Greenhouse environment control method and system based on multi-modal perception and adaptive decision

Abstract

The invention provides a greenhouse environment control method and system based on multi-modal perception and adaptive decision, and the method comprises the steps: carrying out the preprocessing of multi-modal monitoring data, so as to obtain the preprocessing data; determining node weights of the greenhouse sensor nodes, and performing fusion processing on the preprocessed data based on the node weights to obtain fusion data; obtaining a pre-trained environment prediction model, and inputting the fusion data and the action combination into the pre-trained environment prediction model to output predicted environment data; and determining the control score of each action combination based on the predicted environment data, selecting the action combination corresponding to the minimum control score as the optimal action combination, and controlling the greenhouse environment equipment to perform greenhouse environment control according to the final action combination. Meanwhile, key difference information in the environment data can be effectively captured, the complexity of information redundancy and data processing is reduced, and then precise control over the greenhouse environment is achieved.

Description

Technical Field

[0001] This invention belongs to the technical field of greenhouse control, specifically relating to a greenhouse environment control method and system based on multimodal perception and adaptive decision-making. Background Technology

[0002] A greenhouse is a building that controls or partially controls the environment in which plants grow. It is mainly used for non-seasonal or non-regional plant cultivation, scientific research, generational breeding, and ornamental plant cultivation. Greenhouses come in many types, and can be further categorized based on different roof frame materials, lighting materials, shapes, and heating conditions, such as glass greenhouses, polycarbonate greenhouses; single-span greenhouses, multi-span greenhouses; single-roof greenhouses, double-roof greenhouses; heated greenhouses, unheated greenhouses, etc. With the development of the Internet of Things, a large number of control devices are used in greenhouses to control the environment in combination. However, the following problems exist: the fusion of multi-source environmental data in the greenhouse is not ideal; the nonlinearity and lag of environmental parameters cause control problems; and the inability to monitor the operating status of equipment in real time is insufficient. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention provides a greenhouse environment control method and system based on multimodal perception and adaptive decision-making, which solves the technical problems in the prior art.

[0004] In a first aspect, the present invention provides the following technical solution: a greenhouse environment control method based on multimodal sensing and adaptive decision-making, comprising: Acquire multimodal monitoring data collected from different greenhouse sensor nodes, and preprocess the multimodal monitoring data to obtain preprocessed data; The node weights of the greenhouse sensor nodes are determined, and the preprocessed data is fused based on the node weights to obtain fused data. Determine the action combination of the greenhouse environment control equipment, obtain a pre-trained environment prediction model, and input the fused data and the action combination into the pre-trained environment prediction model to output predicted environment data. Based on the predicted environmental data, a control score is determined for each action combination. The action combination with the lowest control score is selected as the optimal action combination, and the greenhouse environment equipment is controlled to perform greenhouse environment control according to the final action combination.

[0005] Compared with existing technologies, the beneficial effects of this invention are as follows: First, this invention acquires multimodal monitoring data collected from different greenhouse sensor nodes, preprocesses the multimodal monitoring data to obtain preprocessed data; determines the node weights of the greenhouse sensor nodes, and fuses the preprocessed data based on the node weights to obtain fused data; determines the action combinations of the greenhouse environmental control equipment, acquires a pre-trained environmental prediction model, and inputs the fused data and action combinations into the pre-trained environmental prediction model to output predicted environmental data; determines the control score for each action combination based on the predicted environmental data, selects the action combination with the minimum control score as the optimal action combination, and controls the greenhouse environmental equipment to perform greenhouse environmental control according to the final action combination. This invention achieves objective quantification of data value and dynamic weight allocation by calculating the node weights of each node, while effectively capturing key differences in environmental data, reducing information redundancy and data processing complexity. Then, the optimal action combination is selected through the model and corresponding scores to achieve precise control of the greenhouse environment. Preferably, the step of preprocessing the multimodal monitoring data to obtain preprocessed data includes: Identify key parameters in the multimodal monitoring data, including the first quartile. Third and quartiles ; Determine abnormal data based on the aforementioned key parameters. : ; The abnormal data is removed from the multimodal monitoring data and the missing values ​​are filled using mean interpolation to obtain preprocessed data.

[0006] Preferably, the step of determining the node weights of the greenhouse sensor nodes and fusing the preprocessed data based on the node weights to obtain fused data includes: A data matrix is ​​determined based on the preprocessed data and the corresponding greenhouse sensor nodes. : ; In the formula, Indicates the first in the preprocessed data The greenhouse sensor node at the... Observations under various data types; The range normalization method is used to map the elements in the data matrix to... Within the range, to obtain the mapping matrix; Calculate data feature values ​​for different data types based on the mapping matrix. : ; In the formula, The first in the mapping matrix Line number Column elements; Based on the data feature values Calculate data type weights : ; Based on the weight of the data type Determine node weights : ; Based on the node weights The preprocessed data is then fused to obtain fused data. : .

[0007] Preferably, the environmental prediction model is a deep neural network.

[0008] Preferably, the step of determining the action combination of the greenhouse environment control equipment includes: Determine the device control action for each of the greenhouse environment control devices, and combine the device control actions of the greenhouse environment control devices to obtain an initial action set; Identify the linkage constraints between the greenhouse environment control devices, and filter the initial set of actions based on the linkage constraints to obtain action combinations.

[0009] Preferably, the step of determining the control score for each action combination based on the predicted environmental data includes: A first control score is determined based on the predicted environmental data. : ; In the formula, These are the weighting coefficients for air temperature, air humidity, light intensity, and carbon dioxide concentration, respectively. These are the normalization constants for air temperature, air humidity, light intensity, and carbon dioxide concentration, respectively. These represent the distances from the predicted environmental data (air temperature, air humidity, light intensity, carbon dioxide concentration) to the boundaries of the preset intervals. A second control score is determined based on the aforementioned action combination. : ; In the formula, These respectively represent greenhouse environmental control equipment. exist , The device control actions at any given time, Greenhouse environmental control equipment energy consumption level These represent the state change penalty coefficient and the energy consumption penalty coefficient, respectively. Based on the first control score Second control score Determine control score : .

[0010] Secondly, the present invention provides the following technical solution: a greenhouse environment control system based on multimodal sensing and adaptive decision-making, the system comprising: The preprocessing module is used to acquire multimodal monitoring data collected from different greenhouse sensor nodes, and to preprocess the multimodal monitoring data to obtain preprocessed data. The fusion module is used to determine the node weights of the greenhouse sensor nodes and perform fusion processing on the preprocessed data based on the node weights to obtain fused data. The prediction module is used to determine the action combination of the greenhouse environment control equipment, obtain a pre-trained environment prediction model, and input the fused data and the action combination into the pre-trained environment prediction model to output predicted environment data. The control module is used to determine the control score of each action combination based on the predicted environmental data, select the action combination with the smallest control score as the optimal action combination, and control the greenhouse environment equipment to perform greenhouse environment control according to the final action combination.

[0011] Preferably, the preprocessing module is specifically used for: Identify key parameters in the multimodal monitoring data, including the first quartile. Third and quartiles ; Determine abnormal data based on the aforementioned key parameters. : ; The abnormal data is removed from the multimodal monitoring data and the missing values ​​are filled using mean interpolation to obtain preprocessed data.

[0012] Thirdly, the present invention provides the following technical solution: a computer, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the greenhouse environment control method based on multimodal perception and adaptive decision-making as described above.

[0013] Fourthly, the present invention provides the following technical solution: a storage medium storing a computer program, wherein when the computer program is executed by a processor, it implements the greenhouse environment control method based on multimodal perception and adaptive decision-making as described above. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 A flowchart of a greenhouse environment control method based on multimodal sensing and adaptive decision-making provided in Embodiment 1 of the present invention; Figure 2 This is a structural block diagram of a greenhouse environment control system based on multimodal perception and adaptive decision-making provided in Embodiment 2 of the present invention; Figure 3 This is a schematic diagram of the hardware structure of a computer provided for another embodiment of the present invention.

[0016] The embodiments of the present invention will be further described below with reference to the accompanying drawings. Detailed Implementation

[0017] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain embodiments of the present invention, and should not be construed as limiting the present invention.

[0018] Example 1 In Embodiment 1 of the present invention, as Figure 1 As shown, a greenhouse environment control method based on multimodal sensing and adaptive decision-making includes: S1. Acquire multimodal monitoring data collected by different greenhouse sensor nodes, and preprocess the multimodal monitoring data to obtain preprocessed data; Step S1 includes: S11. Identify key parameters in the multimodal monitoring data, the key parameters including the first quartile. Third and quartiles ; Specifically, quartiles can reflect the distribution characteristics of data, so they are selected as a key parameter as a benchmark for screening outlier data.

[0019] S12. Determine abnormal data based on the key parameters. : .

[0020] S13. The abnormal data is removed from the multimodal monitoring data and the missing values ​​are filled by mean interpolation to obtain preprocessed data.

[0021] S2. Determine the node weights of the greenhouse sensor nodes, and perform fusion processing on the preprocessed data based on the node weights to obtain fused data; Step S2 includes: S21. Determine the data matrix based on the preprocessed data and the corresponding greenhouse sensor nodes. : ; In the formula, Indicates the first in the preprocessed data The greenhouse sensor node at the... Observations under various data types.

[0022] S22. Map the elements in the data matrix to range normalization using the range standardization method. Within the range, to obtain the mapping matrix.

[0023] S23. Calculate data feature values ​​for different data types based on the mapping matrix. : ; In the formula, The first in the mapping matrix Line number The elements of the column.

[0024] S24. Based on the data feature values Calculate data type weights : .

[0025] S25, Based on the weight of the data type Determine node weights : .

[0026] S26. Based on the node weights The preprocessed data is then fused to obtain fused data. : .

[0027] S3. Determine the action combination of the greenhouse environment control equipment, obtain the pre-trained environment prediction model, and input the fused data and the action combination into the pre-trained environment prediction model to output predicted environment data. Specifically, the environmental prediction model is a deep neural network.

[0028] The step of determining the action combination of the greenhouse environment control equipment includes: S31. Determine the device control action of each greenhouse environment control device, and combine the device control actions of the greenhouse environment control devices to obtain an initial action set; S32. Identify the linkage constraints between the greenhouse environment control devices, and filter the initial action set based on the linkage constraints to obtain action combinations.

[0029] Specifically, in this application, the greenhouse environmental control equipment includes negative pressure fans, wet curtains, top skylights, internal shading, and side shading. The negative pressure fans and wet curtains are used to regulate the temperature and humidity inside the greenhouse, reducing the indoor temperature and increasing the humidity through air flow and water evaporation. The top skylights are used to promote natural air convection and ventilation to improve the quality of the indoor environment. The internal and side shading are used to regulate the light intensity inside the greenhouse to avoid excessive light from adversely affecting the crops.

[0030] In practice, the number of initial action sets is relatively large, so it is necessary to filter them through linkage constraints. For example, when the fan is turned on, the roof window should be closed as much as possible to avoid unnecessary energy consumption or airflow disturbance in the greenhouse; when the wet curtain is turned on, the fan needs to be turned on and the roof window needs to be closed to ensure the synergistic effect of air circulation and humidity control; when the shading is completely closed, the ventilation effect of the roof window will be limited.

[0031] S4. Based on the predicted environmental data, determine the control score for each action combination, select the action combination with the smallest control score as the optimal action combination, and control the greenhouse environment equipment to perform greenhouse environment control according to the final action combination. Step S4 includes: S41. Determine the first control score based on the predicted environment data. : ; In the formula, These are the weighting coefficients for air temperature, air humidity, light intensity, and carbon dioxide concentration, respectively. These are the normalization constants for air temperature, air humidity, light intensity, and carbon dioxide concentration, respectively. These represent the distances from the predicted environmental data (air temperature, air humidity, light intensity, carbon dioxide concentration) to the boundaries of the preset intervals. Specifically, during the model prediction process, the current fused data is input into the environmental prediction model, and based on this, all possible action combinations are input to predict the changes in the greenhouse environment at a given moment under different equipment actions. The control effect is evaluated through control scoring, that is, by quantitatively evaluating the predicted environmental data and the selected action combinations, the action combination with the highest score is selected. In the actual evaluation process, multiple environmental control indicators such as indoor air temperature, relative humidity, CO2 concentration, and light intensity need to be comprehensively considered. At the same time, frequent adjustments to environmental control equipment should be avoided, and actions with lower energy consumption should be selected as much as possible.

[0032] S42. Determine the second control score based on the action combination. : ; In the formula, These respectively represent greenhouse environmental control equipment. exist , The device control actions at any given time, Greenhouse environmental control equipment energy consumption level These represent the state change penalty coefficient and the energy consumption penalty coefficient, respectively. S43, Based on the first control score Second control score Determine control score : .

[0033] The first control score primarily measures the degree to which environmental variables deviate from the target range, reflecting the effectiveness of the control strategy in meeting crop growth needs. The second control score is used to constrain large-scale regulatory actions, avoid causing drastic fluctuations in the greenhouse environment, and impose additional penalties on energy-intensive regulatory actions.

[0034] The greenhouse environment control method based on multimodal perception and adaptive decision-making provided in Embodiment 1 of this invention first acquires multimodal monitoring data collected by different greenhouse sensor nodes, preprocesses the multimodal monitoring data to obtain preprocessed data, determines the node weights of the greenhouse sensor nodes, and fuses the preprocessed data based on the node weights to obtain fused data, determines the action combination of the greenhouse environment control equipment, obtains a pre-trained environment prediction model, and inputs the fused data and action combination into the pre-trained environment prediction model to output predicted environment data, determines the control score of each action combination based on the predicted environment data, selects the action combination with the minimum control score as the optimal action combination, and controls the greenhouse environment equipment to perform greenhouse environment control according to the final action combination. This invention achieves objective quantification of data value and dynamic weight allocation by calculating the node weights of each node, and can effectively capture key difference information in the environmental data, reducing information redundancy and data processing complexity. Then, the optimal action combination is selected through the model and the corresponding score to achieve precise control of the greenhouse environment.

[0035] Example 2 like Figure 2 As shown, in Embodiment 2 of the present invention, a greenhouse environment control system based on multimodal sensing and adaptive decision-making is provided. The system includes: The preprocessing module 1 is used to acquire multimodal monitoring data collected by different greenhouse sensor nodes, and to preprocess the multimodal monitoring data to obtain preprocessed data. Fusion module 2 is used to determine the node weights of the greenhouse sensor nodes, and to perform fusion processing on the preprocessed data based on the node weights to obtain fused data; Prediction module 3 is used to determine the action combination of greenhouse environmental control equipment, obtain a pre-trained environmental prediction model, and input the fused data and the action combination into the pre-trained environmental prediction model to output predicted environmental data. Control module 4 is used to determine the control score of each action combination based on the predicted environmental data, select the action combination with the smallest control score as the optimal action combination, and control the greenhouse environment equipment to perform greenhouse environment control according to the final action combination.

[0036] Specifically, the preprocessing module 1 is used for: Identify key parameters in the multimodal monitoring data, including the first quartile. Third and quartiles ; Determine abnormal data based on the aforementioned key parameters. : ; The abnormal data is removed from the multimodal monitoring data and the missing values ​​are filled using mean interpolation to obtain preprocessed data.

[0037] Specifically, the fusion module 2 is used for: A data matrix is ​​determined based on the preprocessed data and the corresponding greenhouse sensor nodes. : ; In the formula, Indicates the first in the preprocessed data The greenhouse sensor node at the... Observations under various data types; The range normalization method is used to map the elements in the data matrix to... Within the range, to obtain the mapping matrix; Calculate data feature values ​​for different data types based on the mapping matrix. : ; In the formula, The first in the mapping matrix Line number Column elements; Based on the data feature values Calculate data type weights : ; Based on the weight of the data type Determine node weights : ; Based on the node weights The preprocessed data is then fused to obtain fused data. : .

[0038] Specifically, the prediction module 3 is used for: Determine the device control action for each of the greenhouse environment control devices, and combine the device control actions of the greenhouse environment control devices to obtain an initial action set; Identify the linkage constraints between the greenhouse environment control devices, and filter the initial set of actions based on the linkage constraints to obtain action combinations.

[0039] Specifically, the control module 4 is used for: A first control score is determined based on the predicted environmental data. : ; In the formula, These are the weighting coefficients for air temperature, air humidity, light intensity, and carbon dioxide concentration, respectively. These are the normalization constants for air temperature, air humidity, light intensity, and carbon dioxide concentration, respectively. These represent the distances from the predicted environmental data (air temperature, air humidity, light intensity, carbon dioxide concentration) to the boundaries of the preset intervals. A second control score is determined based on the aforementioned action combination. : ; In the formula, These respectively represent greenhouse environmental control equipment. exist , The device control actions at any given time, Greenhouse environmental control equipment energy consumption level These represent the state change penalty coefficient and the energy consumption penalty coefficient, respectively. Based on the first control score Second control score Determine control score : .

[0040] In other embodiments of the present invention, the present invention provides the following technical solution: a computer, including a memory 102, a processor 101, and a computer program stored in the memory 102 and executable on the processor 101, wherein the processor 101 executes the computer program to implement the greenhouse environment control method based on multimodal perception and adaptive decision-making as described above.

[0041] Specifically, the processor 101 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of the present invention.

[0042] The memory 102 may include a large-capacity memory for data or instructions. For example, and not limitingly, the memory 102 may include a hard disk drive (HDD), a floppy disk drive, a solid-state drive (SSD), flash memory, an optical disk drive, a magneto-optical disk drive, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 102 may include removable or non-removable (or fixed) media. Where appropriate, the memory 102 may be internal or external to a data processing device. In a particular embodiment, the memory 102 is non-volatile memory. In a particular embodiment, the memory 102 includes read-only memory (ROM) and random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable read-only memory (PROM), an erasable read-only memory (EPROM), an electrically erasable read-only memory (EEPROM), an electrically alterable read-only memory (EAROM), or flash memory, or a combination of two or more of these. Where appropriate, the RAM can be Static Random-Access Memory (SRAM) or Dynamic Random-Access Memory (DRAM). DRAM can be Fast Page Mode Dynamic Random Access Memory (FPMDRAM), Extended Data Out Dynamic Random Access Memory (EDODRAM), Synchronous Dynamic Random-Access Memory (SDRAM), etc.

[0043] The memory 102 can be used to store or cache various data files that need to be processed and / or used for communication, as well as possible computer program instructions executed by the processor 101.

[0044] The processor 101 reads and executes the computer program instructions stored in the memory 102 to implement the greenhouse environment control method based on multimodal perception and adaptive decision-making.

[0045] In some embodiments, the computer may further include a communication interface 103 and a bus 100. For example, Figure 3 As shown, the processor 101, memory 102, and communication interface 103 are connected through bus 100 and complete communication with each other.

[0046] The communication interface 103 is used to enable communication between the various modules, devices, units, and / or equipment in the embodiments of the present invention. The communication interface 103 can also enable data communication with other components such as external devices, image / data acquisition devices, databases, external storage, and image / data processing workstations.

[0047] Bus 100 includes hardware, software, or both, that couples components of a computer device together. Bus 100 includes, but is not limited to, at least one of the following: data bus, address bus, control bus, expansion bus, and local bus. For example, and not as a limitation, bus 100 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses, or a combination of two or more of these. Where appropriate, bus 100 may include one or more buses. Although specific buses are described and illustrated in the embodiments of the present invention, the present invention is contemplated by any suitable bus or interconnect.

[0048] The computer can acquire a greenhouse environment control system based on multimodal perception and adaptive decision-making, and execute the greenhouse environment control method based on multimodal perception and adaptive decision-making of this invention, thereby realizing greenhouse environment control based on multimodal perception and adaptive decision-making.

[0049] In some further embodiments of the present invention, in conjunction with the above-described greenhouse environment control method based on multimodal perception and adaptive decision-making, the present invention provides the following technical solution: a storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the above-described greenhouse environment control method based on multimodal perception and adaptive decision-making.

[0050] Those skilled in the art will understand that the logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can mean any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0051] More specific examples of readable media (a non-exhaustive list) include: electrical connections (electronic devices) with one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0052] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0053] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0054] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A greenhouse environment control method based on multimodal sensing and adaptive decision-making, characterized in that, include: Acquire multimodal monitoring data collected from different greenhouse sensor nodes, and preprocess the multimodal monitoring data to obtain preprocessed data; The node weights of the greenhouse sensor nodes are determined, and the preprocessed data is fused based on the node weights to obtain fused data. Determine the action combination of the greenhouse environment control equipment, obtain a pre-trained environment prediction model, and input the fused data and the action combination into the pre-trained environment prediction model to output predicted environment data. Based on the predicted environmental data, a control score is determined for each action combination. The action combination with the lowest control score is selected as the optimal action combination, and the greenhouse environment equipment is controlled to perform greenhouse environment control according to the final action combination.

2. The greenhouse environment control method based on multimodal perception and adaptive decision-making according to claim 1, characterized in that, The step of preprocessing the multimodal monitoring data to obtain preprocessed data includes: Identify key parameters in the multimodal monitoring data, including the first quartile. Third and quartiles ; Determine abnormal data based on the aforementioned key parameters. : ； The abnormal data is removed from the multimodal monitoring data and the missing values ​​are filled using mean interpolation to obtain preprocessed data.

3. The greenhouse environment control method based on multimodal perception and adaptive decision-making according to claim 1, characterized in that, The step of determining the node weights of the greenhouse sensor nodes and fusing the preprocessed data based on the node weights to obtain fused data includes: A data matrix is ​​determined based on the preprocessed data and the corresponding greenhouse sensor nodes. : ； In the formula, Indicates the first in the preprocessed data The greenhouse sensor node at the... Observations under various data types; The range normalization method is used to map the elements in the data matrix to... Within the range, to obtain the mapping matrix; Calculate data feature values ​​for different data types based on the mapping matrix. : ； In the formula, The first in the mapping matrix Line number Column elements; Based on the data feature values Calculate data type weights : ； Based on the weight of the data type Determine node weights : ； Based on the node weights The preprocessed data is then fused to obtain fused data. : 。 4. The greenhouse environment control method based on multimodal perception and adaptive decision-making according to claim 1, characterized in that, The environmental prediction model is specifically a deep neural network.

5. The greenhouse environment control method based on multimodal sensing and adaptive decision-making according to claim 1, characterized in that, The steps for determining the action combination of the greenhouse environment control equipment include: Determine the device control action for each of the greenhouse environment control devices, and combine the device control actions of the greenhouse environment control devices to obtain an initial action set; Identify the linkage constraints between the greenhouse environment control devices, and filter the initial set of actions based on the linkage constraints to obtain action combinations.

6. The greenhouse environment control method based on multimodal sensing and adaptive decision-making according to claim 1, characterized in that, The step of determining the control score for each action combination based on the predicted environmental data includes: A first control score is determined based on the predicted environmental data. : ； In the formula, These are the weighting coefficients for air temperature, air humidity, light intensity, and carbon dioxide concentration, respectively. These are the normalization constants for air temperature, air humidity, light intensity, and carbon dioxide concentration, respectively. These represent the distances from the predicted environmental data (air temperature, air humidity, light intensity, carbon dioxide concentration) to the boundaries of the preset intervals. A second control score is determined based on the aforementioned action combination. : ； In the formula, These respectively represent greenhouse environmental control equipment. exist , The device control actions at any given time, Greenhouse environmental control equipment energy consumption level These represent the state change penalty coefficient and the energy consumption penalty coefficient, respectively. Based on the first control score Second control score Determine control score : 。 7. A greenhouse environment control system based on multimodal sensing and adaptive decision-making, characterized in that, The system includes: The preprocessing module is used to acquire multimodal monitoring data collected from different greenhouse sensor nodes, and to preprocess the multimodal monitoring data to obtain preprocessed data. The fusion module is used to determine the node weights of the greenhouse sensor nodes and perform fusion processing on the preprocessed data based on the node weights to obtain fused data. The prediction module is used to determine the action combination of the greenhouse environment control equipment, obtain a pre-trained environment prediction model, and input the fused data and the action combination into the pre-trained environment prediction model to output predicted environment data. The control module is used to determine the control score of each action combination based on the predicted environmental data, select the action combination with the smallest control score as the optimal action combination, and control the greenhouse environment equipment to perform greenhouse environment control according to the final action combination.

8. The greenhouse environment control system based on multimodal perception and adaptive decision-making according to claim 7, characterized in that, The preprocessing module is specifically used for: Identify key parameters in the multimodal monitoring data, including the first quartile. Third and quartiles ; Determine abnormal data based on the aforementioned key parameters. : ； The abnormal data is removed from the multimodal monitoring data and the missing values ​​are filled using mean interpolation to obtain preprocessed data.

9. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the greenhouse environment control method based on multimodal perception and adaptive decision-making as described in any one of claims 1 to 6.

10. A storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the greenhouse environment control method based on multimodal perception and adaptive decision-making as described in any one of claims 1 to 6.

Images