Method and device for irrigation regulation and control, equipment, greenhouse and medium

Abstract

The embodiment of the invention relates to a method and device for irrigation regulation and control, equipment, a greenhouse and a medium. The method is applied to soilless culture based on a greenhouse, and comprises the steps of obtaining production static data for soilless culture, and obtaining production dynamic data for soilless culture at a first preset time frequency. The method further comprises the step of utilizing a machine learning model based on a greenhouse soilless culture irrigation regulation and control algorithm to predict irrigation operation corresponding to an irrigation target for an irrigation machine of the greenhouse based on the obtained static production data and dynamic production data. The method further includes regulating greenhouse-based soilless culture to correspond to the irrigation target by performing the predicted irrigation operation for the irrigation machine. In this way, by combining static description and dynamic observation in greenhouse crop production, automatic irrigation regulation and control for greenhouse soilless culture can be efficiently and accurately carried out, the irrigation efficiency and precision are improved, meanwhile, the management cost is effectively reduced, and resources are saved.

Description

Technical Field

[0001] The present invention relates to the field of agricultural technology, and more particularly, to methods, devices, equipment, greenhouse sheds, and media for irrigation regulation. Background Art

[0002] A greenhouse shed (hereinafter also simply referred to as a shed) is a facility for growing plants such as vegetables and flowers, which can provide suitable environmental conditions to promote the growth and development of plants. Soilless cultivation based on a greenhouse shed is a method of cultivating plants using non-natural soil. It usually uses water, nutrient solution, or solid substrate as the medium for plant growth instead of soil. In a greenhouse shed, soilless cultivation can provide better environmental control and nutrient management, thus helping plants grow better. This cultivation method can reduce soil-borne diseases and pests and can more precisely control the nutrients and water required by plants. Summary of the Invention

[0003] Embodiments of the present disclosure provide methods, devices, equipment, greenhouse sheds, and media for irrigation regulation.

[0004] According to a first aspect of the present disclosure, there is provided a method for irrigation regulation, which is applied to soilless cultivation based on a greenhouse shed and includes obtaining production static data for soilless cultivation and obtaining production dynamic data for soilless cultivation at a first predetermined time frequency. The method further includes using a machine learning model based on an irrigation regulation algorithm for greenhouse shed soilless cultivation to predict an irrigation operation corresponding to an irrigation target for an irrigation machine of the greenhouse shed based on the obtained production static data and production dynamic data. The method further includes regulating soilless cultivation based on the greenhouse shed to correspond to the irrigation target by performing the predicted irrigation operation on the irrigation machine.

[0005] According to a second aspect of the present disclosure, there is provided a device for irrigation regulation, which is applied to soilless cultivation based on a greenhouse shed and includes a data acquisition module configured to obtain production static data for soilless cultivation and obtain production dynamic data for soilless cultivation at a first predetermined time frequency. The device further includes an operation prediction module configured to use a machine learning model based on an irrigation regulation algorithm for greenhouse shed soilless cultivation to predict an irrigation operation corresponding to an irrigation target for an irrigation machine of the greenhouse shed based on the obtained production static data and production dynamic data. The device further includes an operation execution module configured to regulate soilless cultivation based on the greenhouse shed to correspond to the irrigation target by performing the predicted irrigation operation on the irrigation machine.

[0006] According to a third aspect of the present disclosure, an electronic device is provided. The electronic device includes at least one processor. The electronic device further includes a memory that is coupled to the at least one processor and has instructions stored thereon that, when executed by the at least one processor, cause the device to perform the steps of the method in the first aspect of the present disclosure.

[0007] According to a fourth aspect of the present disclosure, a greenhouse is provided. The greenhouse includes the electronic device in the third aspect of the present disclosure.

[0008] According to a fifth aspect of the present disclosure, a computer-readable storage medium is provided. Computer-executable instructions are stored on the computer-readable storage medium, and when the computer-executable instructions are executed by a processor, the steps of the method in the first aspect of the present disclosure are implemented.

[0009] According to the solution of the embodiment of the present disclosure, it is possible to combine the static description and dynamic observation in the crop production in the greenhouse, and efficiently and accurately perform automatic irrigation regulation for soilless cultivation in the greenhouse, while improving the irrigation efficiency and accuracy, effectively reducing the management cost and saving resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] By describing the exemplary embodiments of the present disclosure in more detail in conjunction with the drawings, the above and other objects, features and advantages of the present disclosure will become more obvious. Among them, in the exemplary embodiments of the present disclosure, the same or similar reference numerals generally represent the same or similar components, assemblies, etc.

[0011] Figure 1A A schematic diagram illustrating an example environment in which the method and / or device according to the embodiment of the present disclosure may be implemented;

[0012] Figure 1B A schematic diagram illustrating an example of an irrigation machine of a greenhouse according to the embodiment of the present disclosure;

[0013] Figure 2 A flowchart illustrating a method for irrigation regulation applied to soilless cultivation based on a greenhouse according to the embodiment of the present disclosure;

[0014] Figure 3 A diagram illustrating an input-output stream of a machine learning model based on an irrigation regulation algorithm for soilless cultivation in a greenhouse according to the embodiment of the present disclosure;

[0015] Figure 4 A diagram illustrating the input formation process for a machine learning model based on an irrigation regulation algorithm for soilless cultivation in a greenhouse according to the embodiment of the present disclosure;

[0016] Figure 5Illustrated is a diagram showing the irrigation regulation process of soilless cultivation in a greenhouse according to an embodiment of the present disclosure;

[0017] Figure 6 Shown is a schematic diagram of a device for irrigation regulation according to an embodiment of the present disclosure; and

[0018] Figure 7 Illustrated is a schematic block diagram of an example device suitable for implementing embodiments of the present disclosure.

[0019] In each of the drawings, the same or corresponding reference numerals denote the same or corresponding parts. Detailed implementation manners

[0020] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not used to limit the protection scope of the present disclosure.

[0021] In the description of the embodiments of the present disclosure, the term "including" and its variants should be understood as open inclusion, that is, "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "an embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc. may refer to different or the same objects, unless clearly indicated otherwise.

[0022] As described above, a greenhouse is a facility for growing plants such as vegetables and flowers, which can provide suitable environmental conditions and promote the growth and development of plants. Soilless cultivation based on a greenhouse is an agricultural production method that uses greenhouse facilities and provides a suitable growth environment for crops through soilless cultivation technology, thereby achieving efficient, high-quality, and high-yield production.

[0023] The main purposes of irrigation regulation for soilless cultivation in a greenhouse mainly include the following aspects:

[0024] Saving water resources: In traditional soil cultivation methods, the soil has a certain water retention capacity, but there is also water loss and waste. In soilless cultivation, by using artificial media (such as rock wool, castor shells, etc.) as the substrate for plant growth, the supply and discharge of water can be better controlled, water waste can be reduced, and the water resource utilization efficiency can be improved.

[0025] Precise Irrigation Control: In soilless cultivation, the roots of plants are directly exposed to the medium, and the water supply can be controlled through a precise irrigation system. By adopting modern automated irrigation technologies such as drip irrigation and sprinkler irrigation, the irrigation water volume and frequency can be precisely regulated according to the needs and growth stages of plants, avoiding over-irrigation or under-irrigation and improving water use efficiency.

[0026] Improve Production Efficiency: The irrigation regulation in soilless cultivation can better meet the water requirements of crops, provide suitable water conditions, and is conducive to the growth and development of crops. Reasonable irrigation regulation can increase the yield and quality of crops and enhance agricultural production efficiency.

[0027] Irrigation regulation is a very important part of soilless cultivation based on greenhouse. In soilless cultivation based on greenhouse, the control of irrigation volume directly and / or indirectly affects the growth and yield of crops. If the irrigation volume is insufficient, the crops will be hindered in growth due to water shortage or lack of necessary nutrients; if the irrigation volume is excessive, it will lead to excessive soil humidity and cause problems such as root diseases. Therefore, precise control of irrigation volume is the key to the success of soilless cultivation. Through reasonable irrigation regulation, suitable water conditions can be provided for crops, promoting their normal growth, thereby increasing their yield and quality without heavy labor input and deployment costs. At the same time, reasonable irrigation regulation can also save water resources, reduce waste, and meet the requirements of sustainable development.

[0028] Although the irrigation regulation in soilless cultivation is of great significance in agricultural production and has been continuously developed in recent years, there are still some problems to be improved in these traditional irrigation regulation schemes, resulting in their inability to fully meet the needs of crop production and personnel operations, such as the actual irrigation volume deviating from the target range (e.g., insufficient or excessive), inaccurate irrigation time, etc. By way of example and not limitation, the irrigation for the soilless cultivation can be achieved by enabling and shutting down the irrigation machine. Currently, the operation of the irrigation machine for soilless cultivation irrigation regulation can be combined with artificial intelligence technology, but it requires a large amount of high-quality sample data and cannot guarantee the generalization ability. The common problem of these methods is that when applied to a new greenhouse setting or in the face of a changing production environment, a large amount of data needs to be re-collected for parameter tuning. In the case of slightly insufficient or poor-quality samples, it may lead to deterioration of the model generalization ability, reduction of accuracy, and a long deployment process.

[0029] To address at least the above and other potential issues, embodiments of the present disclosure provide a solution for irrigation regulation. The solution for irrigation regulation according to embodiments of the present disclosure is applied to soilless cultivation based on greenhouse greenhouses and includes obtaining production static data for soilless cultivation and obtaining production dynamic data for soilless cultivation at a first predetermined time frequency. The method further includes using a machine learning model based on an irrigation regulation algorithm for greenhouse soilless cultivation to predict, based on the obtained production static data and production dynamic data, an irrigation operation corresponding to an irrigation target for the irrigation machine of the greenhouse greenhouse. The method further includes regulating the greenhouse greenhouse to correspond to the irrigation target by performing the predicted irrigation operation on the irrigation machine. In this way, it is possible to combine the static description and dynamic observation in greenhouse crop production to efficiently and accurately perform automatic irrigation regulation for greenhouse soilless cultivation, while improving irrigation efficiency and accuracy, reducing management costs, and saving resources.

[0030] Figure 1A FIG. illustrates a schematic diagram of an example environment 100 in which a method and / or process according to an embodiment of the present disclosure may be implemented. As Figure 1A shown, the example environment 100 may include a greenhouse greenhouse 110, greenhouse data 120 from the greenhouse greenhouse 110, a computing device 130, and a storage device 140, and these components may be coupled to each other for interaction, as Figure 1A shown. It should be understood that only a limited number of components are shown in the example environment 100 for implementing embodiments of the present disclosure for ease of understanding and easy illustration, and embodiments of the present disclosure are not limited thereto. For example, the example environment 100 may further include a display (not shown) configured to display model output results.

[0031] According to an embodiment of the present disclosure, the greenhouse greenhouse 110 may be a facility for crop production that can provide suitable environmental conditions to promote the growth and development of plants. The greenhouse greenhouse 110 is configured with an irrigation machine, and for the purpose of irrigation regulation, the irrigation machine of the greenhouse greenhouse 110 may be instructed to perform appropriate irrigation operations, for example, at appropriate time points or at a certain time frequency, so that soilless cultivation based on the greenhouse greenhouse 100 can meet irrigation requirements. The greenhouse greenhouse 110 may include an irrigation machine for performing irrigation operations. According to an embodiment of the present disclosure, the irrigation machine may include one or more irrigation devices, such as drip irrigation devices, sprinkler irrigation devices, subsurface irrigation devices, etc.

[0032] It should be understood that Figure 1AThe shape, style, etc. of the greenhouse 110 shown are merely exemplary and not restrictive, and other different shapes, styles, etc. can also be adopted. In addition, it should be understood that the embodiments of the present disclosure do not limit or restrict the type and style, etc. of the irrigation machine, and other different types and styles, etc. of irrigation machines can also be included. Hereinafter, the sprinkler irrigation machine according to the embodiments of the present disclosure will be further described in detail.

[0033] According to an embodiment of the present disclosure, the greenhouse 110 can be configured with sensors (not shown), which are configured to perform data sensing on the greenhouse 110 and the facilities and equipment inside the greenhouse, etc., to obtain greenhouse data 120 regarding the greenhouse 110. The greenhouse data 120 can include production static data 121 and production dynamic data 122 for soilless cultivation in the greenhouse. In some embodiments, the production static data 121 for soilless cultivation in the greenhouse can include greenhouse data, which includes the greenhouse type, greenhouse area, lowest point height, highest point height, whether there is an earthen wall, air vent position, and air vent area, etc. of the greenhouse 110, to be used to describe the inherent attributes of the greenhouse 110. By way of example and not limitation, the greenhouse type can include warm greenhouses, warm and cold greenhouses, multi-span arch greenhouses, glass greenhouses, etc.

[0034] In some embodiments, the production static data 121 for soilless cultivation in the greenhouse can further include medium data, which can indicate the medium used for soilless cultivation, and the comparison of each medium with other media in terms of water permeability, etc. It includes medium type and medium description. For example, the medium type can include rock wool, castor husks, etc. that serve as the substrate for plant growth, and the medium description can include the text description of rock wool and castor husks. In addition, in some embodiments, the production static data 121 for soilless cultivation in the greenhouse can further include crop data, which includes crop type and crop introduction. For example, the medium type can indicate the types of plants, fungi, etc. to be cultivated soillessly in the greenhouse 110, and the crop description can include the text description of the plants, fungi, etc. to be cultivated soillessly in the greenhouse 110.

[0035] Additionally, according to embodiments of the present disclosure, the production dynamic data 122 for soilless cultivation in a greenhouse can include date and time, medium humidity, crop phenological period, current weather information, and weather information within a predetermined future time period. In some embodiments, the crop phenological period can include: germination, seedling stage, seedling period, extended growth period, tillering stage, branching stage, budding stage, flowering stage, post-flowering stage, fruit swelling stage, and maturity stage. The weather information within a predetermined future time period can be, for example, hourly weather forecasts for the next 24 hours. In some embodiments, the weather information within a predetermined future time period can include weather forecasts every half hour within the next 24 hours starting from the current moment, including weather type, outdoor temperature, outdoor humidity, etc. It should be understood that these production dynamic data 122 for soilless cultivation in a greenhouse may vary dynamically with factors such as time, season, and weather.

[0036] By way of example and not limitation, the sensors configured for the greenhouse 110 can include cameras, ultrasonic or laser rangefinders, infrared or radar sensors, light intensity sensors, temperature sensors, humidity sensors, angle sensors, etc. In addition, the sensors configured for the greenhouse 110 can also include soil moisture sensors and nutrient sensors for monitoring soil moisture and nutrient content, gas sensors for measuring carbon dioxide concentration and oxygen content in the air, sensors for evaluating the plant physiological state (such as leaf moisture, chlorophyll content, and growth rate, etc.) of plants, and water quality sensors for detecting water quality indicators such as soil pH value and conductivity, and so on. It should be understood that only some examples of sensors are given here for ease of understanding. According to embodiments of the present disclosure, other suitable sensors can also be used for soilless cultivation irrigation regulation, and the present disclosure does not impose any limitations or constraints on this.

[0037] Figure 1A The computing device 130 shown in can have computing capabilities corresponding to running a machine learning model for soilless cultivation irrigation regulation algorithm according to embodiments of the present disclosure, which can be arranged locally, or distributed in the cloud, or a combination thereof. The computing device 130 can be configured to perform soilless cultivation irrigation regulation according to embodiments of the present disclosure, and during the execution of soilless cultivation irrigation regulation according to embodiments of the present disclosure, the computing device 130 can access the storage device 140 and utilize the data stored in the storage device 140 to perform corresponding calculations, such as predicting irrigation operations to be performed to meet irrigation requirements. It should be understood that the computing device 130 is Figure 1A schematically shown as one computing device for ease of illustration and understanding, but this is only for the purpose of illustration and easy understanding. In the example environment 100, more computing devices can be arranged according to actual needs. The corresponding operations on the computing device 130 will be described in further detail below.

[0038] By way of example and not limitation, the computing device 130 may include but is not limited to a personal computer, a laptop computer, a server computer, a mobile device (such as a smart phone, a tablet computer, etc.), a wearable electronic device, a multimedia player, a personal digital assistant (PDA), a smart home device, a consumer electronic product, or a distributed computing environment including any one or more of the above devices. In some embodiments, a part of the computing device 130 may be arranged locally, while another part is arranged in the cloud.

[0039] According to an embodiment of the present disclosure, the storage device 140 may be configured to store greenhouse data 120 from the greenhouse 110, models (e.g., a machine learning model based on the greenhouse soilless cultivation irrigation regulation algorithm according to an embodiment of the present disclosure) and their parameters to be used on the computing device 130, prediction results, etc. It should be understood that the storage device 140 is Figure 1A schematically shown as one storage device for the purpose of facilitating illustration and easy understanding, and in the example environment 100, more storage devices may be arranged according to actual needs.

[0040] By way of example and not limitation, the storage device 140 may include but is not limited to a local storage device, a remote storage device, and combinations thereof. In some embodiments, multiple storage devices in the storage device 140 may include but are not limited to a mechanical hard disk drive (HDD), a solid state drive (SSD), etc., and some of the multiple storage devices may be arranged locally, while others may be arranged remotely and are coupled together via, for example, a line or a network.

[0041] Figure 1B A schematic diagram illustrating an example of an irrigation machine (exemplarily, irrigation machines 111, 112, and 113) of the greenhouse 110 according to an embodiment of the present disclosure. As Figure 1B shown, the irrigation machines 111, 112, and 113 may include but are not limited to at least one of the following: a drip irrigation machine, a sprinkler irrigation machine, a subsurface irrigation machine, etc. Based on the irrigation operations for, for example, the irrigation machines 111, 112, and 113, an appropriate humidity environment can be provided for the crops to ensure their healthy growth, thereby significantly improving the yield and quality of the crops. At the same time, this method also significantly reduces the labor input and deployment cost, and is more efficient and convenient. It should be understood that each irrigation machine illustrated thereon is only for the purpose of facilitating understanding and easy illustration, and is not restrictive. The various irrigation machines of the greenhouse 110 may of course have more or fewer numbers and various different styles and shapes, etc., and may be arranged at appropriate positions in the greenhouse 110.

[0042] As described above in connection with Figure 1A andFigure 1B An example environment 100 is described in which the methods and / or processes according to embodiments of the present disclosure may be implemented. The following will be combined with Figure 2 to describe a flowchart of a method 200 for irrigation regulation applied to soilless cultivation based on a greenhouse 110 according to an embodiment of the present disclosure. Through the method 200, it is possible to efficiently and precisely perform automatic irrigation regulation for greenhouse soilless cultivation by combining static descriptions and dynamic observations in crop production in the greenhouse, improving irrigation efficiency while reducing management costs.

[0043] At block 210, production static data 121 for greenhouse soilless cultivation is obtained, and production dynamic data 122 for greenhouse soilless cultivation is obtained at a first predetermined time frequency. As described above, the production static data 121 for greenhouse soilless cultivation may describe the inherent properties of soilless cultivation based on the greenhouse 110, and the production dynamic data 122 for greenhouse soilless cultivation may describe the dynamic changes of soilless cultivation based on the greenhouse 110 at a certain moment or time interval. Therefore, according to an embodiment of the present disclosure, the production static data 121 may be obtained only once, but the production dynamic data 122 needs to be obtained at a certain predetermined frequency to capture the state updates of greenhouse soilless cultivation under changing conditions. In some embodiments, obtaining the production dynamic data 122 at the first predetermined time frequency may be, for example, obtaining the production dynamic data 122 once every 24 hours.

[0044] At block 220, a machine learning model based on a greenhouse soilless cultivation irrigation regulation algorithm is used to predict an irrigation operation corresponding to an irrigation target for the irrigation machines (e.g., irrigation machines 111, 112, and 113) of the greenhouse 110 based on the obtained production static data 121 and production dynamic data 122. According to an embodiment of the present disclosure, the greenhouse soilless cultivation irrigation regulation algorithm may logically represent the target of greenhouse soilless cultivation irrigation regulation. In other words, the greenhouse soilless cultivation irrigation regulation algorithm may be configured to: while minimizing the number of irrigation times of the irrigation machines, regulate the medium humidity obtained at a second predetermined time frequency to be greater than or equal to a predetermined humidity threshold, and regulate the ratio of the liquid return amount to the irrigation amount to correspond to a predetermined ratio threshold, where the second predetermined time frequency is greater than the first predetermined time frequency. In some embodiments, obtaining the medium humidity at the second predetermined time frequency may be, for example, obtaining the medium humidity once every 15 minutes. Thus, even after a certain period of time or other factors, the humidity environment of the greenhouse 110 may decrease and fall below or exceed the irrigation target. To ensure the stability of greenhouse soilless cultivation irrigation regulation, the above re-regulation operation may be performed as a supplementary regulation. In this way, it is possible to keenly detect the medium humidity in a small time interval and timely tune the deviated humidity.

[0045] According to an embodiment of the present disclosure, a reward function may also be included, that is, the reward function: (1

[0046] - Sum the absolute values of the difference between each recharge ratio and a given value (which needs to be normalized)) + (1

[0047] - The number of times the medium humidity is less than the threshold (which needs to be normalized)) * weight coefficient 1 + (1

[0048] - The number of irrigation times (which needs to be normalized)) * weight coefficient 2. It should be understood that the weight coefficient 1 and the weight coefficient 2 can be adaptively configured according to actual usage requirements and are not limited and constrained herein.

[0049] Herein, the described machine learning model can be a machine learning model that can implement the above logic of the machine learning model based on the greenhouse soilless cultivation irrigation control algorithm according to an embodiment of the present disclosure. It can be a neural network model, a context learning model, etc. Specific examples and detailed details will be given below to help better understand the embodiments of the present disclosure. In addition, the described irrigation operation can be an operation for greenhouse soilless cultivation irrigation control for an irrigation machine of the greenhouse 110 (for example, Figure 1A one or more of the irrigation machines 111, 112, and 113 shown in ), and its examples can include but are not limited to controlling one or more irrigation machines to perform drip irrigation, sprinkler irrigation, subsurface irrigation, etc. In some embodiments, the irrigation operation for regulating the environment can include opening the irrigation device to irrigate the same amount of water at a certain time point. It should be understood that the present disclosure does not limit the specific form of the irrigation operation. For example, controlling one or more irrigation machines to perform sprinkler irrigation for a certain period of time and closing the irrigation machine in response to the expiration of the period, etc.

[0050] At block 230, the soilless cultivation based on the greenhouse 110 is adjusted to correspond to the irrigation target by performing the predicted irrigation operation on the irrigation machine. Thus, based on the production static data 121 and the most recent production dynamic data 122 regarding the greenhouse 110, the accurate irrigation operation adapted to the current situation of the soilless cultivation based on the greenhouse 110 can be quickly predicted by using the logic of the greenhouse soilless cultivation irrigation control algorithm according to an embodiment of the present disclosure, so that the soilless cultivation based on the greenhouse 110 can achieve the irrigation target or meet the irrigation requirements after the predicted irrigation operation. The method 200 according to an embodiment of the present disclosure can consider both the production static data 121 and the production dynamic data 122 for greenhouse soilless cultivation, so that it can be adapted to various different types of greenhouses 110 arranged in various regions, even in various seasons or weather conditions, and can efficiently implement greenhouse soilless cultivation irrigation control, thereby contributing to the improvement and promotion of agricultural production.

[0051] Figure 3FIG. illustrates an input-output stream 300 of a machine learning model 310 based on a greenhouse soilless cultivation irrigation regulation algorithm according to an embodiment of the present disclosure. As Figure 3 shown, after obtaining the production static data 121 and the most recent production dynamic data 122 for the soilless cultivation, the production static data 121 and the most recent production dynamic data 122 can be fed as model inputs into a machine learning model 310 based on a greenhouse soilless cultivation irrigation regulation algorithm for performing a prediction of the irrigation operation of the irrigation machine for the greenhouse 110. According to an embodiment of the present disclosure, the production static data 121 and the production dynamic data 122 can be converted into model inputs in a language corresponding to the machine learning model 310.

[0052] Herein, the machine learning model 310 can be arranged at Figure 1A the computing device 130 shown, and can embody the logical purpose of the greenhouse soilless cultivation irrigation regulation algorithm according to an embodiment of the present disclosure. After corresponding calculations of the machine learning model 310 based on the greenhouse soilless cultivation irrigation regulation algorithm, a model output 320 of the machine learning model 310 can be obtained, and the model output 320 indicates the predicted irrigation operation. In some embodiments, hereinafter, this process will be described in more detail by means of more specific examples.

[0053] According to an embodiment of the present disclosure, the machine learning model 310 based on the greenhouse soilless cultivation irrigation regulation algorithm can include a large language model (LLM) of the greenhouse soilless cultivation irrigation regulation algorithm, which is just one example among many examples of the machine learning model 310. Of course, other different types of performance-improved models can also be adopted. Before use, the data can be stored in the form of structured data, such as files like CSV. The obtained production static data 121 and production dynamic data 122 can be converted from structured data into model inputs in natural language. In other words, the obtained production static data 121 and production dynamic data 122 can be stored in, for example, the storage device 140 in the form of structured data, and when needed to be used, they are converted into a language adapted to the model.

[0054] To further improve the accuracy of the model output 320 of the LLM based on the greenhouse soilless cultivation irrigation regulation algorithm, in addition to the above model inputs, additional model inputs can also be adopted. According to an embodiment of the present disclosure, the target description, guidance, reference examples, and status description can be converted from structured data into additional model inputs in natural language. Hereinafter, the model inputs and additional model inputs according to an embodiment of the present disclosure will be further described in detail.

[0055] Figure 4FIG. illustrates a diagram of an input formation process 400 for a machine learning model 310 for a greenhouse soilless cultivation irrigation regulation algorithm according to an embodiment of the present disclosure. For the sake of convenience of explanation only, the above-mentioned LLM based on the greenhouse soilless cultivation irrigation regulation algorithm will be described hereinafter. As Figure 4 As described in, the production static data 121 and production dynamic data 122 associated with the static description and dynamic observation in crop production in the greenhouse 110 can be represented as the current state 410 (also simply referred to as state 410). The current state description 420 (also simply referred to as state description 420) is a description based on the state 410, which indicates the static description and dynamic observation in crop production in the greenhouse 110 based on the acquired production static data 121 and production dynamic data 122, wherein the current state 410 is converted into a corresponding adapted language, such as natural language, by means of a state translator.

[0056] According to an embodiment of the present disclosure, the target description 430 can indicate the target of the greenhouse soilless cultivation irrigation regulation algorithm as described above, and the guidance 440 can include the correlation between irrigation or irrigation operations and the environment, as well as the calculation method of the return score, etc. In addition, the reference example 450 (also simply referred to as example 450) can include historical examples 460, other greenhouse examples 470 and representative examples 480, and each example in the example 450 can include an example identifier, a reference irrigation operation and a reference feedback, wherein the reference feedback can include a reported score to indicate the irrigation regulation effect of the corresponding reference irrigation operation. As Figure 4 As shown in, the corresponding information in the production static data 121 and production dynamic data 122 for greenhouse soilless cultivation is converted into an adapted language, such as natural language, in the example identifier by means of a basic data translator and a state translator. In addition, the corresponding information in the reference irrigation operation and reference feedback is converted into an adapted language corresponding to the LLM based on the greenhouse soilless cultivation irrigation regulation algorithm, such as natural language, by means of an irrigation requirement translator and a feedback translator. It should be understood that the basic data translator, state translator, irrigation requirement translator and feedback translator can be software-based components or systems and can run on a hardware device with computing capabilities (such as the computing device 130).

[0057] By way of example, in the environment, the goal is to control the medium humidity measured every half hour above X, control the backwash ratio after each irrigation at about 15%, and minimize the number of irrigations on the premise that the amount of each irrigation is fixed. The greenhouse used is a warm greenhouse with an area of 100 square meters. The highest point of the warm greenhouse is 2.5 meters, and the lowest point is 1 meter. There is an earthen wall and cotton quilts.

[0058] By way of example, the basic information translator (i.e., basic data translator) translates the basic information of the greenhouse into natural language: the greenhouse is a warm greenhouse with a total area of 100 square meters, a highest point of 2.5 meters, and a lowest point of 1 meter. The greenhouse has earth walls and quilts. The greenhouse irrigation machine is on the roof, with a total area of 4 square meters and a maximum opening of 90 degrees. The medium is coconut coir, and the water permeability is medium to low. In addition, it translates the correlation between irrigation and the environment into natural language: the humidity of the medium is correlated with the frequency of irrigation. Under the same environment, the higher the irrigation frequency, the higher the humidity of the medium. The recharge ratio is correlated with the frequency of irrigation. Under the same environment, the higher the irrigation frequency, the higher the recharge ratio.

[0059] By way of example, the state translator translates the state of the greenhouse into natural language: it is 09:30 now, the medium humidity is X, and the weather forecast every half hour in the next 24 hours is: sunny, 24 degrees Celsius, 30%; sunny, 26 degrees Celsius, 30%; ...; sunny, 20 degrees Celsius, 30%.

[0060] Also by way of example, the irrigation requirement translator translates the time point of the irrigation requirement into natural language: Please irrigate at the following times: [10:00, 12:30, 13:30, ..., 8:00]. In addition, the feedback translator translates the reward score, action, and result into natural language: reward score: 8; irrigate at the following times: [10:00, 12:30, 13:30, ..., 8:00]; result: The corresponding re-irrigation ratio for each irrigation is [15%, 10%, 12%, ..., 20%]. The medium humidity every half hour is [x1, x2, ..., x48].

[0061] According to an embodiment of the present disclosure, the historical examples 460, other greenhouse samples 470, and representative samples 480 may come from a sample data set, which may be stored in, for example, a sample database deployed in the storage device 140. The historical examples 460 are examples of soilless cultivation based on the greenhouse 110 that are instructive for the current action judgment. In some embodiments, all samples of soilless cultivation associated with the greenhouse 110 may be retrieved from the sample database as the historical examples 460.

[0062] Other greenhouse samples 470 are samples of other greenhouses that are of reference value for greenhouse soilless cultivation irrigation control. According to an embodiment of the present disclosure, other greenhouse samples 470 can be obtained by the following actions: based on the other greenhouse sample data subset (i.e. Figure 4 The sample identifier of each sample in the "other shed sample data set" shown in ), clustering the other shed sample data subsets, such sample identifiers include embedded vectors encoded to represent the corresponding samples, and the encoding process of encoding the embedded vectors can be assisted byFigure 4 The "clustering model" shown in Figure 4 . In some embodiments, other subsets of greenhouse sample data are established based on the historical data of soilless cultivation irrigation regulation in greenhouses in other regions, and the clustering model can adopt, but is not limited to, the Kmeans algorithm. Next, based on the model input and additional model input for the LLM based on the soilless cultivation irrigation regulation algorithm in greenhouses, samples within a predetermined distance threshold from the center of the sample cluster for other subsets of greenhouse sample data can be identified as other greenhouse samples 470. The process of identifying samples within a predetermined distance threshold from the center of the sample cluster can be facilitated by Figure 4 The "sampling model" shown in Figure 4 , such as the K-Nearest Neighbor (KNN) model.

[0063] Representative samples 480 are samples that are representative in the field of soilless cultivation irrigation regulation in other greenhouses, such as some important samples with commonalities. According to embodiments of the present disclosure, representative samples 480 can be obtained through the following actions: clustering the other greenhouse sample data subsets (i.e., Figure 4 The "other greenhouse sample dataset" shown in Figure 4 ) in the sample dataset based on the sample identification of each sample, such sample identification including the embedding vector encoded to represent the corresponding sample. The encoding process of encoding into the embedding vector can be facilitated by Figure 4 The "clustering model" shown in Figure 4 . In some embodiments, the clustering model can adopt, but is not limited to, the Kmeans algorithm. Next, the sample closest to the center of the sample cluster for the other greenhouse sample data subset can be identified as representative sample 480. The process of identifying the sample closest to the center of the sample cluster can be facilitated by Figure 4 The "sampling model" shown in Figure 4 , such as the K-Nearest Neighbor (KNN) model.

[0064] After obtaining the historical samples 460, other greenhouse samples 470, and representative samples 480, since these samples may be frequently used, they can be cached for quick retrieval. In some embodiments, the historical samples 460, other greenhouse samples 470, and representative samples 480 can be cached, for example, cached to Figure 4 The "cache (which can be a specific part of the storage device 140)" shown in Figure 4 for retrieval based on the model input and additional model input, and the other samples in the sample dataset are stored in the sample database. Here, the described caching and storage can be performed in the form of a queue, for example, storing the content of recent model and environment interactions, including the environmental status (soil humidity, actual weather) collected every half hour, the irrigation operation requirements given by the large model every half hour, whether each irrigation is executed as planned, and the backflow ratio after each irrigation.

[0065] By way of example, each piece of data in the cache and other greenhouse sample datasets may include: the old state (medium humidity, weather forecast for the next 24 hours), actions (multiple irrigation time points in the next 24 hours), results (the recharge ratio after each irrigation, medium humidity every half hour), and the reward score.

[0066] According to an embodiment of the present disclosure, prompt words for the LLM based on the greenhouse soilless cultivation irrigation regulation algorithm can also be generated, where the prompt words can be extracted, for example, by the prompt word generator 490 from the following: the production static data 121 and production dynamic data 122 in the model input; the target description 430 and state description 420 in the additional model input; the cached historical samples 460, other greenhouse samples 470, and representative samples 480; and other samples in the sample database. In some embodiments, the prompt word generator 490 can integrate the generated prompt words into a text and input it as the model input into the large model.

[0067] It should be understood that the prompt word generator 490 can be a software-based component or system and can run on a device with computing capabilities (such as the computing device 130). In this way, it can significantly reduce the size of the input for the model, and then more valuable information may be added. For example, when the input is text, using the form of prompt words can reduce the length of the input text, and then other meaningful input features can be included, promoting the improvement of the model prediction accuracy. In addition, using the form of prompt words can highlight the key points to help the model understand the semantics.

[0068] Figure 5 A diagram illustrating the irrigation regulation process 500 of greenhouse soilless cultivation according to an embodiment of the present disclosure is shown. For the sake of convenience of explanation only, the above-mentioned LLM based on the greenhouse soilless cultivation irrigation regulation algorithm will be described hereinafter. According to an embodiment of the present disclosure, at 510, based on the current state of the environment, the state translator can translate the state into natural language. At 520, based on the state described in natural language, the embedding model can convert the natural language description into a vector containing semantics. It should be understood that the embodiments of the present disclosure do not limit and restrict the embedding model used, and an appropriate model can be selected according to the usage requirements.

[0069] According to an embodiment of the present disclosure, at 530, based on the embedding vector of the state, the sample sampling model can determine multiple other greenhouse state embedding vectors similar to the given vector, for example, from the other greenhouse sample dataset at 540, which stores the historical data of other greenhouses, including but not limited to greenhouse basic information, old state, actions, new state, reward scores, etc. In some embodiments, the other greenhouse sample dataset can be established based on the historical data of greenhouse soilless cultivation irrigation regulation in other regions.

[0070] According to an embodiment of the present disclosure, at 550, a clustering model can cluster a target data set (e.g., other shed sample data sets at 540), which clusters the data in the target data set into a specified number of clusters through similar embedding vectors. At 560, based on the embedding vectors of other shed states from 530, the representative embedding vectors from 550, and the historical samples of the current shed from 590, a prompt generator can utilize contents such as inputs and target descriptions to generate the input content of the large model regarding the current state, where at 590, the cache can store the recent states, actions, results, etc.

[0071] According to an embodiment of the present disclosure, at 570, based on the content representing the current state, the large model can generate action suggestions. And in the environment at 580, the action can be executed and a reward score can be obtained. Then, the latest state and action are updated and fed into the cache at 590. As Figure 5 shown, the irrigation regulation process 500 of soilless cultivation in a greenhouse according to an embodiment of the present disclosure forms a closed loop, enabling iterative irrigation regulation for soilless cultivation based on a greenhouse via Figure 5 each component or sub-process therein.

[0072] Figure 6 FIG. shows a schematic diagram of a device 600 for irrigation regulation according to an embodiment of the present disclosure. The device 600 can be applied to soilless cultivation based on a greenhouse and includes a plurality of units or modules for performing corresponding steps in the method 200 as discussed in Figure 2 above. As Figure 6 shown, the device 600 includes: a data acquisition module 610 configured to acquire production static data for soilless cultivation and acquire production dynamic data for soilless cultivation at a first predetermined time frequency. The device further includes an operation prediction module 620 configured to utilize a machine learning model based on an irrigation regulation algorithm for greenhouse soilless cultivation to predict irrigation operations corresponding to irrigation targets for the irrigation machine of the greenhouse based on the acquired production static data and production dynamic data. The device further includes an operation execution module 630 configured to regulate the soilless cultivation based on the greenhouse to correspond to the irrigation target by executing the predicted irrigation operations on the irrigation machine.

[0073] In some embodiments, the production static data for soilless cultivation may include: greenhouse data, which includes the greenhouse type, greenhouse area, lowest height, highest height, whether there is an earthen wall, air vent position, and air vent area of the greenhouse; medium data, which includes the medium type and medium description; and crop data, which includes the crop type and crop introduction. In some embodiments, the greenhouse type may include: warm greenhouse, warm and cold greenhouse, multi-span arch greenhouse, and glass greenhouse.

[0074] In some embodiments, the production dynamic data for soilless cultivation may include: date and time, medium humidity, crop phenological period, current weather information, and weather information within a predetermined future time period. In some embodiments, the crop phenological period may include: germination, seedling stage, seedling period, extended growth stage, tillering stage, branching stage, budding stage, flowering stage, post-flowering stage, fruit swelling stage, and maturity stage.

[0075] In some embodiments, the irrigation regulation algorithm for greenhouse soilless cultivation may be configured to: while minimizing the number of irrigation operations of the irrigation machine, regulate the medium humidity obtained at a second predetermined time frequency to be greater than or equal to a predetermined humidity threshold, and regulate the ratio of the liquid return volume to the irrigation volume to correspond to a predetermined ratio threshold, where the second predetermined time frequency is greater than the first predetermined time frequency.

[0076] In some embodiments, the operation of predicting the irrigation corresponding to the irrigation target for the irrigation machine of the greenhouse may include: converting the obtained production static data and production dynamic data into a model input in a language corresponding to a machine learning model; feeding the model input into the machine learning model; and obtaining the model output of the machine learning model, where the model output indicates the irrigation operation and the time point for performing the irrigation operation within a predetermined time period.

[0077] In some embodiments, the machine learning model may include a large language model based on the soilless cultivation irrigation regulation algorithm in a greenhouse, and the model input for converting the acquired production static data and production dynamic data into a language corresponding to the machine learning model may include: converting the acquired production static data and production dynamic data from structured data into natural language as the model input, and the apparatus 600 may further include: an additional model input generation module configured to convert the target description, guidance, reference examples, and status description from structured data into natural language as the additional model input, where: the target description indicates the target of the soilless cultivation irrigation regulation algorithm in the greenhouse, the guidance includes the relevance between the reference irrigation operation and the environment and the calculation method of the return score, the reference examples include historical examples, representative examples, and other greenhouse examples, and each example in the reference examples includes an example identifier, a reference irrigation operation, and a reference feedback, and the status description indicates the production operation environment based on the acquired production static data and production dynamic data.

[0078] In some embodiments, the apparatus 600 further includes a representative example generation module configured to cluster the other greenhouse example data subset based on the example identifier of each example in the other greenhouse example data subset in the example dataset, where the example identifier includes an embedded vector encoded to represent the corresponding example; and identify the example closest to the center of the example cluster for the other greenhouse example data subset as the representative example.

[0079] In some embodiments, the apparatus 600 further includes an other greenhouse example generation module configured to identify the examples within a predetermined distance threshold from the center of the example cluster for the other greenhouse example data subset as the other greenhouse examples based on the model input and the additional model input.

[0080] In some embodiments, the historical examples, representative examples, and other greenhouse examples may be cached for retrieval based on the model input and the additional model input, and the other examples in the example dataset may be stored in the example database.

[0081] In some embodiments, the apparatus 600 further includes a prompt generation module configured to generate prompts for the large language model, where the prompts are extracted from the following: the greenhouse data and greenhouse status data in the model input; the target description and status description in the additional model input; the cached historical examples, representative examples, and other greenhouse examples; and the other examples in the example database.

[0082] Figure 7FIG. 0 shows a schematic block diagram of an example device 700 that can be used to implement embodiments of the present disclosure. As shown, device 700 includes a central processing unit (CPU) 701 that can perform various appropriate actions and processes according to computer program instructions stored in a read-only memory (ROM) 702 or computer program instructions loaded from a storage unit 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of device 700 can also be stored. The CPU 701, ROM 702, and RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0083] Multiple components in device 700 are connected to the I / O interface 705, including: an input unit 706, such as a keyboard, mouse, etc.; an output unit 707, such as various types of displays, speakers, etc.; a storage unit 708, such as a magnetic disk, optical disk, etc.; and a communication unit 709, such as a network card, modem, wireless communication transceiver, etc. The communication unit 709 allows device 700 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0084] Each of the processes and treatments described above, such as method 200 and process 300 and their sub-processes, can be executed by a processing unit 901. For example, in some embodiments, method 200 and process 300 and their sub-processes can be implemented as a computer software program that is tangibly contained in a machine-readable medium, such as storage unit 708. In some embodiments, part or all of the computer program can be loaded and / or installed onto device 700 via the ROM 702 and / or communication unit 709. When the computer program is loaded into the RAM 703 and executed by the CPU 701, one or more actions of method 200 and process 300 and their sub-processes described above can be performed. According to an embodiment of the present disclosure, a greenhouse is provided, which can include the device 600 described above to perform various aspects of the present disclosure.

[0085] The present disclosure can be a method, a device, an electronic device, a vehicle, a computer-readable storage medium, and / or a computer program product. The computer program product can include a computer-readable storage medium having computer-readable program instructions for performing various aspects of the present disclosure loaded thereon.

[0086] A computer-readable storage medium can be a tangible device that can retain and store instructions for use by an instruction execution device. A computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disc (DVD), a memory stick, a floppy disk, a mechanically encoded device such as a punched card or raised structures in grooves having instructions stored thereon, and any suitable combination of the foregoing. The computer-readable storage medium as used herein is not construed as being a transitory signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., an optical pulse through an optical fiber cable), or an electrical signal transmitted through a wire.

[0087] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to respective computing / processing devices, or can be downloaded to an external computer or an external storage device through a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network may include a copper transmission cable, an optical fiber transmission, a wireless transmission, a router, a firewall, a switch, a gateway computer, and / or an edge server. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in each computing / processing device.

[0088] The computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine - related instructions, microcode, firmware instructions, state - setting data, or source code or object code written in any combination of one or more programming languages, including object - oriented programming languages such as Smalltalk, C++, etc., and conventional procedural programming languages such as the "C" language or similar programming languages. The computer - readable program instructions may be executed entirely on the user's computer, partially on the user's computer, executed as a stand - alone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider). In some embodiments, by using the state information of the computer - readable program instructions to customize an electronic circuit, such as a programmable logic circuit, a field - programmable gate array (FPGA), or a programmable logic array (PLA), the electronic circuit can execute the computer - readable program instructions to implement various aspects of the present disclosure.

[0089] Aspects of the present disclosure are described herein with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present disclosure. It should be understood that each block of the flowcharts and / or block diagrams, and combinations of blocks in the flowcharts and / or block diagrams, can be implemented by computer - readable program instructions.

[0090] These computer - readable program instructions can be provided to a processing unit of a general - purpose computer, a special - purpose computer, or other programmable data - processing apparatus to produce a machine such that, when the instructions are executed by the processing unit of the computer or other programmable data - processing apparatus, a device is created that implements the functions / acts specified in one or more blocks of the flowchart and / or block diagram. These computer - readable program instructions can also be stored in a computer - readable storage medium, which causes a computer, a programmable data - processing apparatus, and / or other devices to operate in a particular manner, so that the computer - readable medium storing the instructions includes a manufacture, which includes instructions for implementing various aspects of the functions / acts specified in one or more blocks of the flowchart and / or block diagram.

[0091] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device, causing a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process such that the instructions executed on the computer, other programmable data processing apparatus, or other device implement the functions / acts specified in one or more blocks of the flowchart and / or block diagram.

[0092] The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a portion of an instruction, which contains one or more executable instructions for implementing the specified logical function. In some alternative implementations, the functions noted in the blocks may occur out of the order noted in the figures. For example, two consecutive blocks may in fact be executed substantially in parallel, or they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block of the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented by a dedicated hardware-based system that performs the specified functions or acts, or by a combination of dedicated hardware and computer instructions.

[0093] The embodiments of the present disclosure have been described above. The above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The choice of terms used herein is intended to best explain the principles of the embodiments, the practical application, or the technical improvement of technologies in the market, or to enable other ordinary skilled persons in the art to understand the embodiments disclosed herein.

Claims

1. A method for irrigation regulation, which is applied to soilless cultivation based on a greenhouse, and includes: Obtaining production static data for the soilless cultivation, and obtaining production dynamic data for the soilless cultivation at a first predetermined time frequency; Using a machine learning model based on an irrigation regulation algorithm for greenhouse soilless cultivation, predicting an irrigation operation corresponding to an irrigation target for an irrigation machine of the greenhouse based on the obtained production static data and production dynamic data; And Regulating the soilless cultivation based on the greenhouse to correspond to the irrigation target by performing the predicted irrigation operation on the irrigation machine.

2. The method according to claim 1, wherein the production static data for the soilless cultivation includes: Greenhouse data, which includes the greenhouse type, greenhouse area, lowest point height, highest point height, whether there is an earthen wall, air vent position and air vent area of the greenhouse; Medium data, which includes medium type and medium description; and Crop data, which includes crop type and crop introduction.

3. The method according to claim 2, wherein the greenhouse type includes: Warm greenhouse, cold and warm greenhouse, multi-span arch greenhouse, glass greenhouse.

4. The method according to claim 1, wherein the production dynamic data for the soilless cultivation includes: Date and time, medium humidity, crop phenological period, current weather information and weather information within a predetermined time period in the future.

5. The method according to claim 4, wherein the crop phenological period includes: Germination, seedling stage, seedling period, extended growth stage, tillering stage, branching stage, budding stage, flowering stage, post-flowering stage, fruit expansion stage, maturity stage.

6. The method according to claim 4, wherein the irrigation regulation algorithm for greenhouse soilless cultivation is configured to: while minimizing the number of irrigation times of the irrigation machine, regulate the medium humidity obtained at a second predetermined time frequency to be greater than or equal to a predetermined humidity threshold, and regulate the ratio of the return liquid volume to the irrigation volume to correspond to a predetermined ratio threshold, and the second predetermined time frequency is greater than the first predetermined time frequency.

7. The method according to claim 1, wherein predicting the irrigation operation corresponding to the irrigation target for the irrigation machine of the greenhouse includes: Converting the obtained production static data and production dynamic data into a model input in a language corresponding to the machine learning model; Feeding the model input into the machine learning model; And Obtaining a model output of the machine learning model, where the model output indicates the irrigation operation and the time point for performing the irrigation operation within a predetermined time period.

8. The method according to claim 7, wherein the machine learning model includes a large language model based on the soilless cultivation irrigation regulation algorithm for the greenhouse, and the model input that converts the obtained production static data and production dynamic data into the language corresponding to the machine learning model includes: Converting the obtained production static data and production dynamic data from structured data into the model input in natural language, and wherein the method further includes: converting a target description, instructions, reference examples, and status descriptions from structured data into additional model inputs in natural language, where: The target description indicates the target of the soilless cultivation irrigation regulation algorithm for the greenhouse, and the guidance includes referring to the correlation between irrigation operations and the environment and the calculation method of the return score. The reference examples include historical examples, representative examples, and other greenhouse examples, and each example in the reference examples includes an example identifier, a reference irrigation operation, and a reference feedback, and The state description indicates the production operation environment based on the acquired production static data and production dynamic data.

9. The method according to claim 8, wherein the representative example is obtained by the following actions: Based on the example identifier of each example in the other greenhouse example data subset in the example dataset, clustering the other greenhouse example data subset, and the example identifier includes an embedded vector encoded to represent the corresponding example; and Identifying the example closest to the center of the example cluster for the other greenhouse example data subset as the representative example.

10. The method according to claim 9, wherein the other greenhouse examples are obtained by the following actions: Based on the model input and the additional model input, identifying the examples within a predetermined distance threshold from the center of the example cluster for the other greenhouse example data subset as the other greenhouse examples.

11. The method according to claim 8, wherein: The historical examples, the representative examples, and the other greenhouse examples are cached for retrieval based on the model input and the additional model input, and The other examples in the example dataset are stored in an example database.

12. The method according to claim 11 further comprises: Generating a prompt for the large language model, wherein the prompt is extracted from the following: The production static data and production dynamic data in the model input; The target description and state description in the additional model input; The cached historical examples, representative examples, and other greenhouse examples; And The other examples in the example database.

13. An apparatus for irrigation regulation, the apparatus being applied to soilless cultivation based on a greenhouse and comprising: A data acquisition module configured to acquire production static data for the soilless cultivation and acquire production dynamic data for the soilless cultivation at a first predetermined time frequency; An operation prediction module configured to use a machine learning model based on the soilless cultivation irrigation regulation algorithm for the greenhouse to predict an irrigation operation corresponding to an irrigation target for an irrigation machine of the greenhouse based on the acquired production static data and production dynamic data; And An operation execution module configured to regulate the soilless cultivation based on the greenhouse to correspond to the irrigation target by executing the predicted irrigation operation for the irrigation machine.

14. An electronic device comprising: At least one processor; And A memory coupled to the at least one processor and having instructions stored thereon that, when executed by the at least one processor, cause the electronic device to perform the method according to any one of claims 1-12.

15. A greenhouse, comprising the electronic device according to claim 14.

16. A computer-readable storage medium having computer-executable instructions stored thereon, wherein the computer-executable instructions are executed by a processor of a computer to implement the method according to any one of claims 1-12.

Images