An irrigation management system based on artificial intelligence

Abstract

The application provides an irrigation management and control system based on artificial intelligence, and relates to the technical field of drip irrigation management and control.The system comprises a leaf height information identification module for collecting first crop images and identifying multiple leaf height information; a growth vigor information identification module for collecting second crop images and identifying multiple growth vigor information; a root stem water constraint construction module for analyzing the water demand of root stems and the disease probability of leaves, and constructing root stem water constraints and leaf humidity constraints; and an optimal drip irrigation scheme acquisition module for acquiring an optimal drip irrigation scheme, conducting drip irrigation management and control, and optimizing the process by combining leaf height information to predict leaf humidity. The application can solve the technical problem of low drip irrigation accuracy caused by poor combination of drip irrigation control and actual crop conditions in the prior art, achieve the technical goal of improving plant irrigation accuracy and intelligence, and achieve the technical effect of improving plant crop growth quality.

Description

Technical Field

[0001] This application relates to the field of drip irrigation control technology, and in particular to an irrigation control system based on artificial intelligence. Background Technology

[0002] Irrigation methods for plants include furrow irrigation, sprinkler irrigation, and drip irrigation. Furrow irrigation is a commonly used method. Its advantages include low input, no wetting of plant stems and leaves, and prevention of stem and leaf diseases. However, its disadvantages include requiring more labor, significant water loss, susceptibility to waterlogging, and promotion of soil-borne diseases and plant rot. Sprinkler irrigation is suitable for slopes or sandy soils, effectively controlling irrigation volume and achieving high water use efficiency. However, its disadvantage is that water directly wets plant stems and leaves, potentially leading to fungal diseases. Drip irrigation can utilize limited water resources, reducing soil erosion and evaporation, but it is expensive, has high costs, and can easily promote soil salinization, making it only suitable for greenhouses or net houses.

[0003] Currently, drip irrigation is widely used due to its advantages. However, during the drip irrigation process, if the drippers are too close, the stems and leaves may become wet and diseased; if the drippers are too far apart, the water and nutrients may be insufficient. As a result, drip irrigation still has disease problems. In addition, irrigation management based on the experience of agricultural personnel can lead to inaccuracies in drip irrigation.

[0004] In summary, existing technologies suffer from the problem that manual drip irrigation control leads to a poor integration of drip irrigation control with actual crop conditions, resulting in low drip irrigation accuracy and further impacting crop growth. Summary of the Invention

[0005] The purpose of this application is to provide an artificial intelligence-based irrigation management and control system to solve the technical problem in the prior art where the integration of drip irrigation control with the actual crop conditions is poor due to manual drip irrigation control, resulting in low drip irrigation accuracy and further affecting crop growth.

[0006] In view of the above problems, this application provides an irrigation management and control system based on artificial intelligence.

[0007] Firstly, this application provides an artificial intelligence-based irrigation management system, implemented through an artificial intelligence-based irrigation management method. The system includes: a leaf height information recognition module, used to acquire first crop images of multiple parts of crops within multiple regions, and based on artificial intelligence, to identify multiple leaf height information of the multiple parts of crops based on the multiple first crop images, wherein the multiple regions correspond to the multiple drip irrigation modules; a growth information recognition module, used to acquire second crop images of the multiple parts of crops, and to identify multiple growth information of the multiple parts of crops; and a root and stem water constraint construction module, wherein the root and stem... The water constraint construction module is used to analyze the water requirements of the roots and stems of the multiple crop parts and the disease probability of the leaves based on the multiple growth information. It constructs root and stem water constraints based on the water requirements and leaf humidity constraints based on the disease probability. The optimal drip irrigation scheme acquisition module is used to optimize the drip irrigation scheme of the multiple regions based on the multiple leaf height information, multiple growth information, multiple root and stem water constraints, and multiple leaf humidity constraints to obtain the optimal drip irrigation scheme and perform drip irrigation control. Each optimized drip irrigation scheme includes the dripper, drip irrigation flow rate, drip irrigation frequency, and drip irrigation time in the selected drip irrigation module. During the optimization process, leaf humidity is predicted in combination with leaf height information.

[0008] Secondly, this application also provides an artificial intelligence-based irrigation management method for executing an artificial intelligence-based irrigation management system as described in the first aspect. The method includes: acquiring first crop images of multiple crops in multiple regions; identifying multiple leaf height information of the multiple crops based on artificial intelligence using the multiple first crop images, wherein the multiple regions correspond to the multiple drip irrigation modules; acquiring second crop images of the multiple crops and identifying multiple growth information of the multiple crops; analyzing the water requirements of the roots and stems and the disease probability of the leaves based on the multiple growth information; constructing root and stem water constraints based on the water requirements and constructing leaf humidity constraints based on the disease probability; optimizing the drip irrigation scheme for the multiple regions based on the multiple leaf height information, multiple growth information, multiple root and stem water constraints, and multiple leaf humidity constraints to obtain an optimal drip irrigation scheme for drip irrigation management. Each optimized drip irrigation scheme includes drippers, drip flow rate, drip frequency, and drip time within the selected drip irrigation module, and leaf humidity is predicted in conjunction with leaf height information during the optimization process.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] By acquiring first crop images of multiple crops in multiple regions, and using artificial intelligence to identify multiple leaf height information of the multiple crops based on these first crop images, wherein the multiple regions correspond to multiple drip irrigation modules, second crop images of the multiple crops are acquired, and multiple growth information of the multiple crops is identified, and based on the multiple growth information, the water requirements of the roots and stems and the disease probability of the leaves are analyzed, root and stem water constraints are constructed based on the water requirements, and leaf humidity constraints are constructed based on the disease probability, and the drip irrigation schemes of the multiple regions are optimized based on the multiple leaf height information, multiple growth information, multiple root and stem water constraints, and multiple leaf humidity constraints to obtain the optimal drip irrigation scheme for drip irrigation control, wherein each optimized drip irrigation scheme includes the dripper, drip flow rate, drip frequency, and drip time in the selected drip irrigation module, and leaf humidity is predicted in combination with leaf height information during the optimization process, ultimately achieving the technical goal of improving the accuracy and intelligence of plant irrigation, and achieving the technical effect of improving the growth quality of plants and crops.

[0011] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0013] Figure 1 This is a schematic diagram of the structure of an artificial intelligence-based irrigation control system according to this application;

[0014] Figure 2 This is a flowchart illustrating an artificial intelligence-based irrigation management method according to this application.

[0015] Explanation of reference numerals in the attached figures:

[0016] Leaf height information recognition module 11, growth status information recognition module 12, root and stem water constraint construction module 13, and optimal drip irrigation scheme acquisition module 14. Detailed Implementation

[0017] This application provides an artificial intelligence-based irrigation management system, which solves the technical problem in existing technologies where manual drip irrigation control leads to poor integration between the control method and the actual crop conditions, resulting in low drip irrigation accuracy and further affecting crop growth. It achieves the technical goal of improving the accuracy and intelligence of plant irrigation, thereby enhancing the quality of crop growth.

[0018] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0019] Example 1

[0020] Please see the appendix Figure 1 This application provides an artificial intelligence-based irrigation management and control system, wherein the method is applied to an artificial intelligence-based irrigation management and control method, and the system specifically includes the following steps:

[0021] The leaf height information recognition module 11 is used to collect first crop images of multiple parts of crops in multiple areas, and based on artificial intelligence, to identify multiple leaf height information of the multiple parts of crops, wherein the multiple areas correspond to the multiple drip irrigation modules;

[0022] The growth information recognition module 12 is used to collect a second crop image of the multi-part crop and identify multiple growth information of the multi-part crop.

[0023] The root and stem water constraint construction module 13 is used to analyze the water requirements of the roots and stems of the multiple crop parts and the disease probability of the leaves based on the multiple growth information, construct root and stem water constraints based on the water requirements, and construct leaf humidity constraints based on the disease probability.

[0024] The optimal drip irrigation scheme acquisition module 14 is used to optimize the drip irrigation scheme of the multiple regions based on the multiple leaf height information, multiple growth information, multiple root and stem water constraints and multiple leaf humidity constraints, to obtain the optimal drip irrigation scheme and perform drip irrigation control. Each optimized drip irrigation scheme includes the dripper, drip irrigation flow rate, drip irrigation frequency and drip irrigation time in the selected drip irrigation module. During the optimization process, the leaf humidity is predicted in combination with the leaf height information.

[0025] Furthermore, the blade height information identification module 11 in the system is also used for:

[0026] The first crop image acquisition module is used to acquire mapping images of multiple parts of crops in the multiple regions in a first direction, thereby obtaining multiple first crop images;

[0027] The sample leaf height information set acquisition module is used to acquire a sample first crop image set based on crop height recognition data records, and record the leaf height to obtain a sample leaf height information set.

[0028] The leaf height information acquisition module is used to construct a leaf height recognizer based on an intelligent convolutional neural network using the sample first crop image set and the sample leaf height information set to recognize the multiple first crop images and obtain the multiple leaf height information.

[0029] Furthermore, the growth information identification module 12 in the system is also used for:

[0030] The second crop image acquisition module is used to acquire multiple second crop images of the multi-part crop in a second direction;

[0031] The sample growth progress set acquisition module is used to obtain a sample second crop image set based on crop growth progress identification data records, and to mark the crop growth progress to obtain a sample growth progress set.

[0032] The crop growth progress information acquisition module is used to construct a growth progress recognizer based on an intelligent convolutional neural network using the sample second crop image set and the sample growth progress set to recognize the multiple second crop images and obtain multiple crop growth progress information.

[0033] The crop growth rate information calculation module is used to collect multiple historical crop growth progress information identified in the previous preset time period of the multiple crops, and calculate multiple crop growth rate information by combining the multiple crop growth progress information.

[0034] The growth information classification module is used to classify and obtain multiple growth information of the multiple crops based on the multiple crop growth progress information and multiple crop growth rate information.

[0035] Furthermore, the root and stem water constraint construction module 13 in the system is also used for:

[0036] The sample water requirement set acquisition module is used to acquire the sample growth information set, and to collect and mark the water requirement and leaf disease probability of crops with different sample growth information to acquire the sample water requirement set and the sample disease probability set.

[0037] The disease incidence probability information acquisition module is used to construct a water demand identification branch and a disease incidence probability identification branch by using the sample growth information set and combining the sample water demand set and the sample disease incidence probability set, respectively, to obtain a crop feature recognizer, identify the multiple growth information, and obtain multiple water demand information and multiple disease incidence probability information.

[0038] A root and stem water constraint construction module is used to construct the root and stem water constraint by determining that the water received by the multiple parts of the crop root and stem is greater than or equal to the multiple water demand information.

[0039] A leaf humidity threshold classification module is used to classify and obtain multiple leaf humidity thresholds of crops based on the multiple disease incidence probability information, wherein the classification index is performed by a leaf humidity classifier that uses sample disease incidence probability information and sample leaf humidity thresholds.

[0040] A leaf humidity constraint construction module is used to construct the leaf humidity constraint by setting the humidity of the multiple crop leaves to be less than or equal to the multiple leaf humidity thresholds.

[0041] Furthermore, the optimal drip irrigation scheme acquisition module 14 in the system is also used for:

[0042] A drip irrigation function construction module is used to construct a drip irrigation function to optimize the drip irrigation scheme based on the multiple water demand information and the multiple leaf humidity thresholds, as shown in the following formula:

[0043]

[0044] The drip irrigation function construction and analysis module is used in which di represents the drip irrigation fitness, w1, w2, and w3 are the water weight, leaf weight, and cost weight, respectively, and their sum is 1, M represents the number of multiple regions and multiple parts of crops, and ω i S represents the weights assigned to the i-th part of the crop based on multiple growth information. ai To supply water to the i-th part of the crop roots and stems according to the drip irrigation plan, S bi For the water requirement information of the i-th crop, H bi H represents the leaf humidity threshold for the i-th part of the crop. ai To ensure the humidity of the i-th section of crop leaves is irrigated according to the drip irrigation plan, K r To determine the cost of drip irrigation according to the drip irrigation plan, K m To maximize drip irrigation costs;

[0045] The drip irrigation scheme optimization module is used to optimize the drip irrigation scheme for the multiple regions according to the drip irrigation function, based on the multiple leaf height information, multiple growth information, multiple root and stem water constraints, and multiple leaf humidity constraints.

[0046] Furthermore, the optimal drip irrigation scheme acquisition module 14 in the system is also used for:

[0047] The first drip irrigation scheme obtaining module is used to randomly select multiple drippers based on the multiple drip irrigation modules, and randomly generate multiple drip irrigation flow rates, multiple drip irrigation frequencies and multiple drip irrigation times to obtain a first drip irrigation scheme.

[0048] The first drip irrigation fitness calculation module is used to analyze and obtain multiple water supply and multiple leaf humidity based on the distance between multiple drippers and multiple crops, multiple drip irrigation flow rates, multiple drip irrigation frequencies and multiple drip irrigation times in the first drip irrigation scheme, combined with the multiple leaf height information and the multiple growth information, and calculate the first drip irrigation fitness based on the drip irrigation function.

[0049] The second drip irrigation fitness calculation module is used to continue to randomly generate a second drip irrigation scheme and calculate the second drip irrigation fitness.

[0050] The drip irrigation optimization result acquisition module is used to calculate and obtain an updated probability distribution based on the magnitude of the second drip irrigation fitness and the first drip irrigation fitness, select the first drip irrigation scheme and the second drip irrigation scheme, and obtain the drip irrigation optimization result;

[0051] The drip irrigation optimization result output module is used to continue optimizing the drip irrigation scheme until convergence, and output the final drip irrigation optimization result to obtain the optimal drip irrigation scheme.

[0052] Furthermore, the optimal drip irrigation scheme acquisition module 14 in the system is also used for:

[0053] The drip irrigation record data acquisition module is used to acquire, based on the drip irrigation record data of the drip irrigation device, a set of sample dripper distances, a set of sample drip irrigation flow rates, a set of sample drip irrigation frequencies, a set of sample drip irrigation times, a set of sample leaf height information, a set of sample water supply, and a set of sample leaf humidity.

[0054] A drip irrigation crop analyzer construction module is used to construct a drip irrigation crop analyzer by taking the sample dripper distance set, sample drip flow rate set, sample drip frequency set, sample drip time set, and sample leaf height information set as inputs, and taking the sample water supply set and sample leaf humidity set as outputs.

[0055] The drip irrigation crop analysis module is used to perform drip irrigation crop analysis on the first drip irrigation scheme based on the drip irrigation crop analyzer, and obtain multiple supply water and multiple leaf humidity of the multiple crops.

[0056] Example 2

[0057] Based on the same inventive concept as the AI-based irrigation management system described in the foregoing embodiments, this application also provides an AI-based irrigation management method. Please refer to the appendix. Figure 2 The method includes:

[0058] Step 1: Collect first crop images of multiple parts of crops in multiple regions. Based on the multiple first crop images, identify multiple leaf height information of the multiple parts of crops using artificial intelligence. The multiple regions correspond to the multiple drip irrigation modules.

[0059] Specifically, multiple areas are designated for irrigation, each including a drip irrigation module, and the crops to be irrigated are plants with leaves. First crop images of multiple parts of the crops within these irrigated areas are collected. Based on these first crop images, matching historical crop data and historical height information are extracted. A neural network model is constructed using artificial intelligence. This model is trained using the historical crop data and historical height information, and the trained neural network model is used to identify the first crop images, obtaining multiple leaf height information for various parts of the crops.

[0060] Step 2: Acquire second crop images of the multi-part crops and identify multiple growth information of the multi-part crops;

[0061] Specifically, second crop images of multiple crops are collected, and historical crop data and historical growth progress information are extracted from the multiple second crop images. A neural network model for growth progress recognition is constructed based on artificial intelligence. The neural network model for growth progress recognition is trained using historical crop data and historical growth progress information. Based on the trained neural network model for growth progress recognition, the second crop images are recognized to obtain multiple growth information of multiple crops.

[0062] Step 3: Based on the multiple growth information, analyze the water requirements of the roots and stems of the multiple parts of the crop and the disease probability of the leaves. Construct root and stem water constraints based on the water requirements and construct leaf humidity constraints based on the disease probability.

[0063] Specifically, based on historical crop irrigation records, the water requirements of crop roots and stems and the probability of leaf disease are calculated when multiple parts of the crop have corresponding growth status, according to multiple growth status information. Based on the water requirements, the constraint threshold for irrigation until the water required by the roots and stems is obtained, so that the leaves can reach the corresponding growth status, thus constructing the root and stem water constraint. Based on the disease probability, the constraint threshold for irrigation until the humidity reaches the level required to cause disease is obtained, thus constructing the leaf humidity constraint, thereby ensuring that irrigation satisfies the root and stem water constraint and the leaf humidity constraint.

[0064] Step 4: Based on the multiple leaf height information, multiple growth information, multiple root and stem water constraints, and multiple leaf humidity constraints, optimize the drip irrigation schemes for the multiple regions to obtain the optimal drip irrigation scheme and perform drip irrigation control. Each optimized drip irrigation scheme includes the dripper, drip flow rate, drip frequency, and drip time in the selected drip irrigation module. During the optimization process, leaf humidity is predicted in conjunction with leaf height information.

[0065] Specifically, a fitness calculation method is constructed based on multiple leaf height information, multiple growth status information, multiple root and stem water constraints, and multiple leaf humidity constraints. This method is used to optimize drip irrigation schemes in multiple regions, calculate the drip irrigation scheme with the highest fitness, obtain the optimal drip irrigation scheme, and manage drip irrigation. Each optimized drip irrigation scheme includes the dripper, drip flow rate, drip frequency, and drip time in the selected drip irrigation module. During the optimization process, leaf humidity is predicted in conjunction with leaf height information.

[0066] The aforementioned AI-based irrigation management method is applied to an AI-based irrigation management system, which can achieve the technical goal of improving the accuracy and intelligence of irrigation for plants, thereby improving the quality of plant growth.

[0067] Furthermore, this application also includes the following steps:

[0068] In the first direction, map images of multiple crops within the multiple regions are acquired to obtain multiple first crop images;

[0069] Based on crop height recognition data records, a sample first crop image set is obtained, and the leaf height is recorded to obtain a sample leaf height information set;

[0070] Using the sample set of first crop images and the sample set of leaf height information, a leaf height recognizer is constructed based on an intelligent convolutional neural network to recognize the multiple first crop images and obtain the multiple leaf height information.

[0071] Specifically, the first direction is the direction in which images of the crop are to be acquired, thereby obtaining the crop's height information. In this first direction, image acquisition devices, such as drones or cameras, are used to acquire images of the crop, resulting in multiple initial crop images. These images include mappings of multiple parts of the crop within multiple regions to ensure the comprehensiveness and representativeness of the data for subsequent analysis.

[0072] Then, based on the crop images recorded in the historical crop height identification data, the corresponding crop images are obtained from the first collected crop images, forming a sample first crop image set. The leaf heights are recorded in the sample first crop image set to obtain the crop height information in the sample first crop image set, thus obtaining a sample leaf height information set.

[0073] Next, a leaf height recognizer is constructed using a set of sample first crop images and a set of sample leaf height information as input data, employing an intelligent convolutional neural network. The sample leaf input data is divided into sample leaf training data and sample leaf verification data, with the division ratio customized by those skilled in the art based on actual conditions. For example, a ratio of 7:3. The leaf height recognizer is trained using the sample leaf training data. When the output data of the leaf height recognizer stabilizes, it is verified using the sample leaf verification data to obtain the output accuracy of the leaf height recognizer. The output accuracy of the leaf height recognizer is compared with a threshold value. If the output accuracy of the leaf height recognizer meets the threshold value, the training of the leaf height recognizer is complete. The output accuracy threshold value is customized by those skilled in the art based on actual conditions. For example, a threshold value of 80% is used. Further, multiple first crop images are input into the leaf height recognizer for crop height information recognition and matching, obtaining multiple leaf height information corresponding to multiple first crop images.

[0074] By using artificial intelligence to identify leaf height information from multiple first crop images, the accuracy of obtaining leaf height information can be improved.

[0075] Furthermore, this application also includes the following steps:

[0076] In the second direction, multiple second crop images of the multi-part crop are acquired;

[0077] Based on the crop growth progress identification data record, a sample second crop image set is obtained, and the crop growth progress is marked to obtain a sample growth progress set;

[0078] Using the sample set of second crop images and the sample growth progress set, a growth progress recognizer is constructed based on an intelligent convolutional neural network to recognize the multiple second crop images and obtain multiple crop growth progress information;

[0079] Collect multiple historical crop growth progress information identified in the previous preset time period for the multiple crops, and combine the multiple crop growth progress information to calculate multiple crop growth rate information;

[0080] Based on the multiple crop growth progress information and multiple crop growth rate information, multiple growth status information of the multiple crops is obtained by classification.

[0081] Specifically, the second direction is the direction in which images of the crop are to be acquired, to obtain information on the crop's growth progress. Multiple second crop images are acquired using image acquisition equipment, focusing on multiple parts of the crop.

[0082] Then, based on the crop images identified in the historical crop growth progress identification data records, corresponding crop images are obtained from the collected second crop images, forming a sample second crop image set. Leaf growth progress is then marked on the sample second crop image set to obtain the crop growth progress information within the sample second crop image set, thus acquiring the sample growth progress set.

[0083] Next, a growth progress recognizer is constructed using a set of sample second crop images and a set of sample growth progress information as input data, employing an intelligent convolutional neural network. The sample growth progress input data is divided into sample growth progress training data and sample growth progress verification data, with the division ratio customized by those skilled in the art based on actual conditions. For example, a ratio of 7:3. The growth progress recognizer is trained using the sample growth progress training data. When the output data of the growth progress recognizer stabilizes, it is verified using the sample growth progress verification data to obtain the output accuracy of the growth progress recognizer. The output accuracy of the growth progress recognizer is compared with a threshold value. If the output accuracy of the growth progress recognizer meets the threshold value, the training of the growth progress recognizer is complete. The output accuracy threshold value is customized by those skilled in the art based on actual conditions. For example, an output accuracy threshold value of 80% is used. Further, multiple second crop images are input into the growth progress recognizer for crop growth progress information recognition and matching, obtaining multiple growth progress information corresponding to multiple second crop images.

[0084] Next, the previous preset time period is the time period in which historical crop growth progress information was previously identified, for example, the previous preset time period is the past day. Multiple historical crop growth progress information pieces were collected from various crops identified within the previous preset time period, representing the growth progress information of multiple crops at that time. This information was then combined with the current growth progress information of multiple crops. A difference calculation was performed between the historical crop growth progress information and the multiple crop growth progress information. The ratio of this difference result to the length of the previous preset time period was then calculated to obtain the growth rate information of multiple crops.

[0085] Next, based on multiple crop growth progress information and corresponding crop growth rate information, the crops are categorized according to their growth rate to obtain multiple growth status information for various crops. For example, multiple crops can be clustered based on a speed clustering threshold, which can be customized by those skilled in the art according to actual conditions, thereby obtaining classifications such as crops with faster growth and crops with slower growth.

[0086] By acquiring second crop images of multiple parts of the crop, multiple growth information of the multiple parts of the crop can be identified, thereby improving the accuracy of obtaining growth information.

[0087] Furthermore, this application also includes the following steps:

[0088] Obtain a set of sample growth information, and collect and label the water requirements and leaf disease probability of crops with different sample growth information to obtain a set of sample water requirements and a set of sample disease probabilities.

[0089] Using the sample growth information set, and combining it with the sample water requirement set and the sample disease probability set, a water requirement identification branch and a disease probability identification branch are constructed to obtain a crop feature identifier, which identifies the multiple growth information to obtain multiple water requirement information and multiple disease probability information.

[0090] The water received by the multiple parts of the crop roots and stems is greater than or equal to the multiple water demand information, which is used to construct the root and stem water constraint.

[0091] Based on the multiple disease incidence probability information, multiple leaf humidity thresholds of crops are obtained by classification, wherein the classification index is performed by a leaf humidity classifier that uses sample disease incidence probability information and sample leaf humidity thresholds.

[0092] The leaf humidity constraint is constructed by setting the humidity of the multiple crop leaves to be less than or equal to the multiple leaf humidity thresholds.

[0093] Specifically, multiple historical growth records are obtained to form a sample growth information set. Crops corresponding to different growth conditions within this set are then extracted, along with data on water requirements and leaf disease incidence during the crop's growth process. This yields a sample water requirement set and a sample disease probability set corresponding to the sample growth information set. For example, this can be obtained by calculating the amount of irrigation water received by the crop and the types and frequency of crop diseases.

[0094] Then, a crop feature recognizer is constructed using a neural network model, which includes a water requirement recognition branch and a disease probability recognition branch. Further, a set of sample growth information and a set of sample water requirements corresponding to multiple crops are used as the water requirement input data. This input data is divided into water requirement training data and water requirement validation data, with the division ratio customized by those skilled in the art based on actual conditions. For example, a division ratio of 7:3. The water requirement recognition branch is trained using the water requirement training data. When the output data of the water requirement recognition branch tends to stabilize, the branch is validated using the water requirement validation data to obtain its output accuracy. The output accuracy of the water requirement recognition branch is compared with a threshold value. If the output accuracy of the water requirement recognition branch meets the threshold, the training of the branch is complete. The output accuracy threshold value is customized by those skilled in the art based on actual conditions. For example, a threshold value of 80% is used. Furthermore, a set of sample growth information and a set of sample disease probabilities corresponding to multiple crops are used as the disease probability input data. This disease probability input data is divided into disease probability training data and disease probability verification data, with the division ratio customized by those skilled in the art based on actual conditions. For example, a division ratio of 7:3. The disease probability recognition branch is trained using the disease probability training data. When the output data of the disease probability recognition branch tends to stabilize, the disease probability recognition branch is verified using the disease probability verification data to obtain the output accuracy of the disease probability recognition branch. The output accuracy of the disease probability recognition branch is compared with a threshold value. If the output accuracy of the disease probability recognition branch meets the threshold value, the training of the disease probability recognition branch is complete. The output accuracy threshold value of the disease probability recognition branch is customized by those skilled in the art based on actual conditions. For example, the output accuracy threshold value is 80%. Further, multiple growth information sets are input into a crop feature recognizer to identify and match crop water requirements and disease probabilities, resulting in multiple water requirement information sets and multiple disease probability information sets.

[0095] Then, multiple water demand parameters are established, such as when the water received by the crop roots and stems is greater than or equal to the water demand, to construct root and stem water constraints. This ensures that the water received by the crop meets its water requirements. Further, multiple leaf humidity thresholds for the crop are obtained through classification. These thresholds are obtained by using a leaf humidity classifier that combines sample disease probability information with sample leaf humidity thresholds. These leaf humidity thresholds are used to ensure that the humidity of crop leaves remains within an appropriate range to reduce the risk of disease.

[0096] Next, when the humidity of multiple crop leaves is less than or equal to multiple leaf humidity thresholds, that is, when the humidity of multiple crop leaves does not reach the leaf humidity threshold, leaf humidity constraints are constructed, and leaf irrigation is stopped to avoid the irrigation humidity exceeding the leaf humidity threshold, which would cause leaf disease.

[0097] By analyzing the water requirements of multiple crop roots and stems and the disease probability of leaves, we can obtain root and stem water constraints and leaf humidity constraints, thereby further improving the accuracy of irrigation.

[0098] Furthermore, this application also includes the following steps:

[0099] Based on the multiple water demand information and the multiple leaf humidity thresholds, a drip irrigation function is constructed to optimize the drip irrigation scheme, as shown in the following formula:

[0100]

[0101] Where di represents drip irrigation fitness, w1, w2, and w3 are water weight, leaf weight, and cost weight, respectively, and their sum is 1, M represents the number of multiple regions and multiple parts of crops, and ω i S represents the weights assigned to the i-th part of the crop based on multiple growth information. ai To supply water to the i-th part of the crop roots and stems according to the drip irrigation plan, S bi For the water requirement information of the i-th crop, H bi H represents the leaf humidity threshold for the i-th part of the crop. ai To ensure the humidity of the i-th section of crop leaves is irrigated according to the drip irrigation plan, K r To determine the cost of drip irrigation according to the drip irrigation plan, K m To maximize drip irrigation costs;

[0102] According to the drip irrigation function, the drip irrigation scheme for the multiple regions is optimized based on the multiple leaf height information, multiple growth information, multiple root and stem water constraints, and multiple leaf humidity constraints.

[0103] Specifically, based on multiple water demand information and multiple leaf humidity thresholds, drip irrigation schemes that satisfy multiple water demand information and multiple leaf humidity thresholds are obtained. Among them, the drip irrigation scheme with the highest fitness is obtained by constructing a drip irrigation function to optimize the drip irrigation scheme. The drip irrigation function is as follows:

[0104]

[0105] Where di represents the drip irrigation fitness, the higher the fitness di, the better the drip irrigation scheme, and vice versa.

[0106] Furthermore, M represents the number of crops in multiple regions and parts, ω i S represents the weights assigned to the i-th part of the crop based on multiple growth information. ai To supply water to the i-th part of the crop roots and stems according to the drip irrigation plan, S bi This provides the water requirement information for the i-th part of the crop. Specifically, when the water supply S to the roots and stems of the i-th part of the crop... ai Information on the water requirements of the i-th crop, S bi The larger the difference, the more water is supplied, and the higher the drip irrigation adaptability (di); conversely, the smaller the difference, the lower the adaptability (di). Furthermore, H... bi H represents the leaf humidity threshold for the i-th part of the crop. ai To determine the humidity level at the i-th section of crop leaves according to the drip irrigation plan. Wherein, when the humidity threshold H of the i-th section of crop leaves... bi The humidity H at the i-th part of the crop leaves, as per the drip irrigation plan. ai The larger the difference, the lower the probability of leaf disease, and the higher the drip irrigation adaptability (di); conversely, the smaller the difference, the lower the adaptability. Furthermore, K... r To determine the cost of drip irrigation according to the drip irrigation plan, K m The maximum drip irrigation cost is denoted as K. This represents the cost K when drip irrigation is performed according to the specified plan. r The smaller and maximum drip irrigation cost K m A larger value indicates a lower cost for the drip irrigation scheme, resulting in a higher drip irrigation fitness rate (di), and vice versa. Further, w1, w2, and w3 represent the water weight, leaf weight, and cost weight, respectively, and the sum of the water weight w1, leaf weight w2, and cost weight w3 is 1. The drip irrigation fitness rate is obtained by multiplying the products of multiple weights with their corresponding weight values ​​and summing them.

[0107] Then, based on the drip irrigation function, and taking into account multiple leaf height information, multiple growth information, multiple root and stem water constraints, and multiple leaf humidity constraints, the drip irrigation scheme for each crop in multiple regions is optimized.

[0108] By calculating the drip irrigation fitness using drip irrigation functions, a drip irrigation scheme can be obtained, thereby improving the reliability of the drip irrigation scheme.

[0109] Furthermore, this application also includes the following steps:

[0110] Based on the multiple drip irrigation modules, multiple drippers are randomly selected, and multiple drip irrigation flow rates, multiple drip irrigation frequencies, and multiple drip irrigation times are randomly generated to obtain a first drip irrigation scheme;

[0111] Based on the distances between multiple drippers and multiple crops, multiple drip flow rates, multiple drip frequencies, and multiple drip times in the first drip irrigation scheme, combined with the multiple leaf height information and the multiple growth information, multiple water supply and multiple leaf humidity are analyzed and obtained. Combined with the drip irrigation function, the first drip irrigation fitness is calculated.

[0112] Continue to randomly generate a second drip irrigation scheme and calculate the fitness of the second drip irrigation scheme;

[0113] Based on the magnitudes of the second and first drip irrigation fitness, an updated probability distribution is calculated and used to select between the first and second drip irrigation schemes to obtain the drip irrigation optimization result.

[0114] Continue to optimize the drip irrigation scheme until convergence, output the final drip irrigation optimization result, and obtain the optimal drip irrigation scheme.

[0115] Specifically, multiple drippers are randomly selected to configure multiple crops in multiple drip irrigation modules. Within the drip irrigation flow rate that the multiple drippers can accommodate, multiple drip irrigation flow rates of the multiple drippers are randomly generated, as well as multiple drip irrigation frequencies and multiple drip irrigation time lengths, thereby obtaining a first drip irrigation scheme with drip irrigation flow rate, drip irrigation frequency and drip irrigation time.

[0116] Then, a first drip irrigation scheme is constructed based on the combinations of multiple drippers and their distances to multiple parts of the crop, multiple drip flow rates, multiple drip frequencies, and multiple drip times within the first drip irrigation scheme. Based on the first drip irrigation scheme, multiple leaf height and growth information are obtained, and then multiple water supply and leaf humidity input functions are obtained to calculate the first drip irrigation fitness.

[0117] Next, a second drip irrigation scheme is randomly generated. This second drip irrigation scheme has multiple random distances between drippers and multiple parts of the crop, multiple drip irrigation flow rates, multiple drip irrigation frequencies, and multiple drip irrigation time combinations. Based on the second drip irrigation scheme, multiple leaf height information and multiple growth status information are obtained, and then multiple water supply and multiple leaf humidity input drip irrigation functions are obtained to calculate the second drip irrigation fitness.

[0118] Next, the update probability distribution is calculated, that is, the first and second drip irrigation schemes are updated according to the update probabilities respectively, resulting in updated fitness values ​​for the first and second drip irrigation schemes. Specifically, the updates are performed based on random probabilities; for example, the probability of the first drip irrigation scheme accepting the update is 50%, and the probability of not accepting the update is 50%, and the update probability of the second drip irrigation scheme is randomly set accordingly. Further, the fitness values ​​of the updated first and second drip irrigation schemes are calculated, and the scheme with the higher fitness value is selected to obtain the optimized drip irrigation result.

[0119] Then, the drip irrigation scheme is optimized until convergence, that is, the change in drip irrigation fitness tends to stabilize. The final drip irrigation optimization result is then output to obtain the optimal drip irrigation scheme.

[0120] By optimizing drip irrigation schemes in multiple regions, the optimal drip irrigation scheme can be obtained, thereby further improving the accuracy of drip irrigation control for crops.

[0121] Furthermore, this application also includes the following steps:

[0122] Based on the drip irrigation record data of the drip irrigation device, obtain the sample dripper distance set, sample drip irrigation flow rate set, sample drip irrigation frequency set, sample drip irrigation time set, sample leaf height information set, sample water supply set, and sample leaf humidity set;

[0123] A drip irrigation crop analyzer is constructed by using the sample dripper distance set, sample drip irrigation flow rate set, sample drip irrigation frequency set, sample drip irrigation time set, and sample leaf height information set as inputs, and the sample water supply set and sample leaf humidity set as outputs.

[0124] Based on the drip irrigation crop analyzer, the first drip irrigation scheme is analyzed to obtain multiple water supply and multiple leaf humidity of the multiple crops.

[0125] Specifically, the process involves extracting historical drip irrigation records from the drip irrigation system, randomly extracting distance records between the drip irrigation system and the crop to obtain a sample dripper distance set, extracting drip flow rate records for the same crop to obtain a sample drip irrigation frequency set, extracting drip irrigation execution time records for the same crop to obtain a sample drip irrigation time set, extracting leaf height information records for the same crop to obtain a sample leaf height information set, extracting water supply records for the same crop to obtain a sample water supply set, and extracting leaf humidity records for the same crop to obtain a sample leaf humidity set.

[0126] Then, the drip irrigation analysis input data is obtained using a set of sample dripper distances, sample drip irrigation flow rates, sample drip irrigation frequencies, sample drip irrigation times, and sample leaf height information. The output data is obtained using a set of sample water supply and sample leaf humidity. A drip irrigation crop analyzer is constructed using an intelligent convolutional neural network. The drip irrigation analysis input and output data are divided into training data and validation data, with the division ratio customized by those skilled in the art based on actual conditions. For example, a ratio of 7:3 is used. The drip irrigation crop analyzer is trained using the training data. When the output data of the analyzer stabilizes, it is validated using the validation data to obtain the output accuracy. The output accuracy is then compared to a threshold. If the output accuracy meets the threshold, the training of the drip irrigation crop analyzer is complete. The output accuracy threshold is customized by those skilled in the art based on actual conditions. For example, the output accuracy threshold for the drip irrigation crop analyzer is 80%.

[0127] Next, the first drip irrigation scheme is input into the drip irrigation crop analyzer to perform drip irrigation crop analysis on the first drip irrigation scheme, and obtain multiple water supply and multiple leaf humidity of multiple crops.

[0128] By obtaining multiple supply water and multiple leaf humidity data for various parts of the crop through a drip irrigation crop analyzer, the accuracy of the output data can be improved.

[0129] In summary, the irrigation management method based on artificial intelligence provided in this application has the following technical effects:

[0130] By acquiring first crop images of multiple crops in multiple regions, and using artificial intelligence to identify multiple leaf height information of the multiple crops based on these first crop images, wherein the multiple regions correspond to multiple drip irrigation modules, second crop images of the multiple crops are acquired, and multiple growth information of the multiple crops is identified, and based on the multiple growth information, the water requirements of the roots and stems and the disease probability of the leaves are analyzed, root and stem water constraints are constructed based on the water requirements, and leaf humidity constraints are constructed based on the disease probability, and the drip irrigation schemes of the multiple regions are optimized based on the multiple leaf height information, multiple growth information, multiple root and stem water constraints, and multiple leaf humidity constraints to obtain the optimal drip irrigation scheme for drip irrigation control, wherein each optimized drip irrigation scheme includes the dripper, drip flow rate, drip frequency, and drip time in the selected drip irrigation module, and leaf humidity is predicted in combination with leaf height information during the optimization process, ultimately achieving the technical goal of improving the accuracy and intelligence of plant irrigation, and achieving the technical effect of improving the growth quality of plants and crops.

[0131] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The artificial intelligence-based irrigation management system and specific examples in the aforementioned embodiment one are also applicable to the artificial intelligence-based irrigation management method in this embodiment.

[0132] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0133] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. An irrigation management and control system based on artificial intelligence, characterized in that, The system is connected to a drip irrigation device, which includes multiple drip irrigation modules. Each drip irrigation module includes multiple drippers at different distances from the crop. The system includes: A leaf height information recognition module is used to collect first crop images of multiple parts of crops in multiple regions, and based on artificial intelligence, to identify multiple leaf height information of the multiple parts of crops, wherein the multiple regions correspond to the multiple drip irrigation modules; A growth information recognition module is used to acquire a second crop image of the multi-part crop and identify multiple growth information of the multi-part crop. A root and stem water constraint construction module is used to analyze the water requirements of multiple crop roots and stems and the disease probability of leaves based on the multiple growth information, construct root and stem water constraints based on the water requirements, and construct leaf humidity constraints based on the disease probability, including: The sample water requirement set acquisition module is used to acquire the sample growth information set, and to collect and mark the water requirement and leaf disease probability of crops with different sample growth information to acquire the sample water requirement set and the sample disease probability set. The disease incidence probability information acquisition module is used to construct a water demand identification branch and a disease incidence probability identification branch by using the sample growth information set and combining the sample water demand set and the sample disease incidence probability set, respectively, to obtain a crop feature recognizer, identify the multiple growth information, and obtain multiple water demand information and multiple disease incidence probability information. A root and stem water constraint construction module is used to construct the root and stem water constraint by determining that the water received by the multiple parts of the crop root and stem is greater than or equal to the multiple water demand information. A leaf humidity threshold classification module is used to classify and obtain multiple leaf humidity thresholds of crops based on the multiple disease incidence probability information, wherein the classification index is performed by a leaf humidity classifier that uses sample disease incidence probability information and sample leaf humidity thresholds. A leaf humidity constraint construction module is used to construct the leaf humidity constraint by ensuring that the humidity of the multiple crop leaves is less than or equal to the multiple leaf humidity thresholds. The optimal drip irrigation scheme acquisition module is used to optimize the drip irrigation scheme of the multiple regions based on the multiple leaf height information, multiple growth information, multiple root and stem water constraints, and multiple leaf humidity constraints to obtain the optimal drip irrigation scheme and perform drip irrigation control. Each optimized drip irrigation scheme includes the dripper, drip irrigation flow rate, drip irrigation frequency, and drip irrigation time in the selected drip irrigation module. During the optimization process, leaf humidity is predicted in combination with leaf height information. Based on the multiple leaf height information, multiple growth status information, multiple root and stem water constraints, and multiple leaf humidity constraints, the drip irrigation scheme for the multiple regions is optimized, including: A drip irrigation function construction module is used to construct a drip irrigation function to optimize the drip irrigation scheme based on the multiple water demand information and the multiple leaf humidity thresholds, as shown in the following formula: ； A drip irrigation function construction and analysis module is used within this module. For drip irrigation adaptability, , and These are weights for water, leaves, and cost, respectively, and their sum is 1. M represents the number of crops in multiple regions and multiple parts. The weights of the i-th part of the crop are assigned based on multiple growth information. To supply water to the i-th part of the crop roots and stems according to the drip irrigation plan, This provides water requirement information for the i-th crop. Let i be the leaf humidity threshold for the i-th part of the crop. To ensure the humidity of the i-th section of crop leaves is irrigated according to the drip irrigation plan, To determine the cost of drip irrigation according to the drip irrigation plan, To maximize drip irrigation costs; The drip irrigation scheme optimization module is used to optimize the drip irrigation scheme for the multiple regions according to the drip irrigation function, based on the multiple leaf height information, multiple growth information, multiple root and stem water constraints, and multiple leaf humidity constraints.

2. The system according to claim 1, characterized in that, First crop images of multiple parts of crops within multiple regions are collected. Based on these first crop images and using artificial intelligence, multiple leaf height information of the multiple parts of crops is identified, including: The first crop image acquisition module is used to acquire crop images of multiple parts of crops in the multiple regions in a first direction, thereby obtaining multiple first crop images; The sample leaf height information set acquisition module is used to acquire a sample first crop image set based on crop height recognition data records, and record the leaf height to obtain a sample leaf height information set. The leaf height information acquisition module is used to construct a leaf height recognizer based on an intelligent convolutional neural network using the sample first crop image set and the sample leaf height information set to recognize the multiple first crop images and obtain the multiple leaf height information.

3. The system according to claim 1, acquiring a second crop image of the multi-part crop and identifying multiple growth information of the multi-part crop, includes: The second crop image acquisition module is used to acquire multiple second crop images of the multi-part crop in a second direction; The sample growth progress set acquisition module is used to obtain a sample second crop image set based on crop growth progress identification data records, and to mark the crop growth progress to obtain a sample growth progress set. The crop growth progress information acquisition module is used to construct a growth progress recognizer based on an intelligent convolutional neural network using the sample second crop image set and the sample growth progress set to recognize the multiple second crop images and obtain multiple crop growth progress information. The crop growth rate information calculation module is used to collect multiple historical crop growth progress information identified in the previous preset time period of the multiple crops, and calculate multiple crop growth rate information by combining the multiple crop growth progress information. The growth information classification module is used to classify and obtain multiple growth information of the multiple crops based on the multiple crop growth progress information and multiple crop growth rate information.

4. The system according to claim 1, characterized in that, Based on the drip irrigation function, and considering multiple leaf height information, multiple growth status information, multiple root and stem water constraints, and multiple leaf humidity constraints, the drip irrigation scheme for the multiple regions is optimized, including: The first drip irrigation scheme obtaining module is used to randomly select multiple drippers based on the multiple drip irrigation modules, and randomly generate multiple drip irrigation flow rates, multiple drip irrigation frequencies and multiple drip irrigation times to obtain a first drip irrigation scheme. The first drip irrigation fitness calculation module is used to analyze and obtain multiple water supply and multiple leaf humidity based on the distance between multiple drippers and multiple crops, multiple drip irrigation flow rates, multiple drip irrigation frequencies and multiple drip irrigation times in the first drip irrigation scheme, combined with the multiple leaf height information and the multiple growth information, and calculate the first drip irrigation fitness based on the drip irrigation function. The second drip irrigation fitness calculation module is used to continue to randomly generate a second drip irrigation scheme and calculate the second drip irrigation fitness. The drip irrigation optimization result acquisition module is used to calculate and obtain an updated probability distribution based on the magnitude of the second drip irrigation fitness and the first drip irrigation fitness, select the first drip irrigation scheme and the second drip irrigation scheme, and obtain the drip irrigation optimization result; The drip irrigation optimization result output module is used to continue optimizing the drip irrigation scheme until convergence, and output the final drip irrigation optimization result to obtain the optimal drip irrigation scheme.

5. The system according to claim 4, characterized in that, Based on the distances between multiple drippers and multiple parts of the crop, multiple drip flow rates, multiple drip frequencies, and multiple drip times within the first drip irrigation scheme, combined with the multiple leaf height information, multiple water supply and multiple leaf humidity levels are analyzed and obtained, including: The drip irrigation record data acquisition module is used to acquire, based on the drip irrigation record data of the drip irrigation device, a set of sample dripper distances, a set of sample drip irrigation flow rates, a set of sample drip irrigation frequencies, a set of sample drip irrigation times, a set of sample leaf height information, a set of sample water supply, and a set of sample leaf humidity. A drip irrigation crop analyzer construction module is used to construct a drip irrigation crop analyzer by taking the sample dripper distance set, sample drip flow rate set, sample drip frequency set, sample drip time set, and sample leaf height information set as inputs, and taking the sample water supply set and sample leaf humidity set as outputs. The drip irrigation crop analysis module is used to perform drip irrigation crop analysis on the first drip irrigation scheme based on the drip irrigation crop analyzer, and obtain multiple supply water and multiple leaf humidity of the multiple crops.

6. An irrigation management and control method based on artificial intelligence, characterized in that, The method for implementing the artificial intelligence-based irrigation control system according to any one of claims 1 to 5 includes: First crop images of multiple parts of crops in multiple regions are collected. Based on the multiple first crop images and artificial intelligence, multiple leaf height information of the multiple parts of crops is identified. The multiple regions correspond to the multiple drip irrigation modules. Acquire second crop images of the multiple crop parts and identify multiple growth information of the multiple crop parts; Based on the multiple growth information, the water requirements of the roots and stems of the multiple parts of the crop and the disease probability of the leaves are analyzed. A water requirement for the roots and stems is constructed based on the water requirement, and a humidity requirement for the leaves is constructed based on the disease probability. Based on multiple leaf height information, multiple growth information, multiple root and stem water constraints, and multiple leaf humidity constraints, the drip irrigation schemes for the multiple regions are optimized to obtain the optimal drip irrigation scheme and perform drip irrigation control. Each optimized drip irrigation scheme includes the dripper, drip flow rate, drip frequency, and drip time in the selected drip irrigation module. During the optimization process, leaf humidity is predicted in conjunction with leaf height information.

Images