A comprehensive control method for garlic germplasm innovation and efficient breeding process

Abstract

The present application relates to the garlic breeding field, and discloses a kind of garlic germplasm innovation and the comprehensive control method of efficient breeding process, by real-time collection and store the environment, growth and germplasm data of plant;Data is extracted according to preset time window to construct environment and plant state matrix;Based on the matrix, hydration steady state index, nutrient hunger warning coefficient and biological safety risk index are calculated in turn, and basic maintenance control, nutrient balance control or safety blocking reinspection operation is dynamically triggered and executed, to guarantee the safety of soilless habitat and space-time supply-demand matching;Early screening coefficient is calculated by plant matrix, and those who are not eliminated are used as penalty weight coefficient, and weighted analysis is carried out by combining phenotype and expected genotype matching degree to obtain expression excellence index;The method solves the problem of early disadvantage covering and excellent formula difficult to reuse, triggers exclusive formula solidification instruction when the index meets the standard, realizes the high-dimensional mapping of excellent environment track and genotype, and guarantees the genetic repeatability of breeding experience.

Description

Technical Field

[0001] This invention relates to the field of garlic breeding technology, and more specifically to a comprehensive control method for garlic germplasm innovation and efficient breeding process. Background Technology

[0002] As an important vegetable and medicinal crop, garlic's germplasm innovation and efficient breeding highly depend on the precise coupling of plant growth and development with external environmental factors. Current garlic breeding processes increasingly rely on artificially controlled habitats, especially aeroponics, using sensors for temperature, humidity, light, water, and fertilizer to collect and store environmental data in real time. Based on set empirical thresholds or simple time-series control logic, these processes perform basic habitat stabilization and regulation, nutrient dynamic balance control, and biosafety blocking of circulating fluids and aeroponic delivery habitats. Furthermore, existing technologies periodically extract morphological and physiological growth data from plants to construct a data matrix representing their current state, using a static control approach to assist in the dynamic monitoring of plant phenotypic development and routine breeding management.

[0003] However, existing technologies for controlling the soilless breeding process of garlic still have the following drawbacks:

[0004] Firstly, existing technologies mostly adopt environmental regulation strategies based on static empirical thresholds or single-factor lag, which fail to dynamically map nutrient supply in soilless habitats to the specific developmental rhythm of plants. This leads to a disconnect between water and fertilizer intervention and the actual spatiotemporal needs of garlic at different growth stages, which can easily cause periodic nutrient hunger or redundant supply.

[0005] Secondly, in closed soilless habitats involving circulating fluids and atomized delivery, existing technologies typically rely only on periodic, timed disinfection, lacking real-time quantitative assessment and blocking mechanisms for fluid biosafety risks. This poses a delayed safety hazard where microbial contamination can directly penetrate the delivery channel and infect the root system.

[0006] Third, existing technologies only conduct extensive morphological inferiority elimination in the early stages of breeding, without introducing early developmental disadvantages as a penalty constraint into the later phenotypic evaluation system. This results in the final evaluation results masking the true genetic expression deviations caused by poor early foundations and forced ripening by the environment in the later stages.

[0007] Fourth, existing technologies only record phenotypic data and passively respond to the current environment. They lack the means to perform high-dimensional feature mapping and modeling of the three elements: "standard phenotype - environmental time trajectory - genotype". This results in the inability to transform accidentally obtained excellent formulas into digital breeding models that can be reused from the same source, and there is a risk of irreproducible loss of breeding experience. Summary of the Invention

[0008] In order to overcome the above-mentioned defects of the prior art, the present invention provides a comprehensive control method for garlic germplasm innovation and efficient breeding process, so as to solve the problems existing in the background art.

[0009] This invention provides the following technical solution: a comprehensive control method for garlic germplasm innovation and efficient breeding process, comprising:

[0010] S1: Real-time recording and collection of growth environment data, plant growth data, and germplasm resource data during the growth process of garlic plants, and transmission of the collected growth environment data, plant growth data, and germplasm resource data to the data storage unit for storage.

[0011] S2: Extract the current growth environment data and plant growth data from the data storage unit according to the preset time window, and construct the current environmental state matrix and the current plant state matrix.

[0012] S3: Based on the current environmental state matrix, analyze the dynamic changes of basic environmental survival parameters and calculate the hydration stability index. Determine whether the basic habitat stability control command is triggered. If so, perform the corresponding water replenishment, acidification adjustment or cooling operation and return to S2; otherwise, proceed to S4.

[0013] S4: Based on the current environmental state matrix and the current plant state matrix, combined with germplasm resource data, analyze the spatiotemporal supply and demand matching of current habitat nutrient supply and plant development rhythm, obtain the nutrient hunger warning coefficient, and determine whether the nutrient dynamic balance regulation command is triggered. If so, automatically execute the metering and stirring operation and then proceed to S5; otherwise, directly proceed to S5.

[0014] S5: Based on the current environmental state matrix, analyze the comprehensive risk of biosafety of circulating fluid and atomized delivery habitat, obtain the comprehensive biosafety risk index, determine whether the biosafety blocking control command is triggered, if so, control the atomized delivery channel to perform sterilization and purification operation first, and return to re-execute S5 for re-inspection, otherwise, release the safety blocking, control the atomized delivery channel to directly perform root mist operation, and enter S6.

[0015] S6: By analyzing the plant state matrix at the current moment, the early screening coefficient is calculated to determine whether the inferior elimination instruction is triggered. If so, the breeding control process of the current plant is terminated; otherwise, the candidate qualification is retained, and the early screening coefficient is transmitted to S7 as a penalty weight coefficient.

[0016] S7: Based on the current plant state matrix, germplasm resource data, and penalty weight coefficient, a weighted analysis is performed on the degree of matching expression between the actual phenotypic development of the plant and the expected genotype to obtain an expression excellence index, which is used to determine whether the exclusive formula solidification instruction is triggered.

[0017] Preferably, step S1 involves deploying a multimodal sensor array in the nutrient solution circulation pipeline and root mist habitat of the soilless cultivation system. Various types of sensors within the array are used to collect real-time, high-frequency growth environment data, including liquid level, conductivity, pH value, nutrient solution temperature, and root zone temperature and humidity. A machine vision device deployed on the cultivation rack captures garlic canopy images within a preset time window, and edge computing is used to extract plant growth data, including plant height, pseudostem diameter, and bulb enlargement morphology. Simultaneously, based on the varietal traceability identifier at the time of garlic transplantation, the expected genotype index corresponding to the varietal in a pre-set database is statically mounted as germplasm resource data.

[0018] After being encapsulated via an IoT gateway protocol, growth environment data, plant growth data, and germplasm resource data are aggregated and transmitted to the data storage unit in real time.

[0019] Preferably, S2 uses a preset time window as the time slice boundary, performs timestamp synchronization and mean aggregation on the high-frequency growth environment data within the time window, maps liquid level, conductivity, pH value, nutrient solution temperature and root zone temperature and humidity into the elements of each dimension of the environmental feature vector, and constructs a dynamic environmental state matrix that updates over time.

[0020] Meanwhile, the plant height, pseudostem diameter, and equivalent diameter or projected area values ​​representing the swelling morphology of garlic bulbs extracted by edge computing are normalized in terms of dimensions, arranged in order to construct a phenotypic feature vector, and spatially aligned with the statically mounted expected genotype indicators to generate the plant state matrix at the current moment.

[0021] Preferably, step S3 extracts the liquid level difference and conductivity difference between the current time slice and the adjacent previous time slice from the environmental state matrix at the current moment, and calculates the dynamic water and fertilizer consumption rate, which is used to characterize the intensity of root absorption; at the same time, it extracts the normalized data of pH value and root zone temperature in the current time slice, calculates the absolute deviation of each from the median of the current plant's expected suitable growth range, and performs mean aggregation to obtain the root zone environmental stress degree;

[0022] When the hydration stability index is lower than the preset stability threshold, the basic habitat stability control command is triggered. By comparing the weight ratio of the dynamic water and fertilizer consumption rate and the root zone environmental stress, the specific water replenishment, acidification or cooling operation is matched and locked. After the operation is completed, the process returns to S2 to reconstruct the environmental state matrix. Otherwise, the process directly enters the analysis process of S4.

[0023] Preferably, in step S4, the change rate of the equivalent diameter of the garlic bulb at the current moment is extracted from the plant state matrix at the current moment, and combined with the expected expansion rate genotype index of the current variety in the germplasm resource data at this growth stage, the morphological dynamic demand index is calculated. The morphological dynamic demand index is used to characterize the plant development rhythm. At the same time, the current electrical conductivity mapping value is extracted from the environmental state matrix at the current moment as the base number of habitat basic nutrient supply.

[0024] Multiply the dynamic demand index by a preset time window step size to calculate the expected nutrient consumption increment for the next time slice. Subtract the expected nutrient consumption increment from the basic habitat nutrient supply base to calculate the nutrient supply and demand air conditioning difference value. Perform nonlinear mapping dimensionality reduction on the nutrient supply and demand air conditioning difference value to generate the nutrient hunger early warning coefficient.

[0025] When the nutrient hunger warning coefficient is lower than the preset hunger trigger threshold, the dynamic nutrient balance control command is triggered, and the execution equipment is controlled to automatically perform precise metering and stirring operations according to the expected increase in nutrient consumption. After the operation is completed, the process enters the S5 analysis process. Otherwise, the process directly enters the S5 analysis process.

[0026] Preferably, in step S5, the current turbidity change rate and microbial optical density mapping value of the circulating fluid are extracted from the environmental state matrix at the current moment, and a fluid biological risk index is calculated. The fluid biological risk index is used to characterize the activity of bacterial colony proliferation inside the fluid. At the same time, the current concentration of particulate matter and the dynamic increment of aerosol suspension in the atomized delivery habitat are extracted, and an atomized habitat risk index is calculated. The atomized habitat risk index is used to characterize the space environment impurity interception requirements.

[0027] The fluid biological risk index and the atomization habitat risk index are weighted and coupled to generate a comprehensive biosafety risk index. When the comprehensive biosafety risk index is higher than the preset safety blocking threshold, a biosafety blocking control command is triggered. The atomization delivery channel is controlled to perform sterilization operations to eliminate biological contamination sources. After the operation is completed, the process is forcibly returned to S5 for re-inspection and evaluation. If the re-inspection and evaluation result shows that the comprehensive biosafety risk index is not higher than the preset safety blocking threshold, it indicates that the atomization delivery channel has reached a safe state. The channel is then controlled to perform root mist operations and continues to enter the analysis process of S6.

[0028] Preferably, step S6 extracts the numerical sequence of equivalent diameter or projected area representing the swelling morphology of garlic bulb from the phenotypic feature vectors in multiple consecutive time slices based on the plant state matrix at the current time. Based on this numerical sequence, the actual phenotypic development process is deduced, and the actual phenotypic development process is compared with the expected genotype index statically attached in the plant state matrix at the current time. The degree of deviation between the two is analyzed, and the early screening coefficient is calculated.

[0029] When the early screening coefficient exceeds the preset inferior elimination threshold, an inferior elimination instruction is triggered, the germplasm level of the plant is downgraded to the conventional commercial level, the allocation of high-level environmental regulation resources to the plant is cut off, and the breeding control cycle of the plant is forcibly terminated; otherwise, it is determined that the inferior elimination instruction has not been triggered, its germplasm candidate status is retained, and the early screening coefficient is positively amplified and mapped to be converted into the penalty weight coefficient for environmental formulation assessment, and the analysis process of S7 continues.

[0030] Preferably, in step S7, the phenotypic feature vector, after dimension normalization, is extracted from the plant state matrix at the current moment. This vector is then subtracted item by item from the expected genotype index aligned within the matrix to obtain a multidimensional basic deviation vector. For the deviation component in this multidimensional basic deviation vector that represents the bulb enlargement morphology, a penalty weight coefficient is introduced for product weighting to amplify the early developmental disadvantages in spatial coordinates. Subsequently, the weighted multidimensional basic deviation vector is used as the deviation coordinate point in space. The Euclidean distance between this deviation coordinate point and the baseline coordinate point corresponding to the expected genotype index is calculated. The Euclidean distance is then reciprocally mapped to generate an expression superiority index.

[0031] When the expression superiority index is greater than or equal to the preset solidification threshold, a special formula solidification instruction is triggered. The complete temporal trajectory of the environmental feature vector of the plant, dynamically updated by continuous time slices within the current control period, is extracted in reverse along the time axis. The temporal trajectory of the environmental feature vector is used as the input independent variable, and the expression superiority index is used as the output dependent variable label for feature mapping and binding. A digital breeding model of the garlic germplasm is generated and persistently stored for direct reuse of homologous germplasm. Otherwise, it is determined that the current formula does not have genetic reproducibility, and the current evaluation process is exited and returned to S2 to wait for the next time window to continue dynamic monitoring and evaluation.

[0032] The technical effects and advantages of this invention are as follows:

[0033] (1) By introducing a spatiotemporal supply and demand matching analysis mechanism that combines environmental state matrix and germplasm resource data to calculate the nutrient hunger warning coefficient and trigger dynamic regulation, it is beneficial to accurately map the nutrient supply in soilless habitat to the specific development rhythm of the plant, eliminate the disconnect between water and fertilizer intervention and actual spatiotemporal demand, and avoid staged nutrient hunger or redundant supply.

[0034] (2) By conducting real-time comprehensive risk analysis of circulating fluid and atomized delivery habitat to obtain the comprehensive biosafety risk index, and returning to re-execute the re-inspection after triggering the blocking command, it is beneficial to replace the traditional timed disinfection, realize the real-time quantitative assessment and closed-loop blocking of fluid biosafety risk, and eliminate the delayed safety hazard of microbial contamination penetrating the delivery channel and infecting the root system.

[0035] (3) By calculating the early screening coefficient in the early screening stage and using it as the penalty weight coefficient to directly enter the later phenotypic evaluation for weighted analysis, it is beneficial to break the state of separation between early elimination and later evaluation, apply real constraints to plants with poor foundation in the early stage and forced ripening by the environment in the later stage, and prevent the final evaluation results from masking the real genetic expression bias.

[0036] (4) By triggering the generation of a special formula solidification instruction containing plant identity and high-order parameter locking status after the standard is determined, and extracting the time-series trajectory to generate a breeding model, it is beneficial to realize the high-dimensional feature mapping and modeling of the three elements of "standard phenotype-environmental time-series trajectory-genotype", transforming the accidentally obtained excellent formula into a digital asset that can be reused in the same source, and preventing the loss of breeding experience that cannot be repeated. Attached Figure Description

[0037] Figure 1 This is a diagram illustrating the method steps of the present invention.

[0038] Figure 2 This is a system structure block diagram of the present invention. Detailed Implementation

[0039] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. In addition, the forms of the various structures described in the following embodiments are merely illustrative. The integrated control method for garlic germplasm innovation and efficient breeding process involved in the present invention is not limited to the structures described in the following embodiments. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] like Figure 1 The embodiment shown provides a comprehensive control method for garlic germplasm innovation and efficient breeding processes, including:

[0041] S1: Real-time recording and collection of growth environment data, plant growth data, and germplasm resource data during the growth process of garlic plants, and transmission of the collected growth environment data, plant growth data, and germplasm resource data to the data storage unit for storage.

[0042] In this embodiment, step S1 involves deploying a multimodal sensor array in the nutrient solution circulation pipeline and root mist habitat of the soilless cultivation system. Various types of sensors within the array are used to collect real-time, high-frequency growth environment data, including liquid level, conductivity, pH value, nutrient solution temperature, and root zone temperature and humidity. A machine vision device deployed on the cultivation rack captures garlic canopy images within a preset time window, and edge computing is used to extract plant growth data, including plant height, pseudostem diameter, and bulb enlargement morphology. Simultaneously, based on the varietal traceability identifier at the time of garlic transplantation, the expected genotype index corresponding to the strain in a pre-set database is statically mounted as germplasm resource data.

[0043] After being encapsulated via an IoT gateway protocol, growth environment data, plant growth data, and germplasm resource data are aggregated and transmitted to the data storage unit in real time.

[0044] It should be specifically noted that the high-frequency acquisition of growth environment data involves: immersion-mounting high-precision level sensors, conductivity (EC) sensors, and pH sensors in the nutrient solution storage tank, main supply pipeline, and return pipeline of the hydroponic system; and suspending temperature and humidity probes in the suspended root area below the root mist nozzles. These sensors constitute a "multimodal sensor array," with each sensor transmitting the collected analog electrical signals at a high sampling rate of 1 Hz to the environmental acquisition control cabinet (such as a lower-level computer based on an STM32 chip), converting them into digital quantities of level (accurate to 0.1 mm), EC value (accurate to 0.01 mS / cm), pH value (accurate to 0.01), and root zone temperature and humidity.

[0045] The low-frequency visual extraction of plant growth data is specifically performed as follows: an industrial-grade area array camera (e.g., a 5-megapixel RGB camera with a fixed-focus lens) is fixedly installed 2 meters in front of the cultivation rack, and a uniform LED strip for supplemental lighting is configured. The system is set to trigger image capture every 2 hours. The captured images are not directly uploaded to the cloud, but are transmitted to an edge computing box (e.g., equipped with an NVIDIA Jetson Nano module) deployed in a local rack. Using a pre-trained lightweight YOLOv8 object detection model and a U-Net segmentation model, the edge contour of the garlic pseudostem and the maximum bounding rectangle of the garlic bulb are directly identified from the image locally. Thus, the "plant height (pixel distance converted to physical distance in cm)" and "pseudostem thickness" are directly calculated at the edge, and the extracted pure numerical results (not images) are packaged into JSON format data.

[0046] The static mounting of germplasm resource data specifically involves the operator scanning a QR code (variety traceability identifier, such as "purple-skinned garlic A line from a certain region") on the seedling tray when transplanting virus-free garlic seedlings. Upon receiving this identifier, the system backend automatically retrieves the "expected genotype indicators" for that line from the local MySQL database (e.g., expected maturity period of 90 days, expected bulb diameter ≥ 5.5cm, expected number of cloves 6-8, etc.), which are then mounted in memory as baseline constants for subsequent comparisons and calculations of this batch of garlic.

[0047] The high-frequency environmental data, low-frequency phenotypic JSON data, and statically mounted germplasm baseline data are all sent to an IoT edge gateway. The gateway uses the MQTT communication protocol to timestamp and tag each type of data, and then packages and publishes it to the MQTTBroker message server in the cloud via Wi-Fi / 5G network. After subscribing to these topics, the cloud backend service program associates the environmental data, plant data, and germplasm identifiers, and finally writes them in batches to a time-series database (such as InfluxDB) for persistent storage.

[0048] The "growth environment data" recorded and collected in real time in S1 has a hierarchical feature in the underlying data acquisition architecture: it not only includes conventional apparent parameters such as macroscopic liquid level, conductivity, pH value, nutrient solution temperature, and root zone temperature and humidity, but also, in order to support subsequent habitat safety assessment, the underlying hardware synchronously collects high-frequency fluctuation sequences of light transmittance from the nutrient solution return tube and high-frequency fluctuation sequences of scattered light from the root zone airborne particulate sensor; these underlying high-frequency sequences, as implicit environmental features, are transmitted synchronously with the conventional apparent parameters along with timestamps and stored in the data storage unit.

[0049] S2: Extract the current growth environment data and plant growth data from the data storage unit according to the preset time window, and construct the current environmental state matrix and the current plant state matrix.

[0050] In this embodiment, S2 uses a preset time window as the boundary of the time slice, performs timestamp synchronization and mean aggregation on the high-frequency growth environment data within the time window, and maps the liquid level, conductivity, pH value, nutrient solution temperature and root zone temperature and humidity into the elements of each dimension of the environmental feature vector, and constructs the current environmental state matrix that is dynamically updated over time.

[0051] Meanwhile, the plant height, pseudostem diameter, and equivalent diameter or projected area values ​​representing the swelling morphology of garlic bulbs extracted by edge computing are normalized in terms of dimensions, arranged in order to construct a phenotypic feature vector, and spatially aligned with the statically mounted expected genotype indicators to generate the plant state matrix at the current moment.

[0052] It should be specifically explained that the construction process of the environmental state matrix is ​​as follows: the preset time window is set to 5 minutes (i.e., 300 seconds). Since the conductivity sensor, pH sensor, etc. in the hydroponics system collect data independently at a high frequency of 1 time / second (1Hz), there is a time difference of milliseconds. The system uses this 5-minute time slice boundary and pulls all high-frequency environmental data generated within these 300 seconds into the same time axis queue.

[0053] Taking electrical conductivity (EC) data as an example, the system extracts 300 EC values ​​(e.g., 2.11, 2.12, 2.10, 2.13…2.12 mS / cm) within a 300-second time window. By summing these values ​​and dividing by the number of data points, the mean EC value (e.g., 2.115 mS / cm) is calculated for this time window. This mean aggregation calculation is performed on liquid level, pH value, nutrient solution temperature, and root zone temperature and humidity. Because the five aggregated data have different physical dimensions (e.g., liquid level in cm, temperature in °C, EC in mS / cm), they cannot be directly used for subsequent matrix analysis and require further dimensional normalization. The system has preset safe production extreme value ranges for each environmental parameter; for example, the extreme value range for liquid level is set to 0-50 cm, and the extreme value range for temperature is set to 10-40 °C. The system uses a minimum-maximum normalization formula. ,in This represents the dimensionless pure value of the current parameter after normalization. This represents the actual average value of the parameters calculated through mean aggregation within the current time window (e.g., average liquid level 25cm, average temperature 22℃). This indicates the minimum physical quantity limit of the parameter preset by the system under normal living conditions (e.g., the lower limit of liquid level is set to 0cm, and the lower limit of temperature is set to 10℃). This represents the maximum physical quantity limit of this parameter preset by the system under normal conditions (e.g., the upper limit of liquid level is set to 50cm, and the upper limit of temperature is set to 40℃). Assuming the average liquid level obtained from polymerization within the current 5-minute window is 25cm and the average temperature is 22℃, the normalized liquid level value calculated according to the formula is... Temperature normalized value Similarly, the mean values ​​of the other three parameters are converted into dimensionless pure numerical values ​​between 0 and 1. Finally, these five processed dimensionless numerical values ​​are arranged in a fixed order (e.g., [0.5, 0.6, 0.7, 0.4, 0.8]) to generate a standardized environmental feature vector for that time slice. This vector is continuously added over time to construct a dynamically updated environmental state matrix for the current moment.

[0054] The process of constructing the plant state matrix is ​​as follows:

[0055] Because the physical dimensions of the plant data extracted by machine vision edge computing differ (plant height in cm, pseudostem diameter in mm, and projected area in cm²), it is necessary to refer to the normalization logic of the above environmental data to convert the original measured values ​​into dimensionless pure numerical values ​​to construct the phenotypic feature vector. Specifically, the system presets the theoretical extreme values ​​for the entire growth period of garlic: the maximum plant height is 80cm, the maximum pseudostem diameter is 30mm, and the maximum projected area is 100cm² (the minimum values ​​are all preset to 0). Assuming that the original data extracted by edge computing in the current time window is: plant height 35.5cm, pseudostem diameter 12mm, and projected area 45.2cm², after substituting into the normalization formula with the same logic, the corresponding normalized values ​​are 0.443 for plant height, 0.400 for pseudostem diameter, and 0.452 for projected area. The calculated dimensionless pure numbers are arranged in order to construct the phenotypic feature vector. Subsequently, the expected genotype index of this strain, which is statically mounted in S1, is invoked (the system has pre-converted it into a standard vector using the same normalization logic, such as the expected target vector). ),Will and The components are spliced ​​and aligned within the same feature space, and finally combined to generate the plant state matrix at the current time (e.g., the matrix form is [...]). The expected genotype index refers to the ideal growth form standard target value pre-set based on the genetic characteristics of the garlic variety, specifically including quantitative parameters such as ideal plant height, ideal pseudostem thickness, expected maximum equivalent diameter or projected area of ​​garlic bulb, and standard nutrient conversion inflection point age.

[0056] S3: Based on the current environmental state matrix, analyze the dynamic changes of basic environmental survival parameters and calculate the hydration stability index. Determine whether the basic habitat stability control command is triggered. If so, perform the corresponding water replenishment, acidification adjustment or cooling operation and return to S2; otherwise, proceed to S4.

[0057] In this embodiment, step S3 extracts the liquid level difference and conductivity difference between the current time slice and the adjacent previous time slice from the environmental state matrix at the current moment, and calculates the dynamic water and fertilizer consumption rate. The dynamic water and fertilizer consumption rate is used to characterize the intensity of root absorption. At the same time, the normalized data of pH value and root zone temperature in the current time slice are extracted, and the absolute deviation of the data from the median of the current plant's expected suitable growth range is calculated. The mean is then aggregated to obtain the root zone environmental stress degree.

[0058] When the hydration stability index is lower than the preset stability threshold, the basic habitat stability control command is triggered. By comparing the weight ratio of the dynamic water and fertilizer consumption rate and the root zone environmental stress, the specific water replenishment, acidification or cooling operation is matched and locked. After the operation is completed, the process returns to S2 to reconstruct the environmental state matrix. Otherwise, the process directly enters the analysis process of S4.

[0059] Specifically, the system first extracts the normalized mean liquid level and mean conductivity values ​​from the current time slice (e.g., minute 5) and the adjacent previous time slice (e.g., minute 0) from the current environmental state matrix. By calculating the difference between these two values ​​over time, the liquid level decrease rate and conductivity decrease rate are obtained. These two values ​​are then weighted and summed to calculate the dynamic water and fertilizer consumption rate. The larger this rate, the more intense the water and nutrient absorption activity of the garlic roots during this period. Simultaneously, the system extracts normalized data of pH and root zone temperature from the current environmental state matrix within the current time slice. It also retrieves the median of the expected suitable growth range for this garlic variety during the current growth stage (e.g., a normalized value of 0.5 for the suitable median pH and 0.4 for the suitable median root zone temperature). The system then calculates the absolute deviation of the current pH value from 0.5 (e.g., the current...). The system calculates the absolute deviation of the root zone environment stress by a mean of 0.1 (e.g., pH normalization is 0.6, so the absolute deviation is 0.1) and the absolute deviation of the current root zone temperature from 0.4 (e.g., if the current root zone temperature normalization is 0.45, then the absolute deviation is 0.05). These two absolute deviations are then averaged to obtain the root zone environmental stress degree (i.e., (0.1+0.05) / 2=0.075). The larger this stress degree, the deeper the deviation of the current root zone's basic living environment from the optimal suitable state. After obtaining the dynamic water and fertilizer consumption rate and the root zone environmental stress degree, the system uses an inversely proportional weighted fusion logic to calculate the hydration steady-state index. The hydration steady-state index is inversely proportional to the dynamic water and fertilizer consumption rate (the faster the consumption, the greater the difficulty in maintaining steady state, and the lower the index), and also inversely proportional to the root zone environmental stress degree (the higher the stress, the worse the steady state, and the lower the index). To prevent overflow in the reciprocal calculation when the consumption rate approaches zero, a preset minimum constant is introduced. (like The hydration steady-state index is calculated using a denominator-coupled structure. Specifically, the consumption rate is multiplied by a preset first weighting coefficient, and the stress degree is multiplied by a preset second weighting coefficient. These two coefficients are then summed in the denominator, with a constant 1 used as the numerator. The formula is 1 / (consumption rate × weight 1 + stress degree × weight 2 + ...). This structure strictly ensures from the mathematical foundation of reciprocal logic that: the faster the consumption or the higher the stress, the larger the denominator, the lower the hydration steady-state index, and the index is permanently constrained within the positive range; after calculating the hydration steady-state index, the system compares it with the preset steady-state threshold to determine the subsequent flow direction.

[0060] The steady-state threshold is set as follows: In stage S1, the system pre-imports a large dataset of environmental state matrices of the garlic variety under different growth stages in historical multi-crop hydroponics experiments without any physiological stress or disease symptoms. By extracting the dynamic water and fertilizer consumption rate and root zone environmental stress degree corresponding to each time slice in the healthy dataset, the system back-calculates the hydration steady-state index distribution curve under historical healthy conditions using the same inverse weighted fusion formula. Subsequently, the lower quantile of this distribution curve is extracted using a pre-set confidence interval (e.g., 95% confidence level) as the steady-state threshold. This threshold accurately defines the maximum environmental fluctuation threshold that the garlic plant can tolerate from the data base level. When the actually calculated hydration steady-state index is lower than this steady-state threshold, it means that the current environmental fluctuation has exceeded the historical healthy experience boundary. The system determines that a basic habitat stabilization and regulation command is triggered. At this time, the system further reduces the index through direct comparison. The system uses the original values ​​of two preceding factors to accurately match the operation type. Specifically, if the normalized value of the current dynamic water and fertilizer consumption rate is greater than the normalized value of the root zone environmental stress, it indicates that the homeostasis imbalance is mainly due to the lack of nutrient solution caused by the plant's rapid absorption. The system then matches, locks, and executes watering or fertilizing operations. Conversely, if the latter value is larger, it indicates that the homeostasis imbalance is mainly due to the deterioration of pH or temperature indicators. The system then further compares the absolute deviations of pH and temperature. The parameter corresponding to the larger deviation is the source of deterioration. The system matches, locks, and executes acidification or temperature adjustment operations accordingly. After the corresponding equipment operation is completed, the system forcibly returns to S2 and uses the latest collected data to reconstruct the environmental state matrix to start the next round of closed-loop monitoring. If the comparison result shows that the hydration homeostasis index is not lower than the preset homeostasis threshold, it indicates that the current basic habitat is in a safe range, and the system directly jumps to the analysis process of S4.

[0061] S4: Based on the current environmental state matrix and the current plant state matrix, combined with germplasm resource data, analyze the spatiotemporal supply and demand matching of current habitat nutrient supply and plant development rhythm, obtain the nutrient hunger warning coefficient, and determine whether the nutrient dynamic balance regulation command is triggered. If so, automatically execute the metering and stirring operation and then proceed to S5; otherwise, directly proceed to S5.

[0062] In this embodiment, step S4 extracts the current change rate of the equivalent diameter of garlic bulbs from the plant state matrix at the current moment, and calculates the morphological dynamic demand index by combining the expected expansion rate genotype index of the current variety in the germplasm resource data during the growth period. The morphological dynamic demand index is used to characterize the plant development rhythm. At the same time, the current electrical conductivity mapping value is extracted from the environmental state matrix at the current moment as the base number of habitat nutrient supply.

[0063] Multiply the dynamic demand index by a preset time window step size to calculate the expected nutrient consumption increment for the next time slice. Subtract the expected nutrient consumption increment from the basic habitat nutrient supply base to calculate the nutrient supply and demand air conditioning difference value. Perform nonlinear mapping dimensionality reduction on the nutrient supply and demand air conditioning difference value to generate the nutrient hunger early warning coefficient.

[0064] When the nutrient hunger warning coefficient is lower than the preset hunger trigger threshold, the dynamic nutrient balance control command is triggered, and the execution equipment is controlled to automatically perform precise metering and stirring operations according to the expected increase in nutrient consumption. After the operation is completed, the process enters the S5 analysis process. Otherwise, the process directly enters the S5 analysis process.

[0065] It should be specifically explained that the process in S4 for analyzing the spatiotemporal supply and demand matching of current habitat nutrient supply and plant development rhythm and generating a nutrient starvation early warning coefficient is as follows:

[0066] First, extract the rate of change (assumed to be 0.02 mm / min) of the equivalent diameter of the garlic bulb between the current time slice (e.g., minute 30) and the previous time slice (e.g., minute 25) from the plant state matrix at the current moment. Then, combine this with the expected bulb enlargement rate genotype index of this garlic variety in the germplasm resource data during the current bulb enlargement stage (i.e., the baseline curve parameters representing the theoretically optimal daily or minute enlargement rate that should be achieved during the current growth period, pre-constructed based on the genetic characteristics sequencing results of this variety and the historical multi-crop high-yield standard growth model). By calculating the relative approximation degree or ratio between the actual rate of change and this genotype index, the morphological dynamic demand index is obtained (e.g., calculated to be 0.67). This system characterizes the current vigor of the plant's actual developmental rhythm. Simultaneously, it extracts the current conductivity normalized mapping value (assumed to be 0.6) from the current environmental state matrix, using it as the baseline for the actual habitat nutrient supply. Subsequently, to address the issue of inconsistent dimensions, the system introduces a preset "morphology-conductivity conversion coefficient" (used to convert morphological development requirements into an equivalent conductivity consumption rate). The morphological dynamic demand index (0.67) is multiplied by a preset time window step (5 minutes), and then multiplied again by this conversion coefficient (assumed to be 0.1). This predicts and extrapolates the expected nutrient consumption increment (i.e., 0.67) required for the plant's development in the next time slice, which is consistent with the baseline supply. ×5×0.1=0.335 units of conductivity), then subtract the expected nutrient consumption increment (0.335) from the basic habitat nutrient supply base (0.6) to obtain the nutrient supply and demand time-varying air conditioning difference value (0.265). Assuming the current basic habitat nutrient supply base drops to 0.2, the obtained nutrient supply and demand time-varying air conditioning difference value is -0.135. This negative value directly represents the scale of the nutrient supply gap that will not be able to meet the plant's spatiotemporal expansion needs in the future time period. For this nutrient supply and demand time-varying air conditioning difference value, the system specifically introduces a preset S-shaped nonlinear activation function (such as the Logistic function) to perform nonlinear mapping dimensionality reduction processing. Its essential calculation logic is: divide the aforementioned time-varying air conditioning difference value by a preset... A morphologically sensitive scaling factor is used to adjust the steepness of change. The result is then substituted as an exponent into the negative power fraction of the natural constant e for mathematical transformation. Through the inherent smoothing and saturation characteristics of this S-shaped function, the refined difference with unified dimensions that spans the positive and negative intervals (such as the previously calculated -0.135) is directly compressed and nonlinearly mapped to the standardized numerical range of 0 to 1, thereby eliminating the impact of extreme differences on the control logic and generating a final smooth transition nutrient hunger warning coefficient (after S-shaped function mapping transformation, this warning coefficient is in the range of 0 to 1, and the smaller its value, the larger the current nutrient gap and the higher the risk of nutrient deficiency; assuming that the standardized value output after S-shaped function mapping compression is 0).15) In the judgment stage, the system compares the nutrient hunger warning coefficient with a preset hunger trigger threshold. The hunger trigger threshold is obtained by: extracting environmental and plant matrix data of the strain during a healthy and high-yield cultivation period in the same historical period without nutrient deficiency; using the same logic as above to backtrack and calculate the historical spatiotemporal supply and demand imbalance; and extracting the 10th percentile of its nonlinear mapping distribution curve as a critical benchmark point. When the actual calculated nutrient hunger warning coefficient (0.15) is lower than the hunger trigger threshold (e.g., 0.25), it is determined that the current habitat is facing the risk of nutrient deficiency, triggering a nutrient dynamic balance control command. The system directly controls the execution equipment to automatically perform precise metering pumping and circulating stirring operations according to the absolute value of the aforementioned nutrient supply and demand imbalance (i.e., the actual nutrient gap scale represented by 0.135) corresponding to the mother liquor stock equivalent, in order to fill the gap in advance. After the operation is completed, the system enters the analysis process of S5. If the nutrient hunger warning coefficient is not lower than the threshold, it indicates that the current spatiotemporal nutrient supply can support the development rhythm, and the system directly enters the analysis process of S5.

[0067] S5: Based on the current environmental state matrix, analyze the comprehensive risk of biosafety of circulating fluid and atomization delivery habitat to obtain the comprehensive biosafety risk index. Determine whether the biosafety blocking control command is triggered. If so, control the atomization delivery channel to perform sterilization and purification operation first, and return to re-execute S5 for re-inspection. Otherwise, release the safety blocking, control the atomization delivery channel to directly perform root mist operation, and enter S6.

[0068] In this embodiment, step S5 extracts the current turbidity change rate and microbial optical density mapping value of the circulating fluid from the current environmental state matrix, and calculates the fluid biological risk index, which is used to characterize the colony proliferation activity inside the fluid; at the same time, it extracts the current concentration of particulate matter and the dynamic increment of aerosol suspension in the atomized delivery habitat, and calculates the atomized habitat risk index, which is used to characterize the space environment impurity interception requirements;

[0069] The fluid biological risk index and the atomization habitat risk index are weighted and coupled to generate a comprehensive biosafety risk index. When the comprehensive biosafety risk index is higher than the preset safety blocking threshold, a biosafety blocking control command is triggered. The atomization delivery channel is controlled to perform sterilization operations to eliminate biological contamination sources. After the operation is completed, the process is forcibly returned to S5 for re-inspection and evaluation. If the re-inspection and evaluation result shows that the comprehensive biosafety risk index is not higher than the preset safety blocking threshold, it indicates that the atomization delivery channel has reached a safe state. The channel is then controlled to perform root mist operations and continues to enter the analysis process of S6.

[0070] It should be specifically noted that the specific extraction and calculation execution logic for the comprehensive risk analysis process of circulating fluid biosafety and atomized delivery habitat in S5 is as follows:

[0071] In the analysis phase of S5, the system retrieves the implicit environmental features stored in S1 from the current environmental state matrix for transformation and mapping. Specifically, it calculates the derivative of the transmittance decrease slope of the nutrient solution return tube within the current time slice to extract the current turbidity change rate (e.g., calculating that the current turbidity increases at a rate of 0.5 NTU per minute). Simultaneously, it performs a nonlinear mapping transformation on the absolute attenuation ratio of transmittance using a preset stray light-colony concentration standard calibration curve to estimate and extract the microbial optical density mapping value (e.g., mapping an OD value of 0.12). Based on these two, it calculates the fluid biological risk index (e.g., multiplying the change rate by the OD value to obtain 0.06). At the same time, the system performs pulse counting and particle size discrimination on the high-frequency fluctuation sequence of scattered light from the root zone airborne particulate sensor to extract the current concentration of fine dust particles (assumed to be 150 particles / liter), and combines this with the root zone temperature and humidity data in the current environmental state matrix. The system calculates the increase in the suspension half-life of aerosols under the current temperature and humidity conditions (e.g., a temperature of 28°C and humidity of 95%) to obtain the dynamic increase in aerosol suspension (assuming the calculated increase coefficient is 1.2). Based on this, the atomization habitat risk index (e.g., dust concentration multiplied by the increase coefficient equals 180) is calculated. To eliminate the coupling distortion caused by the difference in magnitude between the two, the system introduces preset maximum tolerance thresholds for fluid risk (e.g., 0.2) and atomization risk (e.g., 200). The fluid biorisk index (0.06) and the atomization habitat risk index (180) are normalized to obtain normalized values ​​for fluid risk (i.e., 0.06 / 0.2 = 0.3) and atomization risk (i.e., 180 / 200 = 0.9). Then, the two are assigned preset fluid and spatial weight coefficients (e.g., 0.4 and 0.6) and added together with standardized weights to generate the final comprehensive biosafety risk index (i.e., the fluid biorisk index is 180). The preset safety blocking threshold (e.g., set to 0.5) is not arbitrarily assigned, but is based on the critical concentration of pathogenic aerosol tolerance of garlic roots in the aerosol cultivation system, combined with the anti-clogging safety margin of the micropore size of the atomizing nozzle. Independent safety critical values ​​for fluid and space are fitted through preliminary orthogonal experiments, and then the system critical safety score is generated by reverse calculation using the same normalization and weighted coupling logic. This index is compared with the preset safety blocking threshold (e.g., set to 0.5). Since 0.66 is higher than 0.5, a biosafety blocking control command is triggered, the system locks the root mist solenoid valve, and the system is adjusted according to the dominant risk. Source-precise scheduling execution equipment: If the fluid risk normalization value is dominant, then the nutrient solution circulation main path is simultaneously controlled to perform ultraviolet / ozone sterilization; if the atomization risk normalization value (such as 0.9 in this example) is dominant, then the atomization delivery channel is controlled to perform air purification and nozzle sterilization purification operations to simultaneously eliminate biological pollution sources from the source and the end, and after the operation is completed, it is forcibly returned to re-execute S5 for re-inspection and judgment; if the biosafety comprehensive risk index recalculated by the re-inspection drops to 0.45 (not higher than the threshold of 0.5), it indicates that the habitat has reached a safe state, the blockage is lifted, the root mist operation is controlled, and the analysis process of S6 continues.

[0072] S6: By analyzing the plant state matrix at the current moment, the early screening coefficient is calculated to determine whether the inferior elimination instruction is triggered. If so, the breeding control process of the current plant is terminated; otherwise, the candidate qualification is retained, and the early screening coefficient is transmitted to S7 as a penalty weight coefficient.

[0073] In this embodiment, S6 extracts the numerical sequence of equivalent diameter or projected area representing the swelling morphology of garlic bulb from the phenotypic feature vectors in multiple consecutive time slices based on the plant state matrix at the current time. Based on the numerical sequence, the actual phenotypic development process is deduced, and the actual phenotypic development process is compared with the expected genotype index statically attached in the plant state matrix at the current time. The degree of deviation between the two is analyzed, and the early screening coefficient is calculated.

[0074] When the early screening coefficient exceeds the preset inferior elimination threshold, an inferior elimination instruction is triggered, the germplasm level of the plant is downgraded to the conventional commercial level, the allocation of high-level environmental regulation resources to the plant is cut off, and the breeding control cycle of the plant is forcibly terminated; otherwise, it is determined that the inferior elimination instruction has not been triggered, its germplasm candidate status is retained, and the early screening coefficient is positively amplified and mapped to be converted into the penalty weight coefficient for environmental formulation assessment, and the analysis process of S7 continues.

[0075] It should be specifically noted that, in the actual calculation, the equivalent diameter numerical sequence representing the garlic bulb enlargement morphology is extracted from multiple consecutive time slices (e.g., the last 5 days). Based on a preset nonlinear growth fitting model (e.g., Logistic growth curve), the instantaneous tangent slope of this sequence at the current moment is extracted as the actual phenotypic development rate (e.g., an instantaneous slope of 0.02 / day). Simultaneously, the "ideal enlargement rate standard value" (e.g., 0.05 / day) is extracted from the statically mounted expected genotype index in the plant state matrix at the current moment. Subsequently, the negative deviation ratio of the two rates is calculated. To prevent abnormal positive fluctuations in actual development from causing a negative deviation value, the system performs non-negative truncation on this ratio, i.e., using the formula... (In this example) The ratio is directly defined as the early screening coefficient. When this coefficient (0.6) is greater than the inferior elimination threshold (0.35), it indicates that the expansion is severely delayed, triggering elimination; if it does not exceed the threshold, the system performs a positive amplification mapping processing on the early screening coefficient (for example, using formula 1 + early screening coefficient value, which in this example is converted to...). This maps the "degree of developmental disadvantage" positively to a "penalty weight coefficient" greater than 1. The larger the penalty weight coefficient, the more obvious the disadvantage in the early stage, and the stronger the penalty amplification force will be as a product factor in subsequent S7. The preset inferior elimination threshold is a critical threshold determined by retrospectively analyzing the plants that ultimately fail to meet the commercial garlic standards based on historical breeding data of homologous germplasm, and taking the lower limit cutoff value of their early developmental delay deviation rate (for example, set to 0.35).

[0076] S7: Based on the current plant state matrix, germplasm resource data, and penalty weight coefficient, a weighted analysis is performed on the degree of matching expression between the actual phenotypic development of the plant and the expected genotype to obtain an expression excellence index, which is used to determine whether the exclusive formula solidification instruction is triggered.

[0077] In this embodiment, step S7 extracts the phenotypic feature vector from the plant state matrix at the current moment after dimension normalization, and subtracts it item by item from the expected genotype index aligned in the matrix to obtain a multidimensional basic deviation vector. For the deviation component in the multidimensional basic deviation vector that represents the garlic bulb enlargement morphology, a penalty weight coefficient is introduced for product weighting to amplify the early developmental disadvantage in spatial coordinates. Subsequently, the weighted multidimensional basic deviation vector is used as the deviation coordinate point in space, and the Euclidean distance between the deviation coordinate point and the baseline coordinate point corresponding to the expected genotype index is calculated. The Euclidean distance is then reciprocally mapped to generate an expression superiority index.

[0078] When the expression superiority index is greater than or equal to the preset solidification threshold, a special formula solidification instruction is triggered. The complete temporal trajectory of the environmental feature vector of the plant, dynamically updated by continuous time slices within the current control period, is extracted in reverse along the time axis. The temporal trajectory of the environmental feature vector is used as the input independent variable, and the expression superiority index is used as the output dependent variable label for feature mapping and binding. A digital breeding model of the garlic germplasm is generated and persistently stored for direct reuse of homologous germplasm. Otherwise, it is determined that the current formula does not have genetic reproducibility, and the current evaluation process is exited and returned to S2 to wait for the next time window to continue dynamic monitoring and evaluation.

[0079] It should be specifically noted that, in the actual calculation, it is assumed that the phenotypic feature vector after dimension normalization is a three-dimensional vector [plant height score, stem diameter score, garlic bulb diameter score] (for example, the currently measured values ​​are [0.8, 0.7, 0.5]), and the expected genotype index (baseline coordinate point) is [0.9, 0.8, 0.9]. First, the multidimensional basic deviation vector is obtained by subtracting each item one by one as [-0.1, -0.1, -0.4]. To avoid the actual phenotypic development exceeding expectations (such as garlic bulb diameter reaching a certain value), the vector is adjusted accordingly. When a positive deviation (1.0) abnormally lowers the excellence index, the system performs a non-positive truncation logic on the deviation vector: It judges each deviation component; if the deviation is positive (i.e., the actual performance is better than expected), it is forcibly set to zero to exempt from deduction; if the deviation is negative, the original value is retained. For the third-dimensional deviation component (-0.4) representing the garlic bulb enlargement morphology, a penalty weighting coefficient generated by S6 (based on S6 logic, assumed to be 1.6 to reflect the amplified penalty effect) is introduced for product weighting, i.e. The updated deviation coordinates become [-0.1, -0.1, -0.64]; subsequently, the Euclidean norm of this multidimensional deviation vector is calculated as the comprehensive deviation distance (in this example, ). Finally, the Euclidean distance is transformed by a reciprocal mapping (e.g., using the smoothed reciprocal formula of 1 / (1+distance value)). In this example, the excellence index is calculated as 1 / (1+0.655)≈0.60. Since this index (0.60) is less than the solidification threshold (0.85), the system determines that the solidification command will not be triggered and returns to S2 to continue monitoring. If the index calculated later is greater than or equal to 0.85, a special formula solidification command containing the identity of the current qualified plant and the locking status of the high-order environmental control parameters is generated. After the system responds to the command, it extracts the time-series trajectory to generate a breeding model. The specific execution process of generating the breeding model is as follows: the time-series trajectory is generated. The temporal trajectory of the complete environmental feature vector extracted inversely (such as a matrix sequence containing the changes of multidimensional environmental parameters such as temperature, humidity, and light over time) is used as the multidimensional input independent variable matrix X. The final fixed expression superiority index is used as the unique output dependent variable label Y. Subsequently, a preset feature mapping binding algorithm is called to perform structured encapsulation and association mapping between the multidimensional input independent variable matrix X and the output dependent variable label Y, generating a digital breeding formula map containing a high-dimensional correspondence between "environmental input and phenotypic output". The parameter matrix and association structure of this map are serialized and persistently stored, thereby generating a digital breeding model of the garlic germplasm. The preset solidification threshold refers to the "minimum excellence threshold" (e.g., set to 0.85) determined by statistically taking the lower quantile of the expression superiority index distribution of successfully solidified formula samples in historical breeding experiments of homologous garlic germplasm.

[0080] like Figure 2 This embodiment provides an implementation system corresponding to a comprehensive control method for garlic germplasm innovation and efficient breeding processes. The system includes a data acquisition module, a dynamic state matrix construction module, a basic habitat stabilization and regulation module, a nutrient dynamic balance regulation module, a biosafety blocking and delivery module, an early screening and elimination assessment module for substandard garlic, and a dedicated formula solidification triggering module. The data acquisition module is connected to the dynamic state matrix construction module, which is also connected to the basic habitat stabilization and regulation module, the nutrient dynamic balance regulation module, the biosafety blocking and delivery module, the biosafety blocking and delivery module, the early screening and elimination assessment module for substandard garlic, and the dedicated formula solidification triggering module.

[0081] Data acquisition module: It records and collects growth environment data, plant growth data and germplasm resource data in real time during the growth process of garlic plants, and transmits the collected growth environment data, plant growth data and germplasm resource data to the data storage unit for storage.

[0082] Dynamic state matrix construction module: Extracts the current growth environment data and plant growth data from the data storage unit according to a preset time window, and constructs the current environmental state matrix and the current plant state matrix.

[0083] Basic habitat stability control module: Based on the current environmental state matrix, analyze the dynamic changes of basic environmental survival parameters and calculate the hydration stability index. Determine whether the basic habitat stability control command is triggered. If so, execute the corresponding water replenishment, acidification or cooling operation and return to S2; otherwise, enter S4.

[0084] Nutrient dynamic balance regulation module: Based on the current environmental state matrix and the current plant state matrix, combined with germplasm resource data, it analyzes the spatiotemporal supply and demand matching of current habitat nutrient supply and plant development rhythm, obtains nutrient hunger warning coefficient, and determines whether the nutrient dynamic balance regulation command is triggered. If so, it automatically executes metering and stirring operations and then enters S5; otherwise, it directly enters S5.

[0085] Biosafety Interception and Delivery Module: Based on the current environmental state matrix, the module analyzes the comprehensive risk of biosafety of circulating fluid and atomized delivery habitat, obtains the comprehensive biosafety risk index, and determines whether a biosafety interruption control command is triggered. If so, the module controls the atomized delivery channel to perform sterilization and purification operations first, and returns to re-execute S5 for re-inspection. Otherwise, the module releases the safety interruption, controls the atomized delivery channel to directly perform root mist operation, and enters S6.

[0086] Early screening and elimination assessment module for inferior plants: By analyzing the plant status matrix at the current moment, the early screening coefficient is calculated to determine whether the inferior elimination instruction is triggered. If so, the breeding control process of the current plant is terminated; otherwise, the candidate qualification is retained, and the early screening coefficient is transmitted to S7 as a penalty weight coefficient.

[0087] The dedicated formula solidification trigger module: Based on the current plant state matrix, germplasm resource data, and penalty weight coefficient, it performs a weighted analysis on the degree of matching expression between the actual phenotypic development of the plant and the expected genotype, and obtains the expression excellence index, which is used to determine whether to trigger the dedicated formula solidification command.

[0088] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0089] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A comprehensive control method for garlic germplasm innovation and efficient breeding process, characterized in that, include: S1: Real-time recording and collection of growth environment data, plant growth data, and germplasm resource data during the growth process of garlic plants, and transmission of the collected growth environment data, plant growth data, and germplasm resource data to the data storage unit for storage. S2: Extract the current growth environment data and plant growth data from the data storage unit according to the preset time window, and construct the current environmental state matrix and the current plant state matrix. S3: Based on the current environmental state matrix, analyze the dynamic changes of basic environmental survival parameters and calculate the hydration stability index. Determine whether the basic habitat stability control command is triggered. If so, perform the corresponding water replenishment, acidification adjustment or cooling operation and return to S2; otherwise, proceed to S4. S4: Based on the current environmental state matrix and the current plant state matrix, combined with germplasm resource data, analyze the spatiotemporal supply and demand matching of current habitat nutrient supply and plant development rhythm, obtain the nutrient hunger warning coefficient, and determine whether the nutrient dynamic balance regulation command is triggered. If so, automatically execute the metering and stirring operation and then proceed to S5; otherwise, directly proceed to S5. S5: Based on the current environmental state matrix, analyze the comprehensive risk of biosafety of circulating fluid and atomized delivery habitat, obtain the comprehensive biosafety risk index, determine whether the biosafety blocking control command is triggered, if so, control the atomized delivery channel to perform sterilization and purification operation first, and return to re-execute S5 for re-inspection, otherwise, release the safety blocking, control the atomized delivery channel to directly perform root mist operation, and enter S6. S6: By analyzing the plant state matrix at the current moment, the early screening coefficient is calculated to determine whether the inferior elimination instruction is triggered. If so, the breeding control process of the current plant is terminated; otherwise, the candidate qualification is retained, and the early screening coefficient is transmitted to S7 as a penalty weight coefficient. S7: Based on the current plant state matrix, germplasm resource data, and penalty weight coefficient, a weighted analysis is performed on the degree of matching expression between the actual phenotypic development of the plant and the expected genotype to obtain an expression excellence index, which is used to determine whether the exclusive formula solidification instruction is triggered.

2. The comprehensive control method for garlic germplasm innovation and efficient breeding process according to claim 1, characterized in that, S1 involves deploying a multimodal sensor array in the nutrient solution circulation pipeline and root mist habitat of the soilless cultivation system. Various types of sensors within the array are used to collect real-time, high-frequency growth environment data, including liquid level, conductivity, pH value, nutrient solution temperature, and root zone temperature and humidity. Machine vision devices deployed on the cultivation racks capture garlic canopy images within a preset time window, and edge computing is used to extract plant growth data, including plant height, pseudostem diameter, and bulb enlargement morphology. Simultaneously, based on the varietal traceability identifier at the time of garlic transplantation, the expected genotype indicators corresponding to the line from a pre-set database are statically mounted as germplasm resource data. After being encapsulated via an IoT gateway protocol, growth environment data, plant growth data, and germplasm resource data are aggregated and transmitted to the data storage unit in real time.

3. The comprehensive control method for garlic germplasm innovation and efficient breeding process according to claim 2, characterized in that, The S2 uses a preset time window as the boundary of the time slice, performs timestamp synchronization and mean aggregation on the high-frequency growth environment data within the time window, and maps the liquid level, conductivity, pH value, nutrient solution temperature and root zone temperature and humidity into the elements of each dimension of the environmental feature vector, and constructs the current environmental state matrix that is dynamically updated over time. Meanwhile, the plant height, pseudostem diameter, and equivalent diameter or projected area values ​​representing the swelling morphology of garlic bulbs extracted by edge computing are normalized in terms of dimensions, arranged in order to construct a phenotypic feature vector, and spatially aligned with the statically mounted expected genotype indicators to generate the plant state matrix at the current moment.

4. The comprehensive control method for garlic germplasm innovation and efficient breeding process according to claim 3, characterized in that, S3 extracts the liquid level difference and conductivity difference between the current time slice and the adjacent previous time slice from the environmental state matrix at the current moment, and calculates the dynamic water and fertilizer consumption rate, which is used to characterize the intensity of root absorption; at the same time, it extracts the normalized data of pH value and root zone temperature in the current time slice, calculates the absolute deviation of each from the median of the current plant's expected suitable growth range, and performs mean aggregation to obtain the root zone environmental stress degree; When the hydration stability index is lower than the preset stability threshold, the basic habitat stability control command is triggered. By comparing the weight ratio of the dynamic water and fertilizer consumption rate and the root zone environmental stress, the specific water replenishment, acidification or cooling operation is matched and locked. After the operation is completed, the process returns to S2 to reconstruct the environmental state matrix. Otherwise, the process directly enters the analysis process of S4.

5. The comprehensive control method for garlic germplasm innovation and efficient breeding process according to claim 4, characterized in that, S4 extracts the current change rate of the equivalent diameter of garlic bulb from the plant state matrix at the current moment, and calculates the morphological dynamic demand index by combining the expected expansion rate genotype index of the current variety in the germplasm resource data during the growth period. The morphological dynamic demand index is used to characterize the plant development rhythm. At the same time, the current electrical conductivity mapping value is extracted from the environmental state matrix at the current moment as the base number of habitat basic nutrient supply. Multiply the dynamic demand index by a preset time window step size to calculate the expected nutrient consumption increment for the next time slice. Subtract the expected nutrient consumption increment from the basic habitat nutrient supply base to calculate the nutrient supply and demand air conditioning difference value. Perform nonlinear mapping dimensionality reduction on the nutrient supply and demand air conditioning difference value to generate the nutrient hunger early warning coefficient. When the nutrient hunger warning coefficient is lower than the preset hunger trigger threshold, the dynamic nutrient balance control command is triggered, and the execution equipment is controlled to automatically perform precise metering and stirring operations according to the expected increase in nutrient consumption. After the operation is completed, the process enters the S5 analysis process. Otherwise, the process directly enters the S5 analysis process.

6. The comprehensive control method for garlic germplasm innovation and efficient breeding process according to claim 5, characterized in that, S5 extracts the current turbidity change rate and microbial optical density mapping value of the circulating fluid from the current environmental state matrix, and calculates the fluid biological risk index, which is used to characterize the activity of bacterial colony proliferation inside the fluid; at the same time, it extracts the current concentration of particulate matter and the dynamic increment of aerosol suspension in the atomized delivery habitat, and calculates the atomized habitat risk index, which is used to characterize the space environment impurity interception requirements. The fluid biological risk index and the atomized habitat risk index are weighted and coupled to generate a comprehensive biosafety risk index. When the comprehensive biosafety risk index is higher than the preset safety blocking threshold, a biosafety blocking control command is triggered. The atomization delivery channel is prioritized to perform sterilization operations to eliminate biological contamination sources. After the operation is completed, the system is forced to return to S5 for re-inspection and evaluation. If the re-inspection and evaluation result shows that the comprehensive biosafety risk index is not higher than the preset safety blocking threshold, it indicates that the atomization delivery channel has reached a safe state. The system is then controlled to perform root mist operations and continues to enter the analysis process of S6.

7. The comprehensive control method for garlic germplasm innovation and efficient breeding process according to claim 6, characterized in that, S6 extracts the equivalent diameter or projected area of ​​garlic bulb enlargement from the phenotypic feature vectors in multiple consecutive time slices based on the plant state matrix at the current time. Based on this numerical sequence, it infers the current actual phenotypic development process and compares the actual phenotypic development process with the expected genotype index statically attached in the plant state matrix at the current time to analyze the degree of deviation between the two and calculate the early screening coefficient. When the early screening coefficient exceeds the preset inferior elimination threshold, an inferior elimination instruction is triggered, the germplasm level of the plant is downgraded to the conventional commercial level, the allocation of high-level environmental regulation resources to the plant is cut off, and the breeding control cycle of the plant is forcibly terminated; otherwise, it is determined that the inferior elimination instruction has not been triggered, its germplasm candidate status is retained, and the early screening coefficient is positively amplified and mapped to be converted into the penalty weight coefficient for environmental formulation assessment, and the analysis process of S7 continues.

8. The comprehensive control method for garlic germplasm innovation and efficient breeding process according to claim 7, characterized in that, S7 extracts the dimensionally normalized phenotypic feature vector from the plant state matrix at the current moment, subtracts it item by item from the expected genotype index aligned in the matrix, and obtains a multidimensional basic deviation vector. For the deviation component in the multidimensional basic deviation vector that represents the garlic bulb enlargement morphology, a penalty weight coefficient is introduced for product weighting to amplify the early developmental disadvantage in spatial coordinates. Subsequently, the weighted multidimensional basic deviation vector is used as the deviation coordinate point in space, and the Euclidean distance between the deviation coordinate point and the baseline coordinate point corresponding to the expected genotype index is calculated. The Euclidean distance is then reciprocally mapped to generate an expression superiority index. When the expression superiority index is greater than or equal to the preset solidification threshold, a special formula solidification instruction is triggered. The complete temporal trajectory of the environmental feature vector of the plant, dynamically updated by continuous time slices within the current control period, is extracted in reverse along the time axis. The temporal trajectory of the environmental feature vector is used as the input independent variable, and the expression superiority index is used as the output dependent variable label for feature mapping and binding. A digital breeding model of the garlic germplasm is generated and persistently stored for direct reuse of homologous germplasm. Otherwise, it is determined that the current formula does not have genetic reproducibility, and the current evaluation process is exited and returned to S2 to wait for the next time window to continue dynamic monitoring and evaluation.

Images