Intelligent greenhouse management system and intelligent management method based on crop growth

Abstract

The invention discloses a greenhouse intelligent management system and management method based on crop growth vigor, and belongs to the technical field of intelligent agriculture. The greenhouse intelligent management system comprises a crop growth vigor and environment monitoring system, a control system and matched operation facilities, the crop growth vigor and environment monitoring system can monitor environment information and crop growth vigor conditions of a greenhouse in real time, and meanwhile, the monitoring information is sent to the control system through a communication module; based on a crop growth identification algorithm fusing RGB and hyperspectral images and a corresponding greenhouse intelligent management method, the control system reasonably regulates and controls application parameters of matched operation facilities and dynamically regulates internal environment elements of the greenhouse, so that rapid growth of greenhouse crops is promoted in an intensive mode, and the greenhouse quality is improved. Meanwhile, the labor intensity and the cost are effectively reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of smart agriculture technology, in particular to a smart greenhouse management system based on crop growth and an intelligent management method. BACKGROUND

[0002] Agricultural greenhouse can realize periodic, all-weather, and seasonal large-scale planting. The promotion and application of greenhouse planting technology greatly improve the land utilization efficiency and effectively make up for the gap in fruit and vegetable production due to the large population.

[0003] Compared with traditional agricultural planting greenhouses, smart agricultural greenhouses integrated with modern planting concepts, agricultural informatization technology, new agricultural materials, and automatic control methods represent a revolutionary progress in agricultural engineering technology. Smart agricultural greenhouses usually use temperature, humidity, and CO2 gas sensors to regularly detect the greenhouse planting environment, and the detection results are used as the basis for the operation of the greenhouse environment facilities, thereby realizing dynamic adjustment of the greenhouse planting environment and effectively improving crop yield. However, the existing smart agricultural greenhouse facilities also have the problem of single detection sensor, which cannot realize high dynamic and multi-sensor fusion of crop growth state monitoring, and there is still a lack of a completely intelligent smart greenhouse management system.

[0004] In summary, smart agricultural greenhouses are an important carrier for future fruit and vegetable planting. By collecting real-time information such as temperature, humidity, soil trace elements, and images in the greenhouse, and effectively fusing multi-sensor information, the growth of crops is evaluated through fusion data, and intelligent management of crops in the greenhouse is implemented, thereby forming a smart greenhouse management system based on crop growth, or one of the important development directions of future smart agricultural greenhouses. SUMMARY

[0005] In order to make up for the technical shortcomings of the smart agricultural greenhouse management system, the present application designs a smart greenhouse management system based on crop growth and an intelligent management method to meet the intelligent planting needs of greenhouse crops.

[0006] A smart greenhouse management system based on crop growth, comprising:

[0007] A crop growth and environment monitoring system, comprising a temperature detection system, a humidity detection system, and a crop growth identification system, the crop growth identification system comprising an RGB camera, a hyperspectral camera, and a microprocessor, based on the data collected by the RGB camera and the hyperspectral camera, the microprocessor judges the crop growth through a crop growth identification algorithm, and transmits the corresponding results to the matching operation facilities;

[0008] The control system comprises a distribution box and an intelligent terminal controller, and the intelligent terminal controller controls the growth environment and required materials of crops in real time based on an intelligent management strategy according to the growth of the crops.

[0009] The matched operation facility comprises a ventilation device, a seedling pressing device, a temperature control device and a water and fertilizer integrated device.

[0010] In the technical scheme, the temperature detection system comprises an LM35 temperature sensor arranged at the middle of the greenhouse and close to one side of the plastic film, which is used for detecting temperature data in the greenhouse, and the Arduino development board reads and stores the temperature data.

[0011] In the technical scheme, the humidity detection system comprises a humidity sensor inserted into the soil in the greenhouse, which is used for detecting humidity data of the soil, and the Arduino development board reads and stores the humidity data.

[0012] In the technical scheme, the RGB camera is arranged at the middle of the greenhouse and close to one side of the plastic film, which is used for monitoring apparent growth information of the crops in real time; and the hyperspectral camera is arranged in the greenhouse, which is used for collecting spectral information of the crops.

[0013] In the technical scheme, the ventilation device is installed at the edge of one side of the greenhouse, which is used for realizing ventilation in the greenhouse; the ventilation device comprises a roller shaft, a coupling A, a stepping motor A, a guide shaft, two rolling bodies, a stepping motor B and a plastic film; the roller shaft is inserted into the edge of the plastic film in the transverse direction, so as to realize rolling and opening of the plastic film; the roller shaft is connected with the input shaft of the stepping motor A through the coupling A, and the stepping motor A is fixed on the connecting plate of the guide shaft and installed vertically on one side of the connecting plate; the stepping motor B is connected with the pulley groove of one rolling body through a transmission belt.

[0014] In the technical scheme, the seedling pressing device is arranged in multiple, and arranged along various planting rows or columns, and specifically comprises a seedling pressing rod, a stepping motor C, a coupling B and a transmission gear set; the output end of the stepping motor C is connected with one end of the coupling B, the other end of the coupling B is connected with the large gear of the transmission gear set through a short shaft, the large gear is engaged with the small gear of the transmission gear set, and the small gear is sleeved on the seedling pressing rod.

[0015] The water and fertilizer integrated device is connected with the seedling pressing device, and specifically comprises a water valve, a rudder engine A, a rudder engine B, a three-way pipe, a nutrient solution control valve, a nutrient solution storage tank, and a spray head installed along the seedling pressing rod; the nutrient solution storage tank is located at one end of the water and fertilizer integrated device, connected with the nutrient solution control valve through a pipeline, and the valve body end of the nutrient solution control valve is connected with the rudder engine B; the nutrient solution control valve is connected with one end of the three-way pipe through a pipeline, the second port of the three-way pipe is connected with the water valve, and the valve body end of the water valve is connected with the rudder engine A; the third port of the three-way pipe is connected with the seedling pressing rod.

[0016] In the above technical solution, the temperature control device comprises a resistance wire, a fan and a motor, the torque output end of the motor is connected with the rotating shaft of the fan, and the whole is installed at the front end of the resistance wire.

[0017] An intelligent management method of a smart greenhouse management system based on crop growth, which identifies the apparent growth state of crops based on RGB image information of crops, is used to distinguish the growth stages of crops, including seedling stage, flowering stage and fruit expansion stage; in each growth cycle, the intelligent identification algorithm based on crop growth fuses RGB image and hyperspectral data to analyze crop growth, and the control system dynamically adjusts the application ratio of various materials and the environmental factors of crop growth.

[0018] Further, the intelligent identification algorithm based on crop growth fuses RGB image and hyperspectral data to analyze crop growth, specifically:

[0019] Extract image features and spectral features of crop growth;

[0020] Construct an image classification factor of crops, i.e., an apparent growth factor:

[0021] delta b = delta LCR + delta HGR + delta SR + delta FCI + delta FSI + delta FCQ

[0022] Wherein, delta LCR represents the proportion of crop leaf area to the whole image, delta HGR represents the proportion of healthy green leaf pixels to total leaf pixels, delta SR represents the proportion of diseased pixels to total leaf pixels, delta FCI represents the normalized fruit number, delta FSI represents the normalized average single fruit area, delta FCQ represents the matching degree score based on the expected color of the target mature stage.

[0023] Constructing a spectral classification factor of crops, i.e. a physiological vigor factor:

[0024] δ s = δ HI - δ WDI - δ SI - δ PI

[0025] Wherein, δ HI represents a normalized combination based on a representative chlorophyll index, a biomass index and a photosynthetic activity index, δ WDI represents a water stress condition, δ SI represents a vegetation reflectance index, a structure-insensitive pigment index and a vegetation senescence reflectance index or a combination thereof, and a specific pest-related absorption feature, δ PI represents an index established for a specific pollutant based on a characteristic waveband;

[0026] Using the above classification factor, the features are weighted and fused:

[0027] T = αδ b + βδ s

[0028] Wherein, α and β are weighting coefficients;

[0029] The channel feature T after weighted fusion is subjected to dimension transformation, and the dimension transformation adopts a deep learning method, including but not limited to shape remodeling, dimension permutation and dimension splicing and splitting; the feature tensor after dimension transformation is input into a deep learning model, and the model is trained and evaluated, so as to obtain a deep learning model taking the weighted and fused feature tensor as input and taking the crop vigor category as output, and then the model is used to evaluate the crop vigor.

[0030] Compared with the prior art, the present application has the beneficial effects that: the present application proposes a smart greenhouse management system and an intelligent management method, in particular, the designed crop vigor recognition algorithm can automatically recognize the vigor of crops through the fusion of RGB images and hyperspectral information, and form autonomous decision-making according to the material requirements of crops in different growth cycles, automatically control the management system of the smart greenhouse based on the decision-making results, and perform real-time adjustment, thereby improving the unit area yield of crops in the greenhouse. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 It is a front view of the overall structure of the present application;

[0032] Figure 2 It is a front view of the overall structure of the present application;

[0033] Figure 3 It is an installation schematic diagram of the RGB camera and the temperature sensor of the present application;

[0034] Figure 4 Schematic diagram of the ventilation device of the present application;

[0035] Figure 5 Schematic diagram of the seedling pressing device of the present application;

[0036] Figure 6 Schematic diagram of the water and fertilizer integrated device of the present application;

[0037] Figure 7 Schematic diagram of the temperature control device of the present application;

[0038] Figure 8 Schematic diagram of the humidity sensor of the present application;

[0039] Figure 9 Flow chart of the intelligent management strategy based on crop growth of the present application;

[0040] Figure 10 Flow chart of the intelligent identification algorithm based on crop growth of the present application;

[0041] Wherein, 1, roller shaft; 2, coupling A; 3, stepper motor A; 4, guide shaft; 5, rolling body; 6, stepper motor B; 7, seedling pressing rod; 8, stepper motor C; 9, coupling B; 10, gear transmission group; 11, water valve; 12, rudder A; 13, three-way pipe; 14, nutrient solution control valve; 15, nutrient solution storage tank; 16, rudder B; 17, spray head; 18, RGB camera; 19, support frame; 20, resistance wire; 21, fan; 22, motor; 23, plastic film; 24, transmission gear box; 25, hyperspectral camera; 26, LM35 temperature sensor; 27, humidity sensor. DETAILED DESCRIPTION

[0042] The specific embodiments of the present application will be described in more detail below with reference to the schematic diagrams, and the specific implementation method takes planting tomatoes as an example. The advantages and features of the present application will be clearer according to the following description. It should be noted that the drawings are very simplified and use non-precise proportions, only for the purpose of facilitating and clearly assisting the description of the embodiments of the present application.

[0043] As shown in Figure 1 and Figure 2 , the embodiment of the present application provides a smart greenhouse management system based on crop growth, which comprises a crop growth and environment monitoring system, a control system and a supporting operation facility.

[0044] The crop growth and environment monitoring system is composed of a temperature and humidity detection system and a crop growth identification system.

[0045] As shown in Figure 1 , Figure 2 and Figure 3As shown, the temperature detection system consists of an LM35 temperature sensor 26, which is mounted on a support frame 19. The support frame 19 is located in the middle of the greenhouse and on the ground near one side of the plastic film 23. The LM35 temperature sensor 26 detects the temperature inside the greenhouse, and the measured temperature data is read and stored using the Arduino development board. The humidity detection system consists of a humidity sensor 27, which is inserted into the greenhouse soil to detect the soil moisture. The measured soil moisture data is read and stored using the Arduino development board, which communicates with the control system.

[0046] like Figure 1 As shown, the crop growth recognition system comprises an RGB camera 18, a hyperspectral camera 25, a microprocessor, and its crop growth recognition algorithm. Figure 3 As shown, the RGB camera 18 is mounted on a support frame 19, which is located on the ground near one side of the plastic film 23 in the middle of the greenhouse. It is used to monitor the apparent growth information of the crops in real time. A hyperspectral camera 25 is installed inside the greenhouse, close to the RGB camera 18. The hyperspectral camera 25 collects the spectral information of the crops and transmits the RGB images and spectral information to the microprocessor. The microprocessor uses a crop growth recognition algorithm to determine the crop growth and transmits the results to the supporting operating facilities for appropriate material control.

[0047] The control system consists of a power distribution box, an intelligent terminal controller inside the power distribution box, and its algorithm. The power distribution box is used to supply power to the smart greenhouse management system. The intelligent terminal controller uses an intelligent management strategy based on crop growth to regulate the crop growth environment and required materials in real time.

[0048] The supporting facilities consist of a ventilation device, a seedling pressing device, a temperature control device, and an integrated water and fertilizer device.

[0049] like Figure 2 As shown, the ventilation device is installed at one edge of the greenhouse to achieve ventilation inside the greenhouse; Figure 4As shown, the ventilation device includes a roller shutter shaft 1, a coupling A2, a stepper motor A3, a guide shaft 4, a pair of rolling elements 5, a stepper motor B6, and a plastic film 23. The roller shutter shaft 1 is inserted laterally into the edge of the plastic film 23, enabling the plastic film 23 to be rolled up and opened. One end of the roller shutter shaft 1 is connected to the coupling A2, and the other end of the coupling A2 is connected to the input shaft of the stepper motor A3. The stepper motor A3 drives the roller shutter shaft 1 to rotate, thereby rolling up the plastic film 23 on one side of the greenhouse. The stepper motor A3 is fixed to the connecting plate of the guide shaft 4 and installed perpendicular to one side of the connecting plate. The stepper motor B6 is connected to the pulley groove of a rolling element 5 through a transmission belt, thereby driving the connected rolling element to rotate in the same direction. The rolling element will form an upward frictional force with the guide rod shaft 4 (the guide rod shaft 4 moves downward when the stepper motor B6 rotates in the opposite direction). The other side of the guide rod shaft 4 is pressed against another rolling element, causing the other rolling element to rotate passively, thereby realizing the operation of rolling up or lowering the greenhouse film and promoting air circulation inside the greenhouse.

[0050] like Figure 5 As shown, this is the seedling pressing device of the supporting operating facility. In actual use, there are multiple seedling pressing devices, arranged along each planting row or column. The seedling pressing device includes a pressing rod 7, a stepper motor C8, a coupling B9, and a transmission gear set 10. The transmission gear set 10 is mounted on the transmission gearbox 24. The output end of the stepper motor C8 is connected to one end of the coupling B9 to realize the torque output of the motor. The other end of the coupling B9 is connected to the large gear of the transmission gear set 10 through a short shaft. The large gear meshes with the small gear of the transmission gear set. The small gear is sleeved on the pressing rod 7, thereby transmitting the rotational motion of the stepper motor C8 to the pressing rod 7. The pressing rod 7 provides a lodging force to the tomato seedlings, realizing the pressing operation.

[0051] By employing the above-mentioned technical solution, pressing down the lower part of the tomato seedlings can prevent excessive growth. During the tomato's growth period, pushing the plants in the same direction suppresses apical dominance and better controls excessive growth. During the process of pushing them down, the overall height of the plant may decrease by 20-30 centimeters, both inhibiting excessive growth and reducing the height of the first layer of tomatoes, thus paving the way for later harvesting of the top tomatoes and flowers.

[0052] like Figure 6As shown, the integrated water and fertilizer device is connected to the seedling pressing device, and specifically includes a water valve 11, a steering servo motor A12 and a steering servo motor B16, a three-way pipe 13, a nutrient solution control valve 14, a nutrient solution storage tank 15, and nozzles 17 distributed along the seedling pressing rod. The nutrient solution storage tank 15 is installed at one end of the integrated water and fertilizer device. The nutrient solution can flow along the pipeline to the nutrient solution control valve 14. The valve body of the nutrient solution control valve 14 is connected to the steering servo motor B16. The servo motor B16 automatically controls the opening and closing of the valve body through 0° and 90° steering. If the valve body of the nutrient solution control valve 14 is in the open state, the nutrient solution can flow along the pipeline to the three-way pipe 13. One end of the three-way pipe 13 is connected to the seedling pressing rod 7 equipped with a nozzle 17, which can realize the spraying of nutrient solution. At the same time, the other end of the three-way pipe 13 is connected to the water valve 11. The valve body of the water valve 11 is connected to the steering servo motor A12. The servo motor A12 automatically controls the opening and closing of the valve body of the water valve 11 through 0° and 90° steering. If the water valve 11 is in the open state, tap water can flow through the water pipe on one side of the valve body into the pipeline of the seedling pressing rod 7 equipped with a nozzle 17, thereby realizing the watering operation of the system. Similarly, when water and fertilizer mixing is required, water valve 11 and nutrient solution control valve 14 are opened simultaneously to achieve water and fertilizer mixing and integrated operation.

[0053] like Figure 7 As shown, the temperature control inside the greenhouse is achieved by a temperature control device consisting of an internal temperature sensor and an external heating device. The external heating device includes a resistance wire 20, a fan 21, and a motor 22. The torque output end of the motor 22 is connected to the rotation shaft of the fan 21, and the entire device is installed at the front end of the resistance wire 20. When the ambient temperature collected by the LM35 temperature sensor 26 is lower than the suitable temperature for tomato cultivation, the control system controls the resistance wire 20 to start heating and simultaneously controls the motor 22 to turn on. The motor 22 drives the fan 21 to blow the hot air generated by the resistance wire 20 into the greenhouse, thereby raising the ambient temperature inside the greenhouse. Similarly, when the ambient temperature inside the greenhouse is lower than the suitable temperature for tomato cultivation, the motor 22 turns on, driving the fan 21 to rotate and blow hot air, thereby lowering the temperature inside the greenhouse.

[0054] The soil moisture detection system inside the tomato greenhouse consists of a moisture sensor 27 and its detection system, such as... Figure 8 As shown, after the humidity sensor 27 is inserted into the soil, its output voltage increases as the soil moisture increases. Users can determine the soil moisture level based on the sensor's output voltage value. Specifically, when the output value is between 0% and 300% RH, the soil is dry; when the output value is between 300% and 700% RH, the soil is moist; and when the output value reaches 700% to 950% RH, the soil moisture content is extremely high, almost saturated. Finally, the data is output to the intelligent decision-making terminal.

[0055] It is particularly noteworthy that when the control system is connected with the Arduino mainboard, a filter capacitor is usually added in the circuit to remove noise interference in the signal transmission process and ensure the stability of the collected information; through the pull-up resistor or the pull-down resistor, the sensor output signal is ensured to be within a suitable voltage range, thereby achieving good electrical matching with the Arduino mainboard; at the same time, a protection diode can also be arranged in the circuit to prevent damage to the temperature and humidity sensor and the Arduino mainboard due to abnormal voltage.

[0056] The monitoring of the apparent information growth of the tomatoes is realized by the crop growth recognition system, as shown in Figure 3 , which specifically comprises a support frame 19 fixed on the ground of the greenhouse and an RGB camera 18 installed at the upper end of the support frame 19, which photographs the apparent growth of the tomato seedlings and transmits the image to the control system for intelligent decision-making. By adopting the technical solution, the differences in different growth periods of the tomato seedlings can be detected, so that the system can accurately determine the growth stage of the tomatoes, thereby realizing scientific management of the tomato planting.

[0057] As shown in Figure 9 , the phase-based management strategy process based on the growth of the tomatoes mainly comprises:

[0058] The RGB camera 18 collects the RGB image information of the tomatoes, and the apparent growth state of the tomatoes is recognized through the improved YOLOv8 algorithm, and the recognition result is used to distinguish the growth stages of the tomatoes. According to the planting requirements in the existing tomato facility cultivation technical regulations, after the germination period, the tomato seedlings can be collected by the RGB camera 18, that is, when the tomato true leaf “heart” state is recognized, the tomato enters the seedling growth period, and the control system starts the seedling period management mode; when the system recognizes that the first inflorescence appears flower buds, it indicates that the tomato enters the flowering period, and the control system switches to the management mode of the flowering period based on the recognition result, and after seven days, the control system issues an instruction to start the seedling pressing device and performs the seedling pressing operation. On this basis, when the system recognizes that the tomato fruit setting-young fruit diameter is about 1-2 cm, it indicates that the tomato planting enters the fruit expansion period, and the system starts the management mode of the fruit expansion period. Based on each growth stage, the tomatoes can be monitored and adjusted in real time. The management modes of the seedling growth period, the inflorescence period and the fruit expansion period are all existing tomato planting technologies, which divide the various planting stages of the tomatoes, provide management strategies such as water and fertilizer and seedling pressing, and are used for reference by farmers during planting.

[0059] As shown in Figure 10As shown, in each growth cycle, the intelligent recognition algorithm based on crop growth provides targeted decisions on environmental and material elements based on effective analysis of tomato growth by fusing RGB images and hyperspectral data. The focus of the present application is to effectively integrate the apparent information and physiological information of tomatoes at the algorithm level.

[0060] The apparent growth information of the crops is monitored in real time by the RGB camera 18, including leaf density, fruit size and color, etc., and the spectral information of the crops is obtained by the hyperspectral camera 25, and based on the two types of information, the growth of the crops is autonomously judged by the intelligent recognition algorithm.

[0061] Firstly, the two types of information are preprocessed, including: preprocessing the image data, including image size scaling, scale normalization and dimension transformation, etc.; preprocessing the spectral data, including wavelength and wave number range interception, noise smoothing, baseline and scattering correction and normalization, etc.

[0062] Then, the preprocessed spectral data and RGB images are fused, and a deep learning framework is adopted, including four key modules of feature extraction, feature dimension reduction, feature weighted fusion and strategy mapping.

[0063] Feature extraction includes extracting image features and spectral features of crop growth, image features are obtained by improved YOLOv8 algorithm, and the apparent growth state of the crops is obtained, including color, density, size and color of tomato leaves and tomato fruits, etc.; spectral features are obtained by intercepting effective feature bands of spectral data, analyzing wave peak, wave trough, steep rise and steep decline of the intercepted bands, forming discrimination of tomato growth conditions such as health, water shortage, disease and pest damage and pollution damage, and effectively reducing the dimension of spectral data.

[0064] Combining the RGB image and spectral information of tomato growth, the image and spectral classification factors of tomato are constructed respectively; the constructed image classification factor, i.e. the expression of apparent growth factor, is as follows:

[0065] δ b = δ LCR + δ HGR + δ SR + δ FCI + δ FSI + δ FCQ

[0066] Wherein, δ LCR represents the proportion of tomato leaf area to the whole image obtained by RGB image and improved YOLOv8 algorithm, and high value represents lush; δ HGRThe proportion of healthy green leaf pixels in total leaf pixels, high value represents healthy leaf; δ SR The proportion of diseased pixels (disease spots, insect holes, pollution spots, obvious yellowing, etc.) in total leaf pixels, high value represents serious problems; δ FCI The normalized fruit number (based on image area or fixed area), high value represents more results; δ FSI The normalized average single fruit area, high value represents large fruit; δ FCQ The matching score based on the expected color of the target maturity stage, high value represents good color (uniform maturity).

[0067] The constructed spectral classification factor, i.e. the expression of the physiological vigor factor, is as follows:

[0068] δ s = δ HI - δ WDI - δ SI - δ PI

[0069] Among them, δ HI represents the combination based on the normalized index representing chlorophyll, biomass and photosynthetic activity, etc. acquired by the hyperspectral camera 25, the higher the value, the healthier; δ WDI represents the water stress situation, high value represents serious water shortage; δ SI represents the absorption characteristics based on vegetation reflectance index, structure-insensitive pigment index and vegetation senescence reflectance index, etc. or their combination, and specific diseases and pests may be related, high value represents (disease or stress) serious; δ PI represents the index established for specific pollutants based on characteristic wavebands, high value represents serious pollution.

[0070] Using the above classification factors, the characteristics are weighted and fused, and the weighted expression is as follows:

[0071] T = αδ b + βδ s

[0072] Among them, α and β respectively represent the relative importance of apparent factors and physiological factors in overall vigor evaluation, tomatoes belong to solanaceous crops, and apparent factors have a greater impact on the vigor and yield of tomatoes, therefore, based on planting experience, α = 0.6, β = 0.4.

[0073] The channel feature T after weighted fusion is transformed in dimension, that is, a dimension transformation operation commonly used in deep learning methods is adopted, including but not limited to shape remodeling, dimension permutation and dimension splicing, splitting and the like, the transformed feature tensor is input into a deep learning model, the model is trained and evaluated, and a deep learning model taking the weighted fused feature tensor as input and taking the crop growth class as output is obtained, the crop growth class includes but is not limited to the following cases: in the tomato seedling stage, both the growth of tomato seedlings and the prevention of excessive water growth of branches and trunks are required, if the output result of the deep learning model is the above growth class, the control system maps to form a corresponding "wet and dry" management strategy, so as to adjust the specific operation of the water and fertilizer integrated device. When the tomato seedlings grow to about 30 cm and the stem base diameter is not less than 0.5 cm, the tomato leaves can be detected by the hyperspectral camera 25 at this time, the crop chlorophyll content is analyzed through red light, near-infrared and other wave bands, if the chlorophyll content is significantly increased, the output result of the crop growth class formed accordingly, the control system maps to form a corresponding seedling pressing management strategy, the seedling pressing device starts operation to realize the seedling pressing operation of the tomato seedlings.

[0074] Based on the algorithm, the system can accurately identify the growth class of tomatoes at each stage, dynamically adjust the application ratio of each type of material and the environmental factors of tomato growth through the control system, and promote the growth of tomato seedlings.

[0075] The embodiments are preferred embodiments of the present application, but the present application is not limited to the above embodiments, any obvious improvement, replacement or modification made by those skilled in the art without departing from the essential content of the present application shall fall within the protection scope of the present application.

Claims

1. A smart greenhouse management system based on crop growth, characterized in that, The application relates to a crop growth and environment monitoring system, a control system and a supporting operation facility. The temperature detection system comprises an LM35 temperature sensor (26) arranged at the middle of a greenhouse and close to one side of a plastic film (23) for detecting temperature data in the greenhouse, and an Arduino development board for reading and storing the temperature data. The humidity detection system comprises a humidity sensor (27) inserted into the soil in the greenhouse for detecting humidity data of the soil, and an Arduino development board for reading and storing the humidity data. The RGB camera (18) is arranged at the middle of the greenhouse and close to one side of the plastic film (23) for monitoring the apparent growth information of crops in real time; and the hyperspectral camera (25) is arranged in the greenhouse for collecting spectral information of crops.

2. The crop growth based smart greenhouse management system as claimed in claim 1, wherein, The ventilation device is installed at the edge of one side of the greenhouse for realizing ventilation in the greenhouse; the ventilation device comprises a roller shaft (1), a coupling A (2), a stepping motor A (3), a guide shaft (4), two rolling bodies (5), a stepping motor B (6) and the plastic film (23); the roller shaft (1) is inserted into the edge of the plastic film (23) transversely to realize rolling and opening of the plastic film (23); the roller shaft (1) is connected with the input shaft of the stepping motor A (3) through the coupling A (2), the stepping motor A (3) is fixed on the connecting plate of the guide shaft (4) and is installed vertically on one side of the connecting plate; the stepping motor B (6) is connected with the belt groove of one rolling body (5) through a transmission belt. 3.The crop growth-based smart greenhouse management system according to claim 1, wherein, The seedling pressing device is arranged in multiple and arranged along various planting rows or columns and specifically comprises a seedling pressing rod (7), a stepping motor C (8), a coupling B (9) and a transmission gear set (10); the output end of the stepping motor C (8) is connected with one end of the coupling B (9), the other end of the coupling B (9) is connected with the large gear of the transmission gear set (10) through a short shaft, the large gear is engaged with the small gear of the transmission gear set (10), and the small gear is sleeved on the seedling pressing rod (7).

4. The crop growth based smart greenhouse management system as claimed in claim 1, wherein, ​ 5. The crop growth based smart greenhouse management system as claimed in claim 1, wherein, ​ 6. The crop growth based smart greenhouse management system as claimed in claim 1, wherein, ​ 7. The crop growth based smart greenhouse management system as claimed in claim 6, wherein, The water and fertilizer integrated device is connected with the seedling pressing device, and specifically comprises a water valve (11), a rudder engine A (12), a rudder engine B (16), a three-way pipe (13), a nutrient solution control valve (14), a nutrient solution storage tank (15), and a spray head (17) installed along the seedling pressing rod (7); the nutrient solution storage tank (15) is located at one end of the water and fertilizer integrated device, is connected with the nutrient solution control valve (14) through a pipeline, and the valve body end of the nutrient solution control valve (14) is connected with the rudder engine B (16); the nutrient solution control valve (14) is connected with one end of the three-way pipe (13) through a pipeline, the second port of the three-way pipe (13) is connected with the water valve (11), and the valve body end of the water valve (11) is connected with the rudder engine A (12); the third port of the three-way pipe (13) is connected with the seedling pressing rod (7). 8.The crop growth condition based smart greenhouse management system according to claim 1, wherein, The temperature control device comprises a resistance wire (20), a fan (21) and a motor (22), the torque output end of the motor (22) is connected with the rotating shaft of the fan (21), and the whole is installed at the front end of the resistance wire (20).

9. An intelligent management method based on the crop growth-based intelligent greenhouse management system according to any one of claims 1-8, characterized in that: Based on the RGB image information of crops, the apparent growth state of crops is identified, which is used to distinguish the growth stages of crops, including seedling stage, flowering stage and fruit expansion stage; in each growth cycle, the intelligent recognition algorithm based on crop growth is used to analyze the crop growth by fusing RGB image and hyperspectral data, and the control system dynamically adjusts the application ratio of various materials and the environmental factors of crop growth.

10. The intelligent management method of claim 9, wherein, The intelligent recognition algorithm based on crop growth is used to analyze the crop growth by fusing RGB image and hyperspectral data, specifically as follows: Extract the image features and spectral features of crop growth; Construct the image classification factor of crops, i.e. apparent growth factor: delta b = delta LCR + delta HGR + delta SR + delta FCI + delta FSI + delta FCQ wherein δ LCR represents the proportion of the crop leaf area to the whole image, δ HGR represents the proportion of healthy green leaf pixels to the total leaf pixels, δ SR represents the proportion of diseased pixels to the total leaf pixels, δ FCI represents the normalized fruit number, δ FSI represents the normalized average single fruit area, δ FCQ represents the matching score based on the expected color of the target maturity stage; Construct the spectral classification factor of crops, i.e. physiological growth factor: delta s = delta HI - delta WDI - delta SI - delta PI where δ HI represents a normalized combination based on the representative chlorophyll index, biomass index and photosynthetic activity index, δ WDI represents a water stress condition, δ SI represents a vegetation reflectance index, a structure-insensitive pigment index and a vegetation senescence reflectance index or a combination thereof, and absorption features related to specific diseases and pests, δ PI represents an index established for a specific pollutant; Use the above classification factors to perform weighted fusion of features: T = a δ b + β δ s Wherein, α and β are weighting coefficients; The channel feature T after weighted fusion is subjected to dimension transformation, the dimension transformation adopts a deep learning method, including but not limited to shape remodeling, dimension permutation and dimension splicing and splitting; the feature vector tensor after dimension transformation is input into a deep learning model, the model is trained and evaluated, a deep learning model taking the weighted fusion feature tensor as input and taking the crop growth category as output is obtained, and then the model is used to evaluate the crop growth.

Images