission technology and intelligent processing technology have entered various fields of agricultural production. Through the intelligentization of agricultural production, the difference of the management field and the all-roundness of the service field, the agricultural industry chain is upgraded, the fine, efficient and green agriculture is realized, the safety of agricultural products is ensured, and the competitiveness and sustainable development ability of agriculture are improved. The existing irrigation system mainly adopts manual irrigation method, and the products mainly involve drip irrigation tape (irrigation nozzle) and pipeline system. With the increase of labor cost, it is urgent to promote the modernization process of agriculture. Nowadays, semi-automatic water and fertilizer integration systems have been developed and applied to irrigation projects. However, the existing system cannot realize fine water and fertilizer optimization combination, and cannot automatically adjust and implement irrigation according to the actual irrigation demand of farmland. It is difficult to truly meet the water demand of crop growth stage and achieve reasonable amount control of irrigation for crops. Therefore, we propose an intelligent irrigation management system. SUMMARY

[0003] In view of the defects in the prior art, the present application provides an intelligent irrigation management system, which comprises a sample acquisition module, a model training module, a small area prediction module and an irrigation module.

[0004] The sample acquisition module acquires crop sample data, pre-processes the crop sample data, and extracts crop growth trend data from the pre-processed crop sample data.

[0005] The model training module constructs a self-attention mechanism model comprising a three-layer Transformer architecture, trains the Transformer architecture self-attention mechanism model using the crop growth trend data, and obtains a basic model for detecting crop growth conditions.

[0006] The small area prediction module collects sample data of the target area and an irrigation pipe network distribution map of the target area, segments the target area with the irrigation pipe network, acquires small area sample data, pre-processes the small area sample data, obtains small area crop growth trend data, inputs the small area crop growth data into the basic model, and obtains a predicted crop growth result.

[0007] Irrigation module: according to the comparison between the predicted crop growth results and the normal crop growth conditions, it is judged whether irrigation water and fertilizer are needed, and the proportion of water and fertilizer is calculated for regional irrigation.

[0008] Further, the crop sample data includes R-stage crop growth trend data, the crop growth trend data includes a growth feature set and a crop growth performance change amount, the growth feature set includes K features, K is an integer greater than zero, and the crop growth change amount includes one of a morphological change amount, a physiological change amount, and a biochemical change amount.

[0009] Further, the crop growth trend data is extracted from the classified data, including:

[0010] A1: Obtain the V-stage crop growth trend data, i=1;

[0011] A2: Use the Pearson correlation coefficient to analyze and calculate the V-stage crop growth trend data to obtain the V-stage growth feature set, and let i=i+1, and return to step A1, the growth feature set includes m growth features;

[0012] A3: Repeat the above steps A1-A2 until i=R, and end the loop to obtain R growth feature sets;

[0013] A4: Calculate the intersection of the R growth feature sets to obtain S common growth features of all crop growth stages, and combine the S common growth features and the crop growth change amount into universal crop growth trend data, S is an integer greater than zero.

[0014] Further, the self-attention mechanism model of the three-layer Transformer architecture includes an embedding layer in the first layer, six Transformer encoder modules in the second layer, and an output layer in the third layer, wherein the Transformer encoder is composed of six repeated encoders, each encoder contains a self-attention sublayer and a feedforward neural network sublayer, and each sublayer is followed by normalization and residual connection.

[0015] Further, the self-attention mechanism model of the Transformer architecture is trained, including: obtaining universal crop growth trend data, dividing the universal crop growth trend data into a model training set and a model test set;

[0016] Label each crop growth change amount in the training set, wherein the label includes "0" and "1", wherein "0" represents normal growth and "1" represents abnormal growth;

[0017] The self-attention mechanism model of the three-layer Transformer architecture is initialized, common growth features in a model training set are taken as inputs of the self-attention mechanism model, and labeled labels in the training set are taken as outputs of the self-attention mechanism model, the self-attention mechanism model is trained, and a trained self-attention mechanism model is obtained.

[0018] Further, the target area is segmented by using the irrigation pipe network to obtain small area sample data, which comprises:

[0019] An irrigation network map and a farmland status map are obtained, all maps are converted into the same plane coordinate system, and the irrigation network map is labeled to indicate the main pipe, branch pipe, capillary, valve, dripper, water source and water storage tank. At the same time, the converted irrigation network map is digitally converted to draw the pipe network map into a point layer and a line layer.

[0020] The farmland status map is obtained, and the region crop polygon map is drawn to obtain the center point of the polygon map.

[0021] The center point of the polygon map is superimposed on the point layer and the line layer, the region crop is matched with the nearest dripper, and the pipeline path to which the region crop belongs is recorded.

[0022] Taking the valve as the center, the service range of each valve is obtained, then the farmland status map is segmented in combination with the terrain to obtain sample data in the segmented region, and the sample data in the segmented region is small area sample data.

[0023] Further, the method for judging whether irrigation water and fertilizer are needed according to the predicted crop growth result compared with the normal crop growth condition comprises: obtaining the result output by the model and the label, and identifying the crop growth condition according to the label; and calculating the proportion of water and fertilizer, and irrigating in regions, which comprises:

[0024] B1: obtaining the actual dry matter growth amount of the non-normal growth region crop and the dry matter growth amount in the normal state of the small region, and calculating the dry matter difference value;

[0025] B2: obtaining the dry matter difference value, and calculating the irrigation water amount and the fertilizer amount needed by using the dry matter difference value;

[0026] The calculation formula of the water and fertilizer amount needed in the small region is:

[0027] I = max {0, 10 · K · (▽t-▽a) - 0.7P - 5 · (θ-θ_W) · Zr},

[0028] N = max {0, 1000 · [0.55 · 25 + 0.40 · 8] · (▽t-▽a) - N_S · ρ_b · Z_r · 0.1} / η_N,

[0029] Where I is the amount of water to be replenished, K is the water consumption per kilogram of dry matter, ▽t is the normal increase in dry matter, ▽a is the actual increase in dry matter, P is the total future rainfall (taken as 70%), θ is the current root zone moisture content, θ_W is the permanent wilting point, Zr is the root zone depth, N_s is the available nitrogen in the soil, ρ_b is the soil bulk density, and η_N is the nitrogen fertilizer utilization rate.

[0030] The beneficial effects of this invention are reflected in:

[0031] 1. Crop growth trend data is extracted from preprocessed crop sample data. The self-attention mechanism model is trained using the crop growth trend data to obtain a trained crop growth prediction model. Then, target area sample data is obtained. The area sample data is segmented according to the irrigation network diagram to obtain small area sample data. The small area sample data is input into the trained crop growth prediction model to obtain the actual crop growth prediction results. The prediction results are compared with the normal crop growth status to calculate the amount of water and fertilizer that should be added to the crops in the small area. Then, irrigation is carried out for the crops in the small area. This invention can accurately predict the crop growth status in small areas through deep learning. It also segments the crops in the middle of the large area into small areas, reducing the amount of calculation of the model at one time, and can accurately predict the plant growth status of each area, which facilitates targeted irrigation and fertilization of crops and improves the accuracy of fertilization of farmland in the area. Attached Figure Description

[0032] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. The elements or parts in the drawings are not necessarily drawn to scale.

[0033] Figure 1 This is a structural diagram of the intelligent irrigation control system in Example 1. Detailed Implementation

[0034] The embodiments of the technical solution of the present invention will now be described in detail with reference to the accompanying drawings. These embodiments are merely illustrative of the technical solution of the present invention and are therefore intended to limit the scope of protection of the present invention.

[0035] It should be noted that, unless otherwise stated, the technical or scientific terms used in this application should have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0036] like Figure 1As shown, an intelligent irrigation management system includes a sample acquisition module, a model training module, a small area prediction module, an irrigation module,

[0037] The sample acquisition module acquires crop sample data, pre-processes the crop sample data, and extracts crop growth trend data from the pre-processed crop sample data.

[0038] The model training module constructs a self-attention mechanism model including a three-layer Transformer architecture, trains the Transformer architecture self-attention mechanism model using crop growth trend data, and obtains a basic model for detecting crop growth conditions.

[0039] The small area prediction module collects sample data of a target area and an irrigation pipe network distribution map of the target area, segments the target area using the irrigation pipe network, acquires small area sample data, pre-processes the small area sample data, obtains small area crop growth trend data, inputs the small area crop growth data into the basic model, and obtains a predicted crop growth result.

[0040] The irrigation module compares the predicted crop growth result with normal crop growth conditions, determines whether irrigation is needed, calculates the ratio of water and fertilizer, and irrigates in different areas.

[0041] In an embodiment, the crop sample data includes R-stage crop growth trend data, the crop growth trend data includes a growth feature set and a crop growth performance change amount, the growth feature set includes K features, K is a positive integer, and the crop growth change amount includes one of a morphological change amount, a physiological change amount, and a biochemical change amount.

[0042] It should be understood that the morphological change amount, the physiological change amount, and the biochemical change amount are specific manifestations of measuring the growth state of crops, wherein the morphological change amount is the total change amount of plant height change, crown expansion, and leaf area index dynamics during the growth process of crops, the physiological change is the change amount of stem flow rate, leaf water potential, chlorophyll relative content, and net photosynthetic rate of crops, and the biochemical change amount is the change of leaf nitrogen, phosphorus, and potassium concentration.

[0043] The k influencing features include, but are not limited to, light intensity, temperature, rainfall, soil moisture, stem flow rate, and leaf area index of the environment to which the plant belongs.

[0044] In an embodiment, the extraction of crop growth trend data from the classified data includes:

[0045] A1: Acquire the crop growth trend data of the Vth stage, i = 1.

[0046] A2: using the Pearson correlation coefficient to analyze and calculate the crop growth trend data of the Vth stage to obtain a growth characteristic set of the Vth stage, and letting i = i + 1, and returning to step A1, the growth characteristic set includes m growth characteristics;

[0047] A3: repeating the above steps A1-A2 until i = R, and ending the loop to obtain R growth characteristic sets;

[0048] A4: performing intersection calculation on the R growth characteristic sets to obtain S common growth characteristics of all crop growth stages, and combining the S common growth characteristics and the crop growth change amount into general crop growth trend data, S is an integer greater than zero;

[0049] Wherein, the formula of the Pearson correlation coefficient is as follows: In the formula, r is the Pearson correlation coefficient; x i is the value of the i th influencing characteristic of the Vth stage crop growth trend; n is the number of data points; p is the value of the growth performance change, and are the mean values of X and Y, respectively.

[0050] In an embodiment, the self-attention mechanism model of the three-layer Transformer architecture includes an embedding layer in the first layer, six Transformer encoder modules in the second layer, and an output layer in the third layer, wherein the Transformer encoder is composed of six repeated encoders, each encoder contains a self-attention sublayer and a feedforward neural network sublayer, and each sublayer is followed by normalization and residual connection.

[0051] In an embodiment, the self-attention mechanism model of the Transformer architecture is trained, including:

[0052] Obtaining general crop growth trend data, dividing the general crop growth trend data into a model training set and a model test set;

[0053] Labeling each crop growth change in the training set, wherein the labeling label includes "0" and "1", wherein "0" represents normal growth and "1" represents abnormal growth;

[0054] Initializing the self-attention mechanism model of the three-layer Transformer architecture, taking the common growth characteristics in the model training set as the input of the self-attention mechanism model, and taking the labeled labels in the training set as the output of the self-attention mechanism model, training the self-attention mechanism model to obtain a trained self-attention mechanism model.

[0055] In the embodiment, the irrigation pipe network is used to segment the target area, and sample data of small areas is obtained, including:

[0056] An irrigation network map and a farmland status map are obtained, and all maps are converted into the same plane coordinate system. The irrigation network map is labeled to indicate the main pipe, branch pipe, capillary, valve, dripper, water source, and water storage tank. The converted irrigation network map is digitally converted to draw the pipe network map into a point layer and a line layer.

[0057] A farmland status map is obtained, and a regional crop polygon map is drawn. The center point of the polygon map is obtained.

[0058] The center point of the polygon map is superimposed on the point layer and the line layer. The regional crop is matched with the nearest dripper, and the pipeline path to which the regional crop belongs is recorded.

[0059] The service range of each valve is obtained with the valve as the center. Then, the farmland status map is segmented in combination with the terrain to obtain sample data in the segmented area. The sample data in the segmented area is small area sample data.

[0060] In the embodiment, the method for determining whether irrigation water and fertilizer are needed according to the predicted crop growth result and the normal crop growth condition includes the following steps: obtaining the result output by the model and the label, and identifying the crop growth condition according to the label; and calculating the proportion of water and fertilizer and irrigating in regions, including:

[0061] B1: obtaining the actual dry matter growth amount of the crop in the non-normal growth small area and the dry matter growth amount in the small area under the normal state, and calculating the dry matter difference value;

[0062] B2: obtaining the dry matter difference value, and calculating the irrigation water amount and the fertilizer amount needed by using the dry matter difference value;

[0063] The calculation formula of the water and fertilizer amounts needed in the small area is as follows:

[0064] I = max{0, 10·K·(▽t-▽a)-0.7P-5·(θ-θ_W)·Zr},

[0065] N = max{0, 1000·[0.55·25+0.40·8]·(▽t-▽a)-N_S·ρ_b·Z_r·0.1} / η_N,

[0066] wherein I is the water amount needed, K is the water consumption per kilogram of dry matter,▽t is the normal dry matter increment,▽a is the actual dry matter increment, P is the total future rainfall (70% is taken), θ is the current root zone water content, θ_W is the permanent wilting point, Zr is the root zone depth, N_s is the soil available nitrogen, ρ_b is the soil bulk density, and η_N is the nitrogen fertilizer utilization rate.

[0067] The embodiment discloses a computer readable storage medium, including a memory, a processor, and a computer program stored in the memory and running on the processor, and the processor implements any one of the intelligent irrigation management systems provided by the above methods when executing the computer program.

[0068] The above formulas are all dimensionless numerical calculations, the formulas are obtained by collecting a large amount of data to simulate a formula of the most recent real situation, and the preset parameters, weights and threshold values in the formula are set by a person skilled in the art according to actual conditions.

[0069] The above embodiments can be realized wholly or partially by software, hardware, firmware or any combination thereof. When realized by software, the above embodiments can be realized wholly or partially in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the flow or function according to the embodiment of the present application is wholly or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable device. The computer instructions can be stored in a computer readable storage medium or transferred from one computer readable storage medium to another, for example, the computer instructions can be transferred from one website, computer, server or data center to another website, computer, server or data center through a wired network or wireless network. The computer readable storage medium can be any available medium accessible by a computer or a data storage device such as a server, data center and the like containing one or more available medium collections. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a DVD) or a semiconductor medium. The semiconductor medium can be a solid state disk.

[0070] Those skilled in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed in the present application can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0071] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described system, device and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0072] In several embodiments provided by the present application, it should be understood that the disclosed system, device and method can be implemented in other manners. For example, the division of the above-described device embodiments is only illustrative, and for example, the division of units can be other division manners, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections between different units, or the among them, can be indirect couplings or communication connections through some interfaces, devices or units, and can be in electrical, mechanical or other forms.

[0073] The units described as separated components can or can not be physically separated, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purposes of the embodiments.

[0074] In addition, each functional unit in the various embodiments of the present application can be integrated in one processing unit, or each unit can be a physically independent unit, or two or more units can be integrated in one unit.

[0075] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand: it can still modify the technical solutions recorded in the above embodiments, or make equivalent replacement for some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application, and they should be covered in the scope of the claims and the specification of the present application.

Claims

1. An intelligent irrigation management and control system, comprising a sample acquisition module, a model training module, a small-area prediction module, and an irrigation module, characterized in that: Sample acquisition module: Acquires crop sample data, preprocesses the crop sample data, and extracts crop growth trend data from the processed crop sample data; Model training module: Construct a self-attention mechanism model with a three-layer Transformer architecture, and train the self-attention mechanism model with the Transformer architecture using crop growth trend data to obtain a basic model for detecting crop growth status. Small area prediction module: Collects sample data and irrigation network distribution map of the target area, segments the target area by irrigation network, obtains small area sample data, preprocesses the small area sample data to obtain small area crop growth trend data, inputs the small area crop growth data into the basic model to obtain the predicted crop growth results; Irrigation module: Based on the predicted crop growth results compared with the normal crop growth status, it determines whether irrigation and fertilization are needed, calculates the water and fertilizer ratio, and irrigates according to different areas.

2. The intelligent irrigation control system according to claim 1, characterized in that: The crop sample data includes crop growth trend data for R stages. The crop growth trend data includes a crop growth feature set and crop growth performance changes. The crop growth feature set includes K features, where K is a positive integer. The crop growth performance changes include one of morphological changes, physiological changes, and biochemical changes.

3. The intelligent irrigation control system according to claim 2, characterized in that: The extraction of crop growth trend data from the classified data includes: A1: Obtain crop growth trend data for stage V, i = 1; A2: Analyze and calculate the crop growth trend data in stage V using the Pearson correlation coefficient to obtain the growth feature set in stage V, and let i = i + 1, and return to step A1. The growth feature set includes m growth features. A3: Repeat steps A1 to A2 above until i = R, then the loop ends, and R types of growth feature sets are obtained; A4: Perform intersection calculation on the R types of growth feature sets to obtain S common growth features for all crop growth stages, and combine the S common growth features and crop growth changes into general crop growth trend data, where S is an integer greater than zero.

4. The intelligent irrigation control system according to claim 1, characterized in that: The self-attention mechanism model of the three-layer Transformer architecture includes an embedding layer in the first layer, six Transformer encoder modules in the second layer, and an output layer in the third layer. The Transformer encoder consists of six repeating encoders, each of which contains a self-attention sublayer and a feedforward neural network sublayer. Each sublayer is followed by normalization and residual connections.

5. The intelligent irrigation control system according to claim 1, characterized in that: The training of the self-attention mechanism model of the Transformer architecture includes: Obtain general crop growth trend data and divide the general crop growth trend data into a model training set and a model test set; The growth change of each crop in the training set is labeled, wherein the label includes "0" and "1", where "0" represents normal growth and "1" represents abnormal growth; Initialize a three-layer Transformer architecture self-attention mechanism model, using the common growth features in the training set as the input of the self-attention mechanism model and the labeled tags in the training set as the output of the self-attention mechanism model, and train the self-attention mechanism model to obtain the trained self-attention mechanism model.

6. The intelligent irrigation control system according to claim 5, characterized in that: The process of segmenting the target area using the irrigation network to obtain small-area sample data includes: Obtain irrigation network diagrams and farmland status maps, convert all maps to the same plane coordinate system, and label the irrigation network diagrams, indicating main pipes, branch pipes, capillary pipes, valves, drippers, water sources, and reservoirs. At the same time, digitally convert the converted irrigation network diagrams into point layers and line layers. Obtain a map of the current state of farmland, draw a polygon map of crops in the area, and obtain the center point of the polygon map; Overlay the midline point of the polygon map with the point layer and line layer, match the crops in the area with the nearest dripper, and record the pipeline path to which it belongs; Centered on the valve, the service range of each valve is obtained. Then, the farmland status map is segmented in combination with the terrain to obtain sample data in the segmented areas. The sample data in the segmented areas are small area sample data.

7. The intelligent irrigation control system according to claim 6, characterized in that: The method for determining whether irrigation and fertilization are needed by comparing the predicted crop growth results with the normal crop growth status is as follows: obtain the model output results and labels, and identify the crop growth status based on the labels; The water-to-fertilizer ratio is calculated, and irrigation is carried out in different areas, including: B1: Obtain the actual dry matter growth of crops in a small area with abnormal growth and the dry matter growth in a small area under normal conditions, and calculate the difference in dry matter. B2: Obtain the dry matter difference and use it to calculate the required irrigation water and fertilizer amount; The formulas for calculating the amount of water and fertilizer to be replenished in a small area are as follows: I=max{0,10·K·(▽t-▽a)-0.7P-5·(θ-θ_W)·Zr}, N=max{0,1000·[0.55·25+0.40·8]·(▽t-▽a)-N_S·ρ_b·Z_r·0.1} / η_N, Where I is the amount of water to be replenished, K is the water consumption per kilogram of dry matter, ▽t is the normal increase in dry matter, ▽a is the actual increase in dry matter, P is the total future rainfall (taken as 70%), θ is the current root zone moisture content, θ_W is the permanent wilting point, Zr is the root zone depth, N_s is the available nitrogen in the soil, ρ_b is the soil bulk density, and η_N is the nitrogen fertilizer utilization rate.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the intelligent irrigation management system according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed, implements the intelligent irrigation control system according to any one of claims 1 to 7.

Images