Efficient frame-building cultivation method for improving cultivation commodity of melons

Abstract

The invention discloses an efficient frame-building cultivation method for improving cultivation commodity of melons, and relates to the technical field of agricultural cultivation, the method comprises the following steps: building a modular high-strength arch-type frame body as a support for melon seedlings; a real-time monitoring unit and a drip irrigation pipeline system are integrated on the support frame body; extending the frame body according to the growth condition of the melon seedlings and carrying out vine binding assistance; collecting multi-source data collected by the real-time monitoring unit and carrying out comprehensive analysis on the multi-source data; inputting the analyzed and processed data result into a crop growth model for deduction; executing a water and fertilizer irrigation strategy based on a deduced result and updating the water and fertilizer irrigation strategy in real time; pruning and fruit retaining management is conducted on the melon seedlings according to the growth state of the melon seedling vines; and collecting growth data of melon seedlings to carry out early warning and prevention and control on disease risks. According to the method, the commodity and the cultivation efficiency of melons and fruits are remarkably improved through optimization of the frame body structure, plant regulation and control and field management.

Description

Technical Field

[0001] This invention relates to the field of agricultural cultivation technology, and in particular to an efficient trellis cultivation method for improving the marketability of melons. Background Technology

[0002] Traditional trellising methods for melons often use simple supports made of bamboo poles and wooden sticks, which suffer from structural instability, poor ventilation and light penetration, inconvenient operation, high labor input, easy fruit deformity, and low marketability. Furthermore, traditional methods heavily rely on manual experience, exhibiting significant inefficiency and delays. Trellising, vine tying, and pruning are entirely based on the farmer's visual judgment and manual execution, resulting in extremely high labor intensity, low efficiency, and low standardization. Improper handling easily leads to vine damage, overlap, or uneven light exposure. Crucial decisions regarding pruning and fruit setting lack scientific data support and cannot be dynamically adjusted based on real-time plant growth and future weather conditions, often resulting in unreasonable plant loads—either excessive vegetative growth affecting fruit setting or excessive fruit production leading to smaller, lower-quality fruit. Problems in water and fertilizer management and disease control are particularly prominent.

[0003] Traditional methods of flood irrigation and regular fertilization result in extremely low water and fertilizer utilization rates, leading not only to resource waste but also to root zone salinization or nutrient imbalances. Disease control largely follows a passive "treat only when disease occurs" approach, failing to monitor the canopy microenvironment and plant sub-health status for early warning, often missing the optimal control period. Once a disease outbreak occurs, increased chemical pesticide use is necessary, increasing costs and threatening agricultural safety and the ecological environment. In summary, existing trellised melon cultivation methods have significant shortcomings in structural reliability, management precision, resource utilization efficiency, and system intelligence, urgently requiring a transformation towards a modern model driven by data, intelligent decision-making, and automated execution. Summary of the Invention

[0004] This invention provides an efficient trellis cultivation method to improve the marketability of melons. By optimizing the trellis structure, plant regulation, and field management, it significantly improves the marketability and cultivation efficiency of melons and fruits.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides an efficient trellis cultivation method for improving the marketability of melons, comprising: S1: Construct a modular, high-strength arched frame to support the melon seedlings; S2: Integrates a real-time monitoring unit and drip irrigation pipeline system on the support frame; S3: Extend the trellis and tie the vines as needed based on the growth of the melon seedlings; S4: Collect and comprehensively analyze multi-source data collected by the real-time monitoring unit; S5: Input the analyzed and processed data results into the crop growth model for extrapolation; S6: Execute water and fertilizer irrigation strategies based on the simulation results and update them in real time; S7: Prune and manage the fruit setting of melon seedlings and vines according to their growth status; S8: Collect growth data of melon seedlings to provide early warning and control of disease risks.

[0006] The beneficial effects of the technical solution provided by this invention include at least the following: The method of this invention, through intelligent seedling growth regulation and integrated water and fertilizer management, can improve the uniformity of subsequent fruits, reduce the rate of deformed fruits, and enhance marketability.

[0007] This invention integrates a real-time monitoring unit with a drip irrigation pipeline system, enabling automatic control. With the assistance of a vine-tying gun, work efficiency is greatly improved. Furthermore, the method of pruning and fruit-setting management based on the growth status of melon seedlings and vines is applicable to various melon crops and different scales of facility cultivation, and has strong scalability.

[0008] This invention enhances wind resistance by optimizing the frame structure and provides early warning of pests and diseases based on the growth data of the melon seedlings, thereby reducing the occurrence of pests and diseases and enhancing the resilience of the melon seedlings during cultivation. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0010] Figure 1 This is a flowchart of an efficient trellis cultivation method for improving the marketability of melons, provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of an archway-type scaffolding structure provided in an embodiment of the present invention. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0012] Example An efficient trellis cultivation method to improve the marketability of melons.

[0013] Please refer to Figure 1 This is a flowchart of an efficient trellis cultivation method for improving the marketability of melons, provided in an embodiment of the present invention.

[0014] S1: Construct a modular, high-strength arched frame to support the melon seedlings; S101: High-strength, weather-resistant composite materials are used as the main arch frame. The surface of the frame is coated with light-reflecting material to enhance the photosynthesis of the lower blades. The structure of the main arch frame is optimized by topology optimization algorithm. Based on the optimized three-dimensional model, composite material molds are developed and the arch frame is manufactured by resin injection process. It should be noted that the calculation process of the topology optimization algorithm is as follows: determine various load conditions including static load and dynamic load (static load includes the self-weight of the frame, the weight of the crop, and the snow load; dynamic load includes wind load and operational impact load). Constraints are set according to the load conditions, including that the maximum deformation of the arch frame under the maximum load does not exceed 1 / 200 of its span and the maximum equivalent stress of the arch frame material does not exceed its allowable stress. Based on the load conditions and constraints, an initial three-dimensional design space is created. By dividing the initial three-dimensional design space into finite element meshes and using the variable density method for iterative calculation, the optimal material distribution model of the arch frame is obtained.

[0015] Please see Figure 2 This is a schematic diagram of an archway-type scaffolding structure provided in an embodiment of the present invention.

[0016] S102: The frame is made of UV-resistant and anti-aging flexible binding tape with quick-connect buckles to form a continuous semi-circular arch structure. S103: Multiple telescopic and adjustable horizontal tie rods are installed at the top and shoulders of the frame to form a stable space truss structure, and solar energy storage units are integrated in the horizontal tie rods to provide in-situ energy for the frame.

[0017] It should be noted that the frame height is 1.6 to 2.0 meters, the span is suitable for a standard 6-meter shed, and the arch spacing is 50 to 55 centimeters, forming a continuous archway structure. This structure facilitates ventilation, light transmission, and mechanical operations. Furthermore, the horizontal tie rods installed at the top and shoulders of the frame enhance overall stability and prevent collapse.

[0018] S2: Integrates a real-time monitoring unit and drip irrigation pipeline system on the support frame; S201: Modular integrated cable trays are installed on the frame and real-time monitoring units are deployed within them, including environmental monitoring sensor units, crop body sensor units, and frame health sensor units. It should be noted that the environmental monitoring sensor unit includes temperature and humidity sensors, soil moisture sensors, and light intensity sensors; the crop body sensor unit includes a multispectral imager and a stem micro-change sensor; and the trellis health sensor unit includes wind speed and direction sensors and pipeline pressure and flow sensors.

[0019] The multispectral imager calculates the normalized chlorophyll index by using the reflectance of the red-edge band to invert the growth status, nutrient stress, and biomass of melon seedlings; the stem micro-change sensor uses micron-level precision to measure the diurnal shrinkage and expansion of the stem diameter (the maximum shrinkage point of the stem occurs at noon, and its variation can reflect the water content inside the melon seedling), and inverts the water stress status of the melon seedlings in real time.

[0020] S202: The drip irrigation pipeline system consists of multiple zones, each zone is set up as a single semi-circular arch structure, and each zone is equipped with a solenoid valve to connect the zones.

[0021] It should be noted that the specific implementation steps of the drip irrigation pipeline system are as follows: configure a normally closed solenoid valve for each micro-irrigation zone, and lay secondary pipelines (including a primary main pipeline for conveying water flow, and secondary capillary branch pipelines connecting the main pipeline to the solenoid valves of each zone), install pressure-compensating drippers for each crop in each micro-zone, and connect each solenoid valve to the controller via the Internet of Things.

[0022] S3: Extend the trellis and tie the vines as needed based on the growth of the melon seedlings; S301: Monitor the top view image of the melon seedling through the crop body sensor and identify it using an image recognition algorithm; It should be noted that the image recognition algorithm's recognition process is as follows: Crop body sensors deployed on the top of the cultivation frame acquire the morphology of the shoot tips and the length of internodes of the melon seedlings at 3-day intervals. A threshold segmentation algorithm based on color features is used to separate the green organs of the plant from the background to generate a binary image. By extracting the stem region of the melon seedling image, the growth point, stem axis, and leaf nodes of the seedling are identified. Simultaneously, the pixel distance between two adjacent leaf nodes on the stem is calculated, and the physical length of the internode is calculated based on a pre-defined pixel-physical size conversion relationship. By comparing this with a preset phenological stage determination threshold, the current phenological stage of the plant is determined.

[0023] S302: When the node length between melon seedling stems reaches 2-3cm and the overall stem length reaches 30-35cm, and the vine-wrapping phenomenon begins to appear, an auxiliary frame should be erected according to the growth status of the stems. S303: When the overall length of the melon seedling stem reaches ≥60cm, use a vine-tying gun every 2 days, and use flexible vine-tying tape to wrap around the trellis in an "S" shape to assist in tying the vines and avoid overlapping of the vines.

[0024] It should be noted that the elasticity of the flexible vine-tying tape can extend as the stems thicken, avoiding the constriction problems that may be caused by traditional vine-tying materials, and promoting even light exposure for the stems and leaves of the melon seedlings.

[0025] S4: Collect multi-source data collected by the real-time monitoring unit and conduct comprehensive analysis on it; S401: Based on the point data of the sensor network, generate a continuous three-dimensional data field of "frame - environment - plant" through a three-dimensional space interpolation algorithm. The specific process is as follows: Consider each data point of the sensor as the core of a Gaussian function in three-dimensional space, and dynamically adjust the covariance matrix of each Gaussian function to simulate the spatial blocking effect of the canopy structure on the environmental physical field. Calculate the value of any point in space as the weighted sum of the corresponding Gaussian functions of all neighboring sensor points, thereby generating a three-dimensional continuous visualization map of canopy temperature, humidity, and light distribution; It should be noted that the adjustment steps of the Gaussian function covariance matrix are as follows: Scan and obtain the three-dimensional distribution map of the leaf area index of the melon seedling canopy through a multi-spectral imaging sensor carried by the frame, and use the local leaf area index density at the position of each sensor data point as the main structural parameter; According to the local leaf area index density, calculate the structural attenuation factors in the X, Y, and Z spatial directions through a predefined attenuation function, and calculate the environmental modulation factors in the three spatial directions according to the real-time environmental data. Generate a dynamic covariance matrix by combining the reference covariance matrix with the structural attenuation factors and environmental modulation factors.

[0026] S402: Use a visual encoder to extract high-dimensional feature vectors from the plant image, distill the high-dimensional feature vectors into low-dimensional knowledge vectors through a lightweight knowledge encoder, and at the same time add noise that conforms to the differential privacy mechanism to the knowledge vectors to complete the desensitization of the image data.

[0027] It should be noted that the distillation process of the lightweight knowledge encoder is as follows: Design a lightweight knowledge encoder with a structure of a multi-layer perceptron, whose input layer dimension is d1, output layer dimension is d2, and d2 << d1, and train the lightweight knowledge encoder through knowledge distillation. Its training objective is: Minimize the mean square error loss between the output low-dimensional knowledge vector and the vector reconstructed by the high-dimensional feature vector through an autoencoder, so that the knowledge vector can retain as much key information as possible required to reconstruct the original features; After training, input the high-dimensional feature vector into the lightweight knowledge encoder, and output the low-dimensional knowledge vector.

[0028] S5: Input the data result after analysis and processing into the crop growth model for deduction; S501: By using a mechanism model including a photosynthesis module, a dry matter production and distribution module, and a phenological development module as the model skeleton, and constructing a deep learning network with the data collected by the real-time monitoring unit as input, and using the prediction residual of the mechanism model as the learning objective, a data-driven crop growth model is obtained. By coupling the output of the data-driven correction model as a dynamic correction term with the prediction result of the mechanism model, a hybrid crop growth model driven by both mechanism and data is obtained. It should be noted that during model initialization, the pre-stored melon seedling gene parameter set is called to configure the mixed crop growth model. The melon seedling gene parameter set is a variety-specific growth parameter extracted by machine learning from the historical cultivation big data of this melon seedling variety.

[0029] S502: The data collected by the real-time monitoring unit within the predetermined time window in the past is used as the assimilated observation value. The observation value is compared with the predicted state of the mixed crop growth model. At the same time, the ensemble Kalman filter algorithm is used to back-optimize and adjust the state variables inside the mixed crop growth model so that the model state is close to the real crop state. S503: Customize candidate management strategies for melon seedlings, input them into the updated mixed crop growth model, and simultaneously input the weather forecast data for the next seven days, so that the mixed crop growth model can simulate the growth trajectory and key physiological indicators of melon seedlings under different candidate management strategies and future weather conditions. Candidate management strategies include a set of water regulation strategies, a set of environmental light strategies, and a set of fertilizer formulation strategies; The water management strategy set includes implementing the baseline irrigation amount calculated by the model, implementing irrigation amount increased by 15% to 20% based on the baseline irrigation amount, and implementing mild water stress by reducing the baseline irrigation amount by 10% to 15%. The ambient light strategy set includes activating the shoulder micro-sprinkler system for evaporative cooling when the predicted canopy temperature exceeds the temperature threshold, activating the supplemental lighting system when the predicted future light is insufficient, and applying warm water at 20-25℃ through the drip irrigation system to warm the root zone when the predicted root zone temperature is below the optimal temperature. The fertilizer formulation strategy set includes a root-promoting and seedling-strengthening formula with high phosphorus, medium nitrogen and a small amount of alginic acid; a balanced growth formula with balanced nitrogen, phosphorus and potassium; a fruit-promoting and quality-improving formula with high potassium, low nitrogen and high calcium; and a stress-resistance and recovery formula with high potassium and a small amount of amino acids and silicon.

[0030] It should be noted that the specific process of selecting the optimal strategy for multi-objective optimization is as follows: A comprehensive score is calculated for the derivation results of each candidate strategy, and the candidate strategy with the highest comprehensive score is selected as the optimal strategy to be executed. The comprehensive score is calculated based on a weighted sum of preset optimization objectives, which at least include predicted output, predicted quality indicators, resource utilization efficiency, and risk aversion probability. It should be noted that the key physiological indicator used in the simulation is the "dynamic allocation index of photosynthetic products," which is used to quantify and predict the priority and proportion of photosynthetic products allocated to fruit sink organs in the future.

[0031] S504: Select the optimal strategy from candidate management strategies based on the crop growth projection results, and convert it into agricultural operation instructions for melon seedlings through natural language generation technology.

[0032] S6: Execute water and fertilizer irrigation strategies based on the simulation results and update them in real time; S601: Transform the selected candidate management strategies into a specific sequence of irrigation events, including irrigation start time, duration, precise speed ratio of each fertilizer mother liquor pump, and target irrigation volume; It should be noted that the candidate management strategy sends irrigation event sequence instructions to the drip irrigation pipeline system and the zone solenoid valves through the Industrial Internet of Things protocol, and automatically executes the irrigation events at preset times.

[0033] S602: Construct a real-time feedback network for integrated water and fertilizer management, and trigger dynamic adjustments to the strategy by comparing the feedback data with the model's expectations; It should be noted that the real-time feedback network includes a root zone feedback loop, a plant physiological feedback loop, and a canopy spectral feedback loop: The root zone feedback loop continuously monitors the dynamic changes in the root zone environment after irrigation using environmental monitoring sensors to confirm whether water and fertilizer are accurately delivered and within the expected range; the plant physiological feedback loop uses stem micro-change sensors to monitor the recovery of stem flow rate and maximum daily shrinkage after irrigation, directly determining whether the plant's water stress has been relieved; the canopy spectral feedback loop uses multispectral sensors to monitor changes in normalized vegetation index and photochemical reflectance index within 24–48 hours after irrigation, assessing the plant's nutritional and photosynthetic activity responses.

[0034] S603: Based on the deviation between the real-time strategy and the prediction results, execute the update cycle of the candidate management strategy, including the first-level fine-tuning mechanism and the second-level reconstruction mechanism, and make judgments based on soil moisture deviation and fertilizer EC deviation.

[0035] It should be noted that if the volumetric water content of the target root zone does not reach the preset target within ±10% or the actual EC value of the irrigation solution deviates from the target EC value by more than ±15% within 2 hours after the irrigation ends, a first-level fine-tuning mechanism will be adopted. By activating the PID feedback control algorithm, the duration of the currently executed irrigation event sequence will be automatically fine-tuned for real-time compensation. If the fine-tuning is ineffective within 1 hour, the model will be re-performed and an alarm will be generated through a buzzer.

[0036] S7: Prune and manage the fruit setting of melon seedlings and vines according to their growth status; S701: The crop growth status is identified by the crop body sensor. When a lateral branch is detected to sprout at any node from the 1st to the 12th node and its length exceeds 3 to 5 cm, a pruning and fruit retention prompt message is generated. S702: Administrators use the single-vine pruning method to manage pruning and fruit retention based on prompts, including: removing all lateral branches below 12 leaves on the main vine, retaining 3 fruits on the main vine in the first batch, topping the main vine 40cm after it climbs over the arch, pinching off the tips of lateral branches with 1 leaf left, and retaining 1 fruit on each lateral branch after harvesting, with a total of 3-4 fruits retained.

[0037] It should be noted that all pruning operations are carried out at midday on sunny days to reduce the risk of wound infection.

[0038] S8: Collect growth data of melon seedlings to provide early warning and control of disease risks; S801: Establish a disease knowledge graph for melon seedlings. By using a time-series deep learning model and with the rules of the precursor knowledge graph as an aid, train the melon seedling growth data and environmental signals to learn the nonlinear relationship between them and disease outbreaks, and form a multi-level early warning model. It should be noted that this map correlates historical disease events with multi-source data from 7 to 10 days prior to their occurrence, forming a rule chain such as "continuous wet leaf duration > 6 hours + abnormally high canopy temperature + PRI index decrease of 10% → high probability of downy mildew".

[0039] S802: Calculate the dynamic disease risk index using a multi-level early warning model. The specific steps are as follows: Comprehensive disease risk index = W × environmental driving index + (1-W) × physiological stress index; Wherein, Sigmoid represents the environmental driving index used to map the environmental driving index to the (0,1) interval and reflect the nonlinear saturation characteristics of its influence; W represents the dynamic weight, the value of which depends on the prediction time; Physiological stress index = 0.3 × PRI anomaly + 0.4 × leaf temperature anomaly + 0.3 × specific spectral feature intensity; Environmental driving index = ƒ (wet leaf duration, temperature suitability, rainfall probability). S803: Disease risk is warned and controlled based on the comprehensive disease risk index. If the calculated probability of the comprehensive disease risk index is less than 30% in the next 3 days, it is classified as a blue warning level; if the calculated probability of disease occurrence is between 30% and 70%, and at least one crop stress signal shows a persistent abnormality, it is classified as a yellow warning level; if the calculated probability of disease occurrence is higher than 70%, and multiple signals reach the alarm threshold in the knowledge graph, it is classified as a red warning level.

[0040] Furthermore, it should be noted that the present invention can be provided as a method, apparatus, or computer program product. Therefore, embodiments of the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, embodiments of the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code.

[0041] The embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0042] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing terminal equipment to cause a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0043] It should also be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0044] Finally, it should be noted that the above description represents a preferred embodiment of the present invention. It should be pointed out that although preferred embodiments have been described, those skilled in the art, once they understand the basic inventive concept of the present invention, can make various improvements and modifications without departing from the principles described herein. These improvements and modifications should also be considered within the scope of protection of the present invention. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the embodiments of the present invention.

Claims

1. A highly efficient trellis cultivation method for improving the marketability of cucurbits, characterized in that, include: S1: Construct a modular, high-strength arched frame to support the melon seedlings; S2: Integrates a real-time monitoring unit and drip irrigation pipeline system on the support frame; S3: Extend the trellis and tie the vines as needed based on the growth of the melon seedlings; S4: Collect and comprehensively analyze multi-source data collected by the real-time monitoring unit; S5: Input the analyzed and processed data results into the crop growth model for extrapolation; S6: Execute water and fertilizer irrigation strategies based on the simulation results and update them in real time; S7: Prune and manage the fruit setting of melon seedlings and vines according to their growth status; S8: Collect growth data of melon seedlings to provide early warning and control of disease risks.

2. The efficient trellis cultivation method for improving the marketability of cucurbits as described in claim 1, characterized in that, The S1 structure is a modular, high-strength arched frame that supports the melon seedlings, wherein: S101: High-strength, weather-resistant composite materials are used as the main arch frame. The surface of the frame is coated with light-reflecting material to enhance the photosynthesis of the lower blades. The structure of the main arch frame is optimized by topology optimization algorithm. Based on the optimized three-dimensional model, composite material molds are developed and the arch frame is manufactured by resin injection process. S102: The frame is made of UV-resistant and anti-aging flexible binding tape with quick-connect buckles to form a continuous semi-circular arch structure. S103: Multiple telescopic and adjustable horizontal tie rods are installed at the top and shoulders of the frame to form a stable space truss structure, and solar energy storage units are integrated in the horizontal tie rods to provide in-situ energy for the frame.

3. The efficient trellis cultivation method for improving the marketability of cucurbits as described in claim 1, characterized in that, The S2 integrates a real-time monitoring unit and a drip irrigation pipeline system on the support frame, wherein: S201: Modular integrated cable trays are installed on the frame and real-time monitoring units are deployed within them, including environmental monitoring sensor units, crop body sensor units, and frame health sensor units. S202: The drip irrigation pipeline system consists of multiple zones, each zone is set up as a single semi-circular arch structure, and each zone is equipped with a solenoid valve to connect the zones.

4. The efficient trellis cultivation method for improving the marketability of cucurbits as described in claim 1, characterized in that, S3 extends the trellis and provides vine tying assistance based on the growth of the melon seedlings, wherein: S301: Monitor the top view image of the melon seedling through the crop body sensor and identify it using an image recognition algorithm; S302: When the node length between melon seedling stems reaches 2-3cm and the overall stem length reaches 30-35cm, and the vine-wrapping phenomenon begins to appear, an auxiliary frame should be erected according to the growth status of the stems. S303: When the overall length of the melon seedling stem reaches ≥60cm, use a vine-tying gun every 2 days, and use flexible vine-tying tape to wrap the vines around the trellis in an "S" shape to assist in tying the vines and avoid overlapping of the vines.

5. The efficient trellis cultivation method for improving the marketability of cucurbits as described in claim 1, characterized in that, The S4 collects and comprehensively analyzes multi-source data gathered by the real-time monitoring unit, including: S401: Based on point data from sensor networks, a continuous three-dimensional data field of "framework-environment-plant" is generated through a three-dimensional spatial interpolation algorithm. The specific process is as follows: each sensor data point is regarded as a Gaussian function core in three-dimensional space, and the covariance matrix of each Gaussian function is dynamically adjusted to simulate the spatial barrier effect of the canopy structure on the physical field of the environment. By calculating the value of any point in space as the weighted sum of the Gaussian functions corresponding to all neighboring sensor points, a three-dimensional continuous visualization map of canopy temperature, humidity and light distribution is generated. S402: A visual encoder is used to extract high-dimensional feature vectors from plant images. A lightweight knowledge encoder is used to distill the high-dimensional feature vectors into low-dimensional knowledge vectors. At the same time, noise that conforms to the differential privacy mechanism is added to the knowledge vectors to complete the desensitization of image data.

6. The efficient trellis cultivation method for improving the marketability of cucurbits as described in claim 1, characterized in that, S5 inputs the analyzed and processed data results into the crop growth model for extrapolation, wherein: S501: By using a mechanism model including a photosynthesis module, a dry matter production and distribution module, and a phenological development module as the model skeleton, and constructing a deep learning network with the data collected by the real-time monitoring unit as input, and using the prediction residual of the mechanism model as the learning objective, a data-driven crop growth model is obtained. By coupling the output of the data-driven correction model as a dynamic correction term with the prediction result of the mechanism model, a hybrid crop growth model driven by both mechanism and data is obtained. S502: The data collected by the real-time monitoring unit within the predetermined time window in the past is used as the assimilated observation value. The observation value is compared with the predicted state of the mixed crop growth model. At the same time, the ensemble Kalman filter algorithm is used to back-optimize and adjust the state variables inside the mixed crop growth model so that the model state is close to the real crop state. S503: Customize candidate management strategies for melon seedlings, input them into the updated mixed crop growth model, and simultaneously input the weather forecast data for the next seven days, so that the mixed crop growth model can simulate the growth trajectory and key physiological indicators of melon seedlings under different candidate management strategies and future weather conditions. S504: Select the optimal strategy from candidate management strategies based on the crop growth projection results, and convert it into agricultural operation instructions for melon seedlings through natural language generation technology.

7. The efficient trellis cultivation method for improving the marketability of cucurbits as described in claim 6, characterized in that, in: The candidate management strategies include a set of water regulation strategies, a set of ambient light strategies, and a set of fertilizer formulation strategies. The water regulation strategy set includes implementing the baseline irrigation amount calculated by the model, implementing irrigation amount increased by 15% to 20% based on the baseline irrigation amount, and implementing mild water stress by reducing the baseline irrigation amount by 10% to 15%. The ambient light strategy set includes activating the shoulder micro-spraying system for evaporative cooling when the predicted canopy temperature exceeds the temperature threshold, activating the supplemental lighting system when the predicted future light is insufficient, and applying warm water at 20-25°C through the drip irrigation system to warm the root zone when the predicted root zone temperature is below the optimal temperature. The fertilizer formulation strategy set includes a root-promoting and seedling-strengthening formula with high phosphorus, medium nitrogen and a small amount of alginic acid, a balanced growth formula with balanced nitrogen, phosphorus and potassium, a fruit-promoting and quality-improving formula with high potassium, low nitrogen and high calcium, and a stress-resistance and recovery formula with high potassium and a small amount of amino acids and silicon.

8. The efficient trellis cultivation method for improving the marketability of cucurbits as described in claim 1, characterized in that, S6 executes and updates the water and fertilizer irrigation strategy based on the simulation results, wherein: S601: Transform the selected candidate management strategies into a specific sequence of irrigation events, including irrigation start time, duration, precise speed ratio of each fertilizer mother liquor pump, and target irrigation volume; S602: Construct a real-time feedback network for integrated water and fertilizer management, and trigger dynamic adjustments to the strategy by comparing the feedback data with the model's expectations; S603: Based on the deviation between the real-time strategy and the prediction results, execute the update cycle of the candidate management strategy, including the first-level fine-tuning mechanism and the second-level reconstruction mechanism, and make judgments based on soil moisture deviation and fertilizer EC deviation.

9. The efficient trellis cultivation method for improving the marketability of cucurbits as described in claim 1, characterized in that, S7 involves pruning and fruit setting management based on the growth status of the melon seedlings and vines, including: S701: The crop growth status is identified by the crop body sensor. When a lateral branch is detected to sprout at any node from the 1st to the 12th node and its length exceeds 3 to 5 cm, a pruning and fruit retention prompt message is generated. S702: Administrators use the single-vine pruning method to manage pruning and fruit retention based on prompts, including: removing all lateral branches below 12 leaves on the main vine, retaining 3 fruits on the main vine in the first batch, topping the main vine 40cm after it climbs over the arch, pinching off the tips of lateral branches with 1 leaf left, and retaining 1 fruit on each lateral branch after harvesting, with a total of 3-4 fruits retained.

10. The efficient trellis cultivation method for improving the marketability of cucurbits as described in claim 1, characterized in that, The S8 collects growth data of melon seedlings for early warning and prevention of disease risks, including: S801: Establish a disease knowledge graph for melon seedlings. By using a time-series deep learning model and with the rules of the precursor knowledge graph as an aid, train the melon seedling growth data and environmental signals to learn the nonlinear relationship between them and disease outbreaks, and form a multi-level early warning model. S802: Calculate the dynamic disease risk index using a multi-level early warning model. The specific steps are as follows: Comprehensive disease risk index = W × environmental driving index + (1-W) × physiological stress index; Wherein, Sigmoid represents the environmental driving index used to map the environmental driving index to the (0,1) interval and reflect the nonlinear saturation characteristics of its influence; W represents the dynamic weight, the value of which depends on the prediction time; Physiological stress index = 0.3 × PRI anomaly + 0.4 × leaf surface temperature anomaly + 0.3 × specific spectral feature intensity; Environmental driving index = ƒ (wet leaf duration, temperature suitability, rainfall probability). S803: Disease risk is warned and controlled based on the comprehensive disease risk index. If the calculated probability of the comprehensive disease risk index is less than 30% in the next 3 days, it is classified as a blue warning level; if the calculated probability of disease occurrence is between 30% and 70%, and at least one crop stress signal shows a persistent abnormality, it is classified as a yellow warning level; if the calculated probability of disease occurrence is higher than 70%, and multiple signals reach the alarm threshold in the knowledge graph, it is classified as a red warning level.

Images