Modularized intelligent planting space management end comprehensive information service system

Abstract

The invention discloses a modularized intelligent planting space management end comprehensive information service system, and relates to the technical field of planting space management. According to the modularized intelligent planting space management end comprehensive information service system, by arranging the planting space module division module, the planting space data management module and the planting crop growth analysis module, the planting space management efficiency and accuracy can be improved, stable growth of crops is guaranteed, comprehensive information service support is provided for intelligent planting, and the intelligent planting space management end comprehensive information service system is suitable for popularization and application. The scientific and intelligent management of the planting process can be realized, and the problems of disordered space division, disordered environmental data management and inaccurate crop growth state monitoring in the traditional planting space management can be effectively solved.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of planting space management, in particular to a modular intelligent planting space management terminal comprehensive information service system. BACKGROUND

[0002] The current modular intelligent planting space management terminal comprehensive information service system has integrated multiple advanced technologies. At the data collection level, various high-precision sensors such as temperature and humidity, light, soil moisture and gas sensors are used to accurately collect multi-dimensional environmental parameters and crop physiological data in the planting space in real time. Relying on Internet of Things technology, with the help of 4G / 5G, NB-IoT, Wi-Fi and other communication methods, data is stably and quickly transmitted to the cloud platform.

[0003] The system adopts modular design, and each functional module such as environment monitoring, crop growth management, water and fertilizer management, and energy consumption management works independently and cooperatively, is convenient to install, expand and maintain, has good software and hardware compatibility, can adapt to different brands and types of agricultural facilities and equipment, and supports multi-terminal collaboration. Through the command center LED large screen, mobile phone APP and Web platform, management personnel can obtain information and perform operations at any time and anywhere.

[0004] However, there are problems of disordered space division, chaotic environmental data management and inaccurate crop growth state monitoring in traditional planting space management. SUMMARY

[0005] In view of the deficiencies of the prior art, the application provides a modular intelligent planting space management terminal comprehensive information service system, which solves the problems of disordered space division, chaotic environmental data management and inaccurate crop growth state monitoring in traditional planting space management.

[0006] To achieve the above purpose, the application is implemented by the following technical scheme: a modular intelligent planting space management terminal comprehensive information service system, comprising a planting space module division module, a planting space data management module and a planting crop growth analysis module, wherein: The planting space module division module is used for numbering the planting space based on crop types and crop growth stages; The planting space data management module is used for acquiring planting space information and managing the planting space based on the planting space information to determine the environmental stability result of the planting space; The planting crop growth analysis module is used for acquiring crop growth information and analyzing and managing the crop growth state in combination with the planting space information.

[0007] Further, the numbering of the planting space based on the crop types and the crop growth stages comprises: acquiring the encoding of each type of crop stored in the database; dividing the planting space into physical regions based on the physical space where each type of crop is located, and encoding the divided physical regions based on the encoding of each type of crop, to obtain a physical region encoding; acquiring the growth number of the growth stage of each type of crop; concatenating the growth number corresponding to the current growth stage of each type of crop to the rear of the physical region encoding to obtain a planting space encoding, thereby realizing the numbering of the planting space.

[0008] Further, the planting space data management module comprises a space information acquisition unit, an environmental parameter management unit, and an environmental stability analysis unit, wherein: The space information acquisition unit is configured to acquire the planting space information of each physical region according to the planting space encoding, wherein the planting space information comprises static information and dynamic information; The environmental parameter management unit is configured to acquire the environmental parameter setting range corresponding to the planting space encoding of each physical region stored in the database: If each parameter in the dynamic information of any physical region is within the environmental parameter setting range, no environmental parameter management is performed; If any parameter in the dynamic information of a certain physical region is not within the corresponding environmental parameter setting range, the parameter is adjusted until it is within the corresponding environmental parameter setting range; The environmental stability analysis unit is configured to perform data analysis based on the dynamic information when each parameter in the dynamic information of any physical region is within the environmental parameter setting range, to obtain an environmental stability evaluation coefficient of each physical region as the environmental stability result of each physical region.

[0009] Further, the static information comprises physical parameters of the planting space and device configuration information, wherein the physical parameters comprise planting space encoding, space area, space type, and belonging region; and the device configuration information comprises sensor type and quantity, and execution device type and quantity; The dynamic information comprises initial environmental data when the planting space is enabled, wherein the initial environmental data comprises initial air temperature, initial air humidity, initial soil humidity, initial light intensity, and initial CO2 concentration.

[0010] Further, the environmental stability evaluation coefficient acquisition process is: Standardizing the dynamic information based on the environmental parameter setting range to obtain standardized dynamic information values, wherein the standardized dynamic information values comprise air temperature standardized value, air humidity standardized value, soil humidity standardized value, light intensity standardized value, and CO2 concentration standardized value; The standardized dynamic information values collected based on the set collection frequency are subjected to mean square difference processing to obtain dynamic information mean square difference values, including air temperature mean square difference values, air humidity mean square difference values, soil humidity mean square difference values, light intensity mean square difference values and CO2 concentration mean square difference values; The dynamic information mean square difference values are subjected to weighted summation to obtain an environmental stability evaluation coefficient.

[0011] Further, the crop growth information includes dynamic information, plant height proportion, leaf number proportion, leaf area proportion, number of environmental abnormalities in the last 3 days, number of irrigation in the last 7 days, number of fertilization in the last 7 days and crop leaf RGB image; The crop growth analysis module includes a crop growth state analysis unit and a current growth stage end time prediction unit, wherein: The crop growth state analysis unit is configured to determine whether the crop has a disease or pest based on the crop growth information: If not, determine the crop growth normal state compliance degree based on the crop growth information; If so, obtain a corresponding remediation scheme from the database based on the determined disease or pest type; The current growth stage end time prediction unit is configured to obtain crop growth data and determine the expected end time of the current growth stage based on the crop growth data.

[0012] Further, determining whether the crop has a disease or pest based on the crop growth information includes: The dynamic information, plant height proportion, leaf number proportion, leaf area proportion, number of environmental abnormalities in the last 3 days, number of irrigation in the last 7 days and number of fertilization in the last 7 days are input into the trained random forest classifier to obtain random forest growth state classification probabilities, including a first normal probability, a first mild pest probability, a first moderate pest probability and a first severe pest probability; The crop leaf RGB image is input into the CNN feature extraction model to output an image feature vector; The image feature vector is processed by the support vector machine model to output a mixed growth state classification probability, including a second normal probability, a second mild pest probability, a second moderate pest probability and a second severe pest probability; The random forest growth state classification probability and the mixed growth state classification probability are subjected to weighted summation to obtain a prediction probability, including a growth normal probability, a mild pest probability, a moderate pest probability and a severe pest probability; The class with the highest probability is taken as the final prediction result to determine whether there is a disease or pest.

[0013] Further, determining the crop growth normal state compliance degree based on the crop growth information includes: According to the physical region where the crop is located, the corresponding planting space code is obtained; obtain crop growth reference information corresponding to the planting space code from the database; fuse and analyze the crop growth information and the crop growth reference information to obtain a similarity, denoted as a crop growth normal state coincidence degree.

[0014] Further, the calculation formula of the similarity is: wherein F is the similarity, is the crop growth information, is the crop growth reference information, is a cosine similarity function.

[0015] Further, determining an expected end time of a current growth stage based on the crop growth data, comprising: inputting the crop growth data into the trained gradient boosting regression tree model to obtain the expected end time of the current growth stage.

[0016] The present application has the following advantages: The modular intelligent planting space management terminal comprehensive information service system can improve the efficiency and accuracy of planting space management, ensure stable growth of crops, provide comprehensive information service support for intelligent planting, and help realize scientific and intelligent management of the planting process, which can effectively solve the problems of disordered space division, chaotic environmental data management and inaccurate crop growth state monitoring in traditional planting space management.

[0017] Of course, implementing any product of the present application does not necessarily require all the advantages described above. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 The present application is a modular intelligent planting space management terminal comprehensive information service system flowchart. DETAILED DESCRIPTION

[0019] Please refer to Figure 1 The present application provides a technical solution: a modular intelligent planting space management terminal comprehensive information service system, which includes a planting space module division module, a planting space data management module and a planting crop growth analysis module, wherein: ​The planting space module division module is used for numbering the planting space based on crop types and crop growth stages; the standardized crop codes stored in the database are directly called, the repeated work of creating codes for different crops is avoided, the uniformity and standardization of the crop codes are ensured, the problem of low management efficiency caused by the confusion of crop identification and the lack of fixed codes in traditional planting is solved, and a standardized foundation is laid for subsequent space coding.

[0020] The codes of various types of crops stored in the database are obtained; the planting space is physically divided into physical regions based on the physical space where various types of crops are located, and the physical regions after the division are coded based on the codes of various types of crops to obtain physical region codes; the standardized crop codes stored in the database are directly called, the repeated work of creating codes for different crops is avoided, the uniformity and standardization of the crop codes are ensured, the problem of low management efficiency caused by the confusion of crop identification and the lack of fixed codes in traditional planting is solved, and a standardized foundation is laid for subsequent space coding.

[0021] The growth numbers of the growth stages of various types of crops are obtained; the growth numbers can accurately identify the current growth stage (such as the germination stage, the seedling stage, etc.) of the crops, and solve the problem of "mixing different growth stages of the same crop in the same space and being unable to manage them specifically" in traditional planting, and provide an identification basis for the subsequent need to distinguish different growth stages of the same crop.

[0022] The growth numbers of the current growth stages of various types of crops are spliced at the back of the physical region codes to obtain planting space codes, and the numbering of the planting space is realized. The finally generated planting space code contains not only the physical region information corresponding to the crop type, but also the current growth stage information of the crop, and realizes the fine management of the planting space "precise to crop type + growth stage", and provides clear and unique identification support for the subsequent planting space data management module to obtain space information according to the code and the planting crop growth analysis module to analyze the growth state.

[0023] The planting space data management module is used for obtaining planting space information and managing the planting space based on the planting space information to determine the environmental stability result of the planting space. The planting space data management module includes a space information acquisition unit, an environmental parameter management unit and an environmental stability analysis unit, wherein: the space information acquisition unit is used for obtaining the planting space information of each physical region according to the planting space code, and the planting space information includes static information and dynamic information. The environment parameter management unit is configured to acquire the environment parameter setting range corresponding to the planting space code of each physical region stored in the database; if each parameter in the dynamic information of any physical region is within the environment parameter setting range, no environment parameter management is performed; if any parameter in the dynamic information of a physical region is not within the corresponding environment parameter setting range, the parameter is adjusted until it is within the corresponding environment parameter setting range; the environment parameter setting range corresponding to each physical region code in the database is acquired first, and different management strategies are adopted according to whether the dynamic information meets the range; if all parameters are within the range, no intervention is performed, and if any parameter exceeds the range, it is adjusted to meet the requirement. The advantage is that, first, the problem of "different crops, different growth stages, different environmental requirements, but unified management according to fixed standards" in traditional planting is solved through "code matching setting range", for example, different temperature setting ranges for tomato seedling stage and fruiting stage can be set to manage the temperature of the corresponding code area; second, the mode of "dynamic adjustment + intervention as needed" can avoid unnecessary parameter adjustment, reduce energy consumption, and timely correct abnormal environment parameters to prevent crops from being hindered due to unsuitable environment and ensure that crops are in a suitable growing environment.

[0024] The environment stability analysis unit is configured to perform data analysis based on the dynamic information when each parameter in the dynamic information of any physical region is within the environment parameter setting range, to obtain an environment stability evaluation coefficient of each physical region as an environment stability result of each physical region. The environment stability evaluation coefficient can intuitively reflect the degree of environmental fluctuation; at the same time, the coefficient can provide more detailed decision basis for subsequent planting management, for example, when the stability coefficient is too low, equipment failure or management strategy can be checked in advance to further reduce the potential impact of environmental fluctuation on crop growth and improve the predictability and accuracy of planting environment management.

[0025] The static information includes physical parameters and device configuration information of the planting space, wherein the physical parameters include planting space code, space area (unit: m2), space type (such as greenhouse, plant factory, vertical planting frame, etc.), and belonging area (such as A greenhouse group, B plant factory); the device configuration information includes sensor type and quantity (temperature and humidity sensor, light sensor, CO2 sensor, etc.), and execution device type and quantity (ventilation equipment, irrigation equipment, light supplementing equipment, etc.); The dynamic information includes initial environment data when the planting space is started, and the initial environment data includes initial air temperature (unit: ℃), initial air humidity (unit: %RH), initial soil humidity (unit: %), initial light intensity (unit: lux), and initial CO2 concentration (unit: ppm).

[0026] The static information (including planting space code, area, type, belonging area and equipment configuration information) and dynamic information (including initial air temperature, humidity, soil humidity and other environmental data) of the corresponding physical area are accurately obtained. The advantages are that, on the one hand, through the coding of the associated information, the problem of "space information being difficult to accurately match with crops and growth stages" in traditional planting is solved, for example, the space equipment configuration and real-time environmental data of a specific growth stage of a crop can be directly retrieved through the code; on the other hand, the classification of static information and dynamic information can clearly distinguish the fixed properties (such as area and equipment) of the planting space and the changing environmental parameters, lay a data foundation for subsequent targeted management of environmental parameters, and avoid the problem of low management efficiency caused by mixed information.

[0027] The environmental stability evaluation coefficient obtaining process is: The dynamic information is standardized based on the environmental parameter setting range to obtain standardized dynamic information values, including air temperature standardized value, air humidity standardized value, soil humidity standardized value, light intensity standardized value and CO2 concentration standardized value; the dynamic information (such as air temperature and humidity) is converted into standardized values, which solves the problem that different environmental parameters cannot be directly compared and analyzed due to different units and value ranges (such as temperature unit °C and humidity unit %RH). For example, air temperature 25°C and air humidity 60%RH originally belong to different dimensional data, and after standardization, they can be unified in the same numerical interval, laying a unified data foundation for subsequent environmental stability calculation and avoiding analysis errors caused by parameter dimension differences.

[0028] The standardized dynamic information values collected based on the set collection frequency are subjected to mean square deviation processing to obtain dynamic information mean square deviation values, including air temperature mean square deviation value, air humidity mean square deviation value, soil humidity mean square deviation value, light intensity mean square deviation value and CO2 concentration mean square deviation value; the mean square deviation can reflect the dispersion degree of the data, and by calculating the mean square deviation of the standardized values of air temperature, humidity and other parameters, the fluctuation of each environmental parameter can be accurately captured. This solves the problem in traditional planting that "only whether the parameter is within the qualified range can be judged, but the fluctuation range of the parameter within the range cannot be known", for example, the temperature of a certain area is always within the qualified range of 20-28°C, but through the mean square deviation, it can be found whether it frequently fluctuates greatly within the range, providing data support for identifying "hidden environmental instability".

[0029] The environmental stability assessment coefficient is obtained by weighted summation of the mean squared errors of dynamic information. This weighted summation can incorporate the varying degrees of influence of different environmental parameters on crop growth (e.g., temperature has a greater impact on seedlings and can be assigned a higher weight) to generate a comprehensive environmental stability assessment coefficient. This coefficient addresses the problem that "fluctuations in a single parameter cannot reflect overall environmental stability," providing a direct and quantitative representation of the overall environmental stability level of the planting space. Subsequent managers can use this coefficient to quickly compare the environmental stability of different planting areas, prioritizing equipment maintenance or parameter optimization in areas with low coefficients (poor stability), thereby improving the targeting and efficiency of environmental management and further ensuring crop growth in a stable environment.

[0030] The crop growth analysis module is used to acquire crop growth information and analyze and manage the crop growth status in conjunction with planting space information.

[0031] The crop growth information includes dynamic information, plant height ratio, leaf number ratio, leaf area ratio, number of environmental anomalies in the past 3 days, number of irrigations in the past 7 days, number of fertilizations in the past 7 days, and RGB images of crop leaves. Defined crop growth information includes dynamic information (such as environmental data like air temperature and humidity), plant height ratio, leaf number ratio, leaf area ratio, number of environmental anomalies in the past 3 days, number of irrigations in the past 7 days, number of fertilizations in the past 7 days, and RGB images of crop leaves. This information encompasses environmental data, crop morphology data, management operation data, and image data, solving the problem in traditional planting where "relying solely on single data points (such as irrigation frequency) to assess crop growth leads to incomplete assessments due to insufficient data dimensions." For example, combining plant height ratio and leaf area ratio can determine whether the crop growth rate is normal; combining the number of environmental anomalies in the past 3 days can analyze the potential impact of environmental fluctuations on the crop; and combining leaf RGB images allows for a direct observation of whether there are any abnormalities in the crop's appearance. These multi-dimensional data provide a comprehensive basis for accurately assessing the crop's growth status.

[0032] The crop growth analysis module includes a crop growth status analysis unit and a current growth stage end time prediction unit, wherein: The crop growth status analysis unit is used to determine whether the crop is affected by pests or diseases based on crop growth information. If not, the degree of conformity to the normal growth status of crops is determined based on crop growth information; If it exists, the corresponding treatment plan will be retrieved from the database based on the identified pest or disease type; The current growth stage end time prediction unit is used to acquire crop growth data and determine the expected end time of the current growth stage based on the data. By acquiring crop growth data and determining the expected end time of the current growth stage, it solves the problem in traditional planting where "judging the end time of crop growth stages based solely on experience leads to large errors and makes it impossible to accurately plan subsequent agricultural operations (such as harvesting and fertilization)." For example, predicting the end time of tomato fruiting in advance allows for advance arrangement of harvesting personnel and preparation of storage equipment, avoiding untimely harvesting or resource waste due to inaccurate timing judgments, improving the planning and efficiency of agricultural management, and ensuring the orderly progress of the planting process.

[0033] Determining whether a crop is affected by pests or diseases based on crop growth information includes: Dynamic information, plant height ratio, leaf number ratio, leaf area ratio, number of environmental anomalies in the past 3 days, number of irrigations in the past 7 days, and number of fertilizations in the past 7 days are input into a trained random forest classifier to obtain random forest growth state classification probabilities, including the first normal probability, the first minor pest probability, the first moderate pest probability, and the first severe pest probability. Dynamic information, plant height ratio, and other multi-dimensional growth data are input into the trained random forest classifier to obtain the first normal probability and the first minor / moderate / severe pest probability. Compared to the traditional single-dimensional judgment method that relies solely on manual observation of morphology, the input of multi-dimensional data can cover key factors affecting pests and diseases, such as environment, crop morphology, and management operations, solving the problem of "missed judgments due to one-sided judgment dimensions." At the same time, the output of classification probabilities quantifies the likelihood of different growth states, providing data support for subsequent judgments and avoiding subjective errors caused by relying solely on experience.

[0034] The RGB image of crop leaves is input into a CNN feature extraction model, which outputs an image feature vector. Image feature vectors are processed by a support vector machine (SVM) model to output mixed growth state classification probabilities, including a second normal probability, a second minor pest probability, a second moderate pest probability, and a second severe pest probability. RGB images of crop leaves are input into a CNN feature extraction model to obtain image feature vectors, which are then processed by the SVM model to obtain the second set of growth state classification probabilities. This step fully utilizes the visual features of the image data (such as leaf spots and yellowing levels), solving the problem of "difficulty in accurately capturing subtle pest and disease features on leaves in traditional judgments." The image features complement the multi-dimensional data mentioned earlier, further enriching the judgment criteria and improving the ability to identify easily overlooked conditions such as minor pest infestations.

[0035] The predicted probabilities are obtained by weighted summation of the probabilities of random forest growth status classification and mixed growth status classification, including the probability of normal growth, the probability of minor pest infestation, the probability of moderate pest infestation, and the probability of severe pest infestation. The category with the highest probability is taken as the final prediction result to determine whether pests or diseases exist.

[0036] By weighted summing of the two sets of classification probabilities and selecting the category with the highest probability as the result, the limitations of single-model judgment (such as the insensitivity of random forests to image data and the insufficient utilization of environmental data by CNNs) are addressed. Weighted fusion combines the advantages of both models, balancing the judgment results from both data and visual dimensions, significantly improving the accuracy of pest and disease identification. Furthermore, the explicit probability ranking makes the final judgment more credible, avoiding management errors caused by misjudgments from a single model, and providing a precise basis for timely pest and disease control measures.

[0037] Determining the conformity of crop growth to normal status based on crop growth information includes: obtaining the corresponding planting space code based on the physical region where the crop is located; obtaining crop growth parameter information corresponding to the planting space code from the database; and performing a fusion analysis of the crop growth information and the crop growth parameter information to obtain the similarity, which is recorded as the conformity of crop growth to normal status.

[0038] The formula for calculating similarity is: ; Where F represents the similarity score. For crop growth information, Provide information for crop growth. This is the cosine similarity function.

[0039] By integrating actual growth information (such as the proportion of actual plant height and the proportion of leaves) with reference data, the qualitative judgment of "whether it is normal" is transformed into a quantitative assessment of "similarity level," solving the problem that "the concept of 'normal' in traditional assessments is vague and cannot measure the degree of growth achievement." For example, the difference between 90% and 70% similarity can clearly reflect the gap between crop growth and the ideal state, providing a clear quantitative basis for whether to adjust management measures (such as increasing fertilization or optimizing light) in the future.

[0040] Determining the expected end time of the current growth stage based on crop growth data includes: inputting crop growth data into a trained gradient boosting regression tree model to obtain the expected end time of the current growth stage.

[0041] For crops in "normal growth status" (based on the results of the crop growth status and pest prediction model mentioned above), the remaining time of their current growth stage is accurately predicted, and the "end time of the current growth stage" (accurate to the day) is finally output, providing a time basis for agricultural operation planning (such as fertilization and harvest preparation), and adapting to the "refined cycle management" needs of smart planting.

[0042] Key concept definition: The current growth stage is a segment of the growth cycle based on the characteristics of the crop variety. For example, tomatoes are divided into "germination period (1-7 days), seedling period (8-30 days), flowering and fruit setting period (31-55 days), and fruiting period (56-90 days)". The model needs to confirm the current stage of the crop by inputting data.

[0043] The end time of the phase is the last day of the current phase. The formula is "current date + predicted remaining time". For example, if the current date is August 10 and the predicted remaining time is 5 days, then the end time is August 15.

[0044] Crop growth data includes basic crop attribute data (crop variety (e.g., tomato 'Pink Crown No. 1', lettuce 'Italian Lettuce'), current growth stage, planting date), environmental characteristic data (daily average air temperature, daily average air humidity, daily average soil moisture, daily average sunshine duration, daily average CO2 concentration), and crop growth status data (daily increase in plant height (cm / day), daily increase in leaf number (leaf / day), daily increase in leaf area (cm² / day)).

[0045] The dataset was constructed based on 24 months of historical data from 5 planting modules (covering 3 crops and 8 varieties), containing a total of 3600 valid samples (only samples with "normal growth status" were retained, with at least 300 samples for each variety at each growth stage). It was divided into a training set (2880 samples) and a test set (720 samples) in an 8:2 ratio, ensuring that the proportion of samples for "each crop variety - growth stage" was consistent in the training and test sets (e.g., if the proportion of samples in the tomato seedling stage is 20%, then both the training and test sets should maintain 20%). The "3σ principle" was used to remove outlier samples (e.g., if the daily average temperature of a sample is 45℃, which is more than 3 times the standard deviation of the temperature for that crop at that stage, it is judged as an outlier and removed), ensuring the quality of the dataset.

[0046] We selected the Gradient Boosting Regression Tree (LightGBM) as the core model. The model structure and core parameter settings include: (1) Input layer: Feature vector construction The preprocessed data from the four categories are integrated into an 18-dimensional feature vector, with the specific dimensions as follows: Crop basic attributes (5 dimensions): Variety unique heat vector (2 dimensions), current growth stage (1 dimension), number of days grown in the current stage (1 dimension), historical average duration (1 dimension). Environmental deviation characteristics (5 dimensions): temperature deviation, humidity deviation, soil moisture deviation, light duration deviation, and CO2 concentration deviation; Growth increment characteristics (3 dimensions): daily increase in plant height, daily increase in number of leaves, and daily increase in leaf area; Historical fluctuation characteristics (5 dimensions): standard deviation of historical duration, deviation of current planting duration from historical mean, shortest historical duration in the same period, longest historical duration in the same period, and average historical duration of the same season in the past 3 years.

[0047] (2) Core parameter optimization (based on validation set tuning) By using 5-fold cross-validation, key parameters are optimized on the training set to ensure that the model balances accuracy and generalization ability.

[0048] (3) Output layer: Calculation of prediction results The model output is "remaining duration of the current growth stage (days)", which requires the following post-processing to ensure its reasonableness: Boundary constraints: Set a reasonable range for the remaining time based on historical data (e.g., the remaining time for tomato seedlings should not be less than 3 days or greater than 20 days). If the predicted value exceeds the range, take the nearest boundary value (e.g., if the predicted value is 25 days, correct it to 20 days). Integer Conversion: The remaining duration must be an integer (days). The prediction result (decimal) is converted using the rounding method (e.g., if the prediction is 5.3 days, it is converted to 5 days; if the prediction is 5.6 days, it is converted to 6 days). End time calculation: Combined with the current system date, the "current growth stage end time" (format: YYYY-MM-DD) is automatically calculated and output.

[0049] Using mean absolute error (MAE) as the loss function, training is performed in stages: Phase 1: Basic Model Training Load the training set data, train the basic regression model using LightGBM, optimize the parameters using 5-fold cross-validation, record the MAE and feature importance for each fold, stop training when the mean 5-fold MAE does not decrease for 10 consecutive rounds, and save the basic model parameters. Phase Two: Variety-Specific Optimization The models were grouped by "crop variety - growth stage" (e.g., tomato seedling stage, lettuce mature stage), and each group was individually fine-tuned (adjusting only the learning rate and the number of leaf nodes) to address the issue of "significant differences in growth patterns among different varieties." For example, tomatoes are sensitive to temperature during the fruiting stage, so the learning rate needs to be reduced to enhance the model's ability to capture temperature deviation features; while lettuce seedlings are more sensitive to light, so the number of leaf nodes needs to be increased to improve feature fitting accuracy.

[0050] Key performance optimization measures include: Feature engineering optimization: New features were constructed using "feature interaction", such as "temperature deviation × light duration deviation" (capturing the effect of temperature and light synergy on growth) and "daily increase in plant height × historical duration deviation" (combining current growth rate with historical patterns). After adding 5-dimensional interactive features, the model's MAE decreased by 12%. Incorporating temporal trends: The "average environmental deviation over the past 7 days" and "average growth increment over the past 7 days" are added as new features to replace daily data, highlighting growth trends and avoiding interference from daily fluctuations, thus improving the model's predictive stability by 15%. Incremental training mechanism: Add "normal crop" growth data (about 150 samples) to the system every month to incrementally train the model (only update the decision tree weights, do not retrain the entire tree) to ensure that the model can adapt to seasonal changes (such as the difference in growth rate between summer and winter) and maintain prediction accuracy in the long term.

[0051] An electronic device includes: a processor; and a memory storing computer program instructions, which, when executed by the processor, cause the processor to perform the modular smart planting space management terminal integrated information service system as described above.

[0052] A computer-readable storage medium for storing a program that, when executed by a processor, implements the modular smart planting space management terminal integrated information service system as described above.

[0053] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0054] This invention is described with reference to flowchart illustrations and / or block diagrams of systems, apparatus (systems), and computer program products according to embodiments of the invention. It will 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, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0055] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function 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 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0056] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable 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.

[0057] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. 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 invention.

[0058] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A modular intelligent planting space management terminal integrated information service system, characterized in that, It includes a planting space module division module, a planting space data management module, and a crop growth analysis module, among which: The planting space module division module is used to number the planting spaces based on crop type and crop growth stage; The planting space data management module is used to acquire planting space information, manage the planting space based on the planting space information, and determine the environmental stability results of the planting space. The crop growth analysis module is used to acquire crop growth information and analyze and manage the crop growth status in conjunction with planting space information.

2. The modular intelligent planting space management terminal integrated information service system according to claim 1, characterized in that, Planting spaces are numbered based on crop type and crop growth stage, including: Retrieve the codes for various crop types stored in the database; The planting space is divided into physical regions based on the physical space where each type of crop is located, and the divided physical regions are coded based on the codes of each type of crop to obtain physical region codes. Obtain the growth number of each type of crop at its growth stage; The growth number corresponding to the current growth stage of each type of crop is appended to the physical area code to obtain the planting space code, thereby realizing the numbering of the planting space.

3. The modular intelligent planting space management terminal integrated information service system according to claim 2, characterized in that, The planting space data management module includes a spatial information acquisition unit, an environmental parameter management unit, and an environmental stability analysis unit, wherein: The spatial information acquisition unit is used to acquire planting space information of each physical region according to the planting space code. The planting space information includes static information and dynamic information. The environmental parameter management unit is used to obtain the environmental parameter setting range corresponding to the planting space code of each physical area stored in the database. If all parameters in the dynamic information of any physical area are within the range of environmental parameter settings, then environmental parameter management will not be performed. If any parameter in the dynamic information of a certain physical area is not within the corresponding environmental parameter setting range, then adjust that parameter until it is within the corresponding environmental parameter setting range; The environmental stability analysis unit is used to perform data analysis based on dynamic information when all parameters in the dynamic information of any physical area are within the set range of environmental parameters, and to obtain the environmental stability evaluation coefficient of each physical area as the environmental stability result of each physical area.

4. The modular intelligent planting space management terminal integrated information service system according to claim 3, characterized in that, The static information includes the physical parameters of the planting space and the equipment configuration information, wherein the physical parameters include the planting space code, space area, space type, and region; the equipment configuration information includes the sensor type and quantity, and the execution device type and quantity. The dynamic information includes the initial environmental data when the planting space is activated, including the initial air temperature, initial air humidity, initial soil moisture, initial light intensity, and initial CO2 concentration.

5. The modular intelligent planting space management terminal integrated information service system according to claim 4, characterized in that, The process for obtaining the environmental stability assessment coefficient is as follows: Based on the environmental parameter setting range, the dynamic information is standardized to obtain standardized dynamic information values, which include standardized values ​​for air temperature, air humidity, soil moisture, light intensity, and CO2 concentration. The standardized dynamic information values ​​collected based on the set collection frequency are subjected to standard deviation processing to obtain the standard deviation values ​​of dynamic information, including the standard deviation values ​​of air temperature, air humidity, soil moisture, light intensity and CO2 concentration. The environmental stability assessment coefficient is obtained by weighted summation of the mean squared errors of the dynamic information.

6. The modular intelligent planting space management terminal integrated information service system according to claim 4, characterized in that, The crop growth information includes dynamic information, plant height ratio, leaf number ratio, leaf area ratio, number of environmental anomalies in the past 3 days, number of irrigations in the past 7 days, number of fertilizations in the past 7 days, and RGB images of crop leaves; The crop growth analysis module includes a crop growth status analysis unit and a current growth stage end time prediction unit, wherein: The crop growth status analysis unit is used to determine whether the crop is affected by pests or diseases based on crop growth information. If not, the degree of conformity to the normal growth status of crops is determined based on crop growth information; If it exists, the corresponding treatment plan will be retrieved from the database based on the identified pest or disease type; The current growth stage end time prediction unit is used to acquire crop growth data and determine the expected end time of the current growth stage based on the crop growth data.

7. The modular intelligent planting space management terminal integrated information service system according to claim 6, characterized in that, Determining whether a crop is affected by pests or diseases based on crop growth information includes: The dynamic information, plant height ratio, leaf number ratio, leaf area ratio, number of environmental anomalies in the past 3 days, number of irrigations in the past 7 days, and number of fertilizations in the past 7 days are input into the trained random forest classifier to obtain the random forest growth status classification probability, including the first normal probability, the first slight pest probability, the first moderate pest probability, and the first severe pest probability. The RGB image of crop leaves is input into a CNN feature extraction model, which outputs an image feature vector. The image feature vectors are processed by a support vector machine model to output the mixed growth state classification probability, including the second normal probability, the second slight pest probability, the second moderate pest probability, and the second severe pest probability. The predicted probabilities are obtained by weighted summation of the probabilities of random forest growth status classification and mixed growth status classification, including the probability of normal growth, the probability of slight pest infestation, the probability of moderate pest infestation and the probability of severe pest infestation. The category with the highest probability is taken as the final prediction result to determine whether pests or diseases exist.

8. The modular intelligent planting space management terminal integrated information service system according to claim 6, characterized in that, Determining the degree of conformity to normal crop growth status based on crop growth information includes: Obtain the corresponding planting space code based on the physical region where the crop is located; Retrieve crop growth parameter information corresponding to the planting space code from the database; The similarity score is obtained by integrating and analyzing crop growth information with crop growth parameter information, and is recorded as the crop growth normal state conformity score.

9. The modular intelligent planting space management terminal integrated information service system according to claim 8, characterized in that, The formula for calculating similarity is: ； Where F represents the similarity score. For crop growth information, Provide information for crop growth. This is the cosine similarity function.

10. The modular intelligent planting space management terminal integrated information service system according to claim 6, characterized in that, Determine the expected end time of the current growth stage based on crop growth data, including: By inputting crop growth data into a trained gradient boosting regression tree model, the expected end time of the current growth stage can be obtained.

Images