Garden data intelligent management method and system based on cloud computing

Abstract

The invention discloses an intelligent garden data management method and system based on cloud computing, and relates to the crossing field of cloud computing technologies, and the method comprises the steps: collecting and preprocessing multi-dimensional garden data through a distributed terminal, and adding classification and space-time labels; using a hybrid cloud architecture to realize data security storage and collaborative management, and establishing a classification index; a deep learning model is trained based on historical data, an analysis result is output, a personalized maintenance scheme is generated, and dynamic optimization is performed through real-time feedback; secure data intercommunication between the platform and various terminals is realized by means of a multi-terminal collaborative interaction module, and data security is guaranteed by adopting encrypted transmission and authority control; monitoring and encryption protection are implemented in the whole process, a traceable log is generated and filed regularly, and continuous iterative upgrading is performed in combination with a model effect. The method has the advantages that based on the hybrid cloud architecture and distributed data processing, garden management is promoted to be transformed to data intelligent driving through full-process closed-loop optimization, and the management efficiency and the refinement level are greatly improved.

Description

Technical Field

[0001] This invention relates to the intersection of cloud computing technology, and in particular to a cloud-based intelligent management method and system for garden data. Background Technology

[0002] With the acceleration of urbanization and the increasing demand for ecological construction, park management faces challenges such as large data volumes, high real-time requirements, and difficulties in integrating multi-source heterogeneous information. Traditional management models rely on manual inspections and decentralized systems, resulting in low efficiency, poor coordination, and delayed decision-making. The maturity of cloud computing technology provides a new approach to solving these pain points.

[0003] Current landscape data management methods, including those mentioned above, largely rely on a single storage architecture, making it difficult to balance the security of sensitive data with the efficient handling of massive amounts of data. This can easily lead to information silos or security vulnerabilities. Data collection dimensions are limited, and preprocessing is often crude, relying heavily on manual recording or single-device collection. There is a lack of standardized classification and spatiotemporal labeling management, resulting in inconsistent data quality. Analysis models are mostly traditional statistical methods or single-function models, making it difficult to achieve deep integration of plant growth assessment, pest and disease early warning, and maintenance prediction. This results in low decision-making accuracy and insufficient personalization. Multi-terminal collaboration capabilities are weak, and data transmission encryption and access control are inadequate, posing a risk of data leakage. Furthermore, there is a lack of end-to-end data traceability and dynamic optimization mechanisms. Management models are rigid and difficult to iteratively adjust based on maintenance results and environmental changes, leading to low overall management efficiency and a low level of precision. Summary of the Invention

[0004] To improve existing methods and systems, this paper proposes a cloud-based intelligent management method and system for garden data. This method relies on a hybrid cloud architecture and distributed data processing, combined with deep learning models to achieve precise and intelligent decision-making for garden maintenance. It takes into account the efficiency of multi-terminal collaboration and the security of data throughout the entire life cycle. Through closed-loop optimization of the entire process, it promotes the transformation of garden management towards data intelligence-driven development, and significantly improves management efficiency and precision.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: Cloud-based intelligent management methods for garden data include: Multi-dimensional data of the garden is collected through distributed acquisition terminals and preprocessed. Based on the garden data classification rules, category labels and spatiotemporal labels are added to different types of data. A garden data storage and collaboration platform is built based on a hybrid cloud architecture, which includes private cloud and public cloud. Data interaction and access control between private cloud and public cloud are realized through cloud gateway. Preprocessed data is classified and stored, and data index is established. A deep learning-based intelligent analysis model for garden data is constructed. The intelligent analysis model includes a plant growth status assessment sub-model, an early warning sub-model for pests and diseases, and a maintenance demand prediction sub-model. The intelligent analysis model is trained and optimized using historical garden data. Based on the output of the intelligent analysis model and combined with the garden management standards, a personalized maintenance plan is generated. The maintenance effect data is collected in real time for feedback, and the maintenance plan is dynamically adjusted. By constructing a multi-terminal collaborative interaction module, the hybrid cloud platform enables bidirectional data interaction with garden management terminals, mobile maintenance terminals, and public inquiry terminals, and protects the interactive data through an encrypted transmission protocol; Real-time monitoring of the hybrid cloud platform and intelligent analysis model; data encryption technology is used to encrypt and protect sensitive data throughout its entire lifecycle; and different operation permissions are assigned to different users. Data traceability logs are generated based on the entire process of information collection, preprocessing, storage, analysis and application. Historical data is archived and cleaned regularly, and iterative optimization is carried out in combination with the running effect of intelligent analysis models.

[0006] Preferably, the step of collecting multi-dimensional garden data through distributed acquisition terminals and performing data preprocessing, and adding category labels and spatiotemporal labels to different types of data based on garden data classification rules, specifically includes: The distributed data acquisition terminal includes an environmental sensing terminal, an image acquisition terminal, a manual data entry terminal, and an equipment operation terminal. The types of data collected cover environmental perception data, plant status data, maintenance operation data, and equipment operation data. Data cleaning algorithms are used to remove abnormal data and missing values. Image data, sensor data, and text data are uniformly converted into structured data format through format conversion. Based on the garden data classification rules, category labels and spatiotemporal labels are added to different types of data. The spatiotemporal labels include the collection timestamp and collection location coordinates.

[0007] Preferably, the garden data storage and collaboration platform built on a hybrid cloud architecture includes a private cloud and a public cloud. A cloud gateway enables data interaction and access control between the private and public clouds. The platform categorizes and stores pre-processed data and establishes a data index. Specifically, this includes: A garden data storage and collaboration platform is built based on a hybrid cloud architecture, which includes a private cloud and a public cloud. The private cloud stores sensitive maintenance data, plant germplasm resource data and core management data, while the public cloud stores massive amounts of environmental monitoring data, image data and non-sensitive shared data. Deploy a cloud gateway at the boundary between private and public clouds to perform protocol adaptation and data format conversion. By configuring access permissions, only authorized users can achieve cross-cloud data interaction through the cloud gateway. Create structured data partitions and unstructured data partitions, and store the preprocessed structured data and unstructured data into the corresponding partitions; An index mapping table is built based on the category labels and spatiotemporal labels of the data to enable rapid data location and retrieval.

[0008] Preferably, the construction of a deep learning-based intelligent analysis model for garden data includes a plant growth status assessment sub-model, an early warning sub-model for pests and diseases, and a maintenance requirement prediction sub-model. Training and optimizing the intelligent analysis model using historical garden data specifically includes: The plant growth status assessment sub-model is based on a convolutional neural network to fuse and analyze plant image data and environmental perception data, and outputs a plant growth health score. The early warning sub-model for pests and diseases is based on a recurrent neural network to perform time-series analysis on historical pest and disease data, environmental data and plant status data, identify potential risks of pest and disease occurrence and output warning levels. The maintenance demand prediction sub-model uses the gradient boosting tree algorithm to perform correlation analysis on plant growth cycle data, environmental data and maintenance history data to predict maintenance projects and maintenance timing within a preset time period in the future. Historical garden data was retrieved from a private cloud to train three sub-models. The hyperparameters of the models were iteratively adjusted using cross-validation. The trained three sub-models were then integrated, and their overall performance was verified using a validation set.

[0009] Preferably, the output results based on the intelligent analysis model, combined with garden management standards, generate a personalized maintenance plan. The plan is then dynamically adjusted by collecting real-time maintenance effect data for feedback. Specifically, this includes: Based on the growth health score and pest and disease early warning level output by the intelligent analysis model, combined with garden management standards, human and equipment resource information, a personalized maintenance plan is generated. The personalized maintenance plan includes watering frequency, fertilizer type and amount, pruning time, pest and disease control measures and equipment scheduling plan. The system collects real-time data on changes in plant growth status and the execution of maintenance operations, adjusts the parameters of the intelligent analysis model, and iteratively updates the maintenance plan.

[0010] Preferably, the step of constructing a multi-terminal collaborative interaction module to enable bidirectional data interaction between the hybrid cloud platform and garden management terminals, mobile maintenance terminals, and public inquiry terminals, and protecting the interactive data through an encrypted transmission protocol, specifically includes: By constructing a multi-terminal collaborative interaction module, a two-way data flow channel between the platform and the terminal is established. When the platform pushes decision-making schemes and maintenance instructions to the terminal, a push service protocol is used, and when the terminal uploads data to the platform, an incremental upload mechanism is used. An identity authentication mechanism is used for all terminal access, which uses dual verification through the device's unique identifier and the user's account; data transmission uses the SSL / TLS encryption protocol and performs CRC verification on the transmitted data; The terminals include garden management terminals, mobile maintenance terminals, and public inquiry terminals.

[0011] Preferably, the real-time monitoring of the hybrid cloud platform and intelligent analysis model, the use of data encryption technology to encrypt and protect sensitive data throughout its entire lifecycle, and the allocation of different operating permissions to different users specifically include: Real-time monitoring of hybrid cloud platforms and intelligent analysis models; collection of server CPU utilization, memory usage, storage capacity hardware parameters, model inference latency, data transmission rate performance indicators. During the data transmission phase, SSL / TLS encryption protocol is used; during the storage phase, AES-256 symmetric encryption algorithm is used for the core data of the private cloud; and during the data usage phase, decryption and access are achieved through an encryption sandbox. By dividing users into administrators, maintenance supervisors, frontline maintenance personnel, and public users, differentiated operation permissions are assigned to each role, and all users' login, data access, and operation behaviors are recorded.

[0012] Preferably, the process of generating data traceability logs based on information from the entire process of collection, preprocessing, storage, analysis to application, regularly archiving and cleaning historical data, and iteratively optimizing the data in conjunction with the performance of the intelligent analysis model specifically includes: Based on information from collection, preprocessing, storage, analysis to application, standardized traceability logs are generated. The logs include the operation time, operation subject, data source, data status and change records for each stage. Inactive historical data exceeding 3 years is categorized and archived to private cloud offline storage nodes, and compressed storage technology is used to save space. Based on the feedback from the intelligent analysis model, the completeness and accuracy of the existing collected data are evaluated, and the deployment location and collection frequency of the distributed collection terminals are adjusted. The training set of the model is expanded based on archived historical data, and the parameters of the intelligent analysis model are iteratively adjusted by combining maintenance effect data and plant growth status data.

[0013] Furthermore, a cloud-based intelligent management system for garden data is proposed, including: Hybrid cloud storage and collaborative management module: Core sensitive data is stored in the private cloud, non-sensitive data is processed in the public cloud, and cross-cloud secure interaction and classification index construction are achieved using the cloud gateway; Intelligent analysis model module: Deploys three AI sub-models for plant growth assessment, pest and disease early warning, and maintenance prediction, which are trained and optimized based on historical data; Maintenance plan generation module: Combines model output with landscape standards to generate personalized maintenance plans, and dynamically adjusts plan parameters through real-time feedback data; Multi-terminal collaborative interaction module: Supports two-way data interaction between management terminal, maintenance terminal and public terminal, and uses encrypted transmission and identity authentication to ensure communication security; Monitoring and Access Control Module: Monitors platform operation status in real time, implements full lifecycle data encryption, and assigns operation permissions based on roles; Data traceability and optimization module: records the entire process operation log, archives cold data regularly, and optimizes models and collection strategies using historical data; Processor: The processor is used to handle the calculation process of each formula and the construction calculation process of each model.

[0014] Compared with the prior art, the advantages of the present invention are: Leveraging a distributed data acquisition and preprocessing mechanism to ensure data accuracy and standardization, a hybrid cloud architecture achieves a balance between secure storage of sensitive data and efficient handling of massive amounts of data. Through multi-dimensional deep learning models, it accurately assesses plant growth, provides early warnings of pests and diseases, and predicts maintenance needs. Combined with standardized data, it generates personalized solutions and dynamically optimizes them, significantly improving the scientific rigor and targeted nature of maintenance. Multi-terminal collaborative interaction and full lifecycle encryption protection balance management efficiency and data security. A full-process traceability log and iterative optimization mechanism help continuously improve data acquisition strategies and model performance, forming a closed loop of "collection-analysis-application-optimization." This drives the transformation of garden management from experience-driven to data-intelligent driven, improving overall management efficiency and precision. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the method proposed in this invention; Figure 2 This is a schematic diagram of the data acquisition and preprocessing proposed in this invention; Figure 3 This is a schematic diagram of the hybrid cloud architecture platform proposed in this invention; Figure 4 This is a schematic diagram of the intelligent analysis model for garden data proposed in this invention; Figure 5 This is a schematic diagram of the intelligent garden management decision-making proposed in this invention; Figure 6 This is a schematic diagram of the multi-terminal collaborative interaction proposed in this invention; Figure 7 This is a schematic diagram illustrating the cloud platform operation and security assurance proposed in this invention; Figure 8 This is a schematic diagram of the full lifecycle management of garden data proposed in this invention. Detailed Implementation

[0016] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0017] A cloud-based intelligent management system for garden data includes: Hybrid cloud storage and collaborative management module: Core sensitive data is stored in the private cloud, non-sensitive data is processed in the public cloud, and cross-cloud secure interaction and classification index construction are achieved using the cloud gateway; Intelligent analysis model module: Deploys three AI sub-models for plant growth assessment, pest and disease early warning, and maintenance prediction, which are trained and optimized based on historical data; Maintenance plan generation module: Combines model output with landscape standards to generate personalized maintenance plans, and dynamically adjusts plan parameters through real-time feedback data; Multi-terminal collaborative interaction module: Supports two-way data interaction between management terminal, maintenance terminal and public terminal, and uses encrypted transmission and identity authentication to ensure communication security; Monitoring and Access Control Module: Monitors platform operation status in real time, implements full lifecycle data encryption, and assigns operation permissions based on roles; Data traceability and optimization module: records the entire process operation log, archives cold data regularly, and optimizes models and collection strategies using historical data; Processor: The processor is used to handle the calculation process of each formula and the construction calculation process of each model.

[0018] See Figure 1 As shown, the cloud-based intelligent management method for garden data includes: Step 1: Collect multi-dimensional data of the garden through distributed acquisition terminals and perform data preprocessing. Add category labels and spatiotemporal labels to different types of data based on garden data classification rules. Step 2: Build a garden data storage and collaboration platform based on a hybrid cloud architecture. The hybrid cloud architecture includes private cloud and public cloud. Data interaction and access control between private cloud and public cloud are realized through cloud gateway. The pre-processed data is classified and stored, and a data index is established. Step 3: Construct a deep learning-based intelligent analysis model for garden data. The intelligent analysis model includes a plant growth status assessment sub-model, an early warning sub-model for pests and diseases, and a maintenance demand prediction sub-model. The intelligent analysis model is trained and optimized using historical garden data. Step 4: Based on the output of the intelligent analysis model and combined with the garden management standards, generate a personalized maintenance plan. Feedback is provided by collecting maintenance effect data in real time, and the maintenance plan is dynamically adjusted. Step 5: By constructing a multi-terminal collaborative interaction module, the hybrid cloud platform enables bidirectional data interaction with garden management terminals, mobile maintenance terminals, and public inquiry terminals, and protects the interactive data through an encrypted transmission protocol; Step Six: Monitor the hybrid cloud platform and intelligent analysis model in real time, use data encryption technology to encrypt and protect sensitive data throughout its entire lifecycle, and assign different operation permissions to different users; Step 7: Generate a data traceability log based on the entire process of information collection, preprocessing, storage, analysis and application, archive and clean up historical data regularly, and iterate and optimize the data in combination with the performance of the intelligent analysis model.

[0019] See Figure 2 As shown, multi-dimensional data of the garden is collected and preprocessed through distributed acquisition terminals. Based on the garden data classification rules, category labels and spatiotemporal labels are added to different types of data, specifically including: The distributed data acquisition terminal includes an environmental sensing terminal, an image acquisition terminal, a manual data entry terminal, and an equipment operation terminal. The types of data collected cover environmental perception data, plant status data, maintenance operation data, and equipment operation data. Data cleaning algorithms are used to remove abnormal data and missing values. Image data, sensor data, and text data are uniformly converted into structured data format through format conversion. Based on the garden data classification rules, category labels and spatiotemporal labels are added to different types of data. The spatiotemporal labels include the collection timestamp and collection location coordinates.

[0020] See Figure 3 As shown, a garden data storage and collaboration platform is built based on a hybrid cloud architecture. This architecture includes both private and public clouds. A cloud gateway enables data interaction and access control between the private and public clouds. Pre-processed data is categorized and stored, and a data index is established. Specifically, this includes: A garden data storage and collaboration platform is built based on a hybrid cloud architecture, which includes a private cloud and a public cloud. The private cloud stores sensitive maintenance data, plant germplasm resource data and core management data, while the public cloud stores massive amounts of environmental monitoring data, image data and non-sensitive shared data. Deploy a cloud gateway at the boundary between private and public clouds to perform protocol adaptation and data format conversion. By configuring access permissions, only authorized users can achieve cross-cloud data interaction through the cloud gateway. Create structured data partitions and unstructured data partitions, and store the preprocessed structured data and unstructured data into the corresponding partitions; An index mapping table is built based on the category labels and spatiotemporal labels of the data to enable rapid data location and retrieval.

[0021] Specifically, the architecture is deployed as a "private cloud core + public cloud extension". The private cloud nodes are built using blade server clusters, configured with redundant storage arrays and local backup nodes, and equipped with a Linux operating system and virtualization management platform to achieve physical isolation storage of core data. On the public cloud side, a compatible public cloud service provider is selected to complete the elastic rental of computing and storage resources through cloud service interfaces, and build an elastically scalable cloud resource pool. A distributed file system is used to classify and index preprocessed structured and unstructured data. Indexes are built for structured data and feature indexes are built for unstructured image data. A cache server cluster is deployed to cache frequently accessed data, and a load balancing algorithm is used to distribute retrieval requests to improve the response speed of cross-cloud data retrieval.

[0022] See Figure 4 As shown, a deep learning-based intelligent analysis model for garden data is constructed. This model includes a plant growth status assessment sub-model, an early warning sub-model for pests and diseases, and a maintenance requirement prediction sub-model. The training and optimization of the intelligent analysis model using historical garden data specifically includes: The plant growth status assessment sub-model is based on a convolutional neural network to fuse and analyze plant image data and environmental perception data, and outputs a plant growth health score. The early warning sub-model for pests and diseases is based on a recurrent neural network to perform time-series analysis on historical pest and disease data, environmental data and plant status data, identify potential risks of pest and disease occurrence and output warning levels. The maintenance demand prediction sub-model uses the gradient boosting tree algorithm to perform correlation analysis on plant growth cycle data, environmental data and maintenance history data to predict maintenance projects and maintenance timing within a preset time period in the future. Historical garden data was retrieved from a private cloud to train three sub-models. The hyperparameters of the models were iteratively adjusted using cross-validation. The trained three sub-models were then integrated, and their overall performance was verified using a validation set.

[0023] Specifically, a "multi-sub-model parallelism + feature fusion" architecture is adopted, and the model training environment is built based on a private cloud GPU cluster. Among them, the plant growth status assessment sub-model adopts a residual network architecture to adapt to the deep extraction of plant image features; the early warning sub-model of pests and diseases adopts a long short-term memory network to adapt to the trend mining of time series data; and the maintenance demand prediction sub-model adopts a gradient boosting tree architecture to improve the accuracy of multi-feature association analysis. The collaborative calling of the three sub-models is realized through model interface encapsulation. Three sub-models were trained step-by-step on a private cloud GPU cluster. The plant growth status assessment sub-model takes preprocessed plant image features and environmental data as input, optimizes the convolution kernel parameters through backpropagation, and outputs a health score. The early warning sub-model for pests and diseases takes time-series environmental data and historical pest and disease records as input, optimizes the model by adjusting the number of hidden layer nodes in the LSTM network, and outputs three warning levels: low, medium, and high. The maintenance demand prediction sub-model takes growth cycle, environmental parameters, and other features as input, suppresses overfitting through regularization, and predicts maintenance projects and optimal timing for the next 1-3 months.

[0024] See Figure 5 As shown, based on the output of the intelligent analysis model and combined with garden management standards, a personalized maintenance plan is generated. Feedback is provided through real-time collection of maintenance effect data, allowing for dynamic adjustments to the maintenance plan. Specifically, this includes: Based on the growth health score and pest and disease early warning level output by the intelligent analysis model, combined with garden management standards, human and equipment resource information, a personalized maintenance plan is generated. The personalized maintenance plan includes watering frequency, fertilizer type and amount, pruning time, pest and disease control measures and equipment scheduling plan. The system collects real-time data on changes in plant growth status and the execution of maintenance operations, adjusts the parameters of the intelligent analysis model, and iteratively updates the maintenance plan.

[0025] Specifically, first, the basic maintenance standards corresponding to the plant type and growth stage in the maintenance rule library are matched; second, the maintenance parameters are adjusted in combination with environmental perception data; then, targeted prevention and control measures are matched according to the pest and disease warning level; finally, the time window and resource allocation plan for maintenance execution are determined in combination with the personnel configuration, equipment distribution and material inventory information fed back by the resource scheduling module, and a personalized maintenance plan including operation details, responsible persons and time nodes is generated. A feedback collection channel for maintenance effectiveness is established, using mobile maintenance terminals to collect maintenance execution data and environmental change data in real time, and then feeding this data back to the decision support module. The decision engine compares the actual maintenance effect with the expected effect. If the improvement in plant health does not meet expectations or the pest and disease warning is not lifted, the solution optimization process is triggered, maintenance parameters are adjusted, the optimized solution is synchronized to relevant terminals, and the solution iteration log is recorded.

[0026] See Figure 6 As shown, by constructing a multi-terminal collaborative interaction module, the hybrid cloud platform enables bidirectional data interaction with garden management terminals, mobile maintenance terminals, and public inquiry terminals. Furthermore, the interaction data is protected through an encrypted transmission protocol, specifically including: By constructing a multi-terminal collaborative interaction module, a two-way data flow channel between the platform and the terminal is established. When the platform pushes decision-making schemes and maintenance instructions to the terminal, a push service protocol is used, and when the terminal uploads data to the platform, an incremental upload mechanism is used. An identity authentication mechanism is used for all terminal access, which uses dual verification through the device's unique identifier and the user's account; data transmission uses the SSL / TLS encryption protocol and performs CRC verification on the transmitted data; The terminals include garden management terminals, mobile maintenance terminals, and public inquiry terminals.

[0027] Specifically, customized adaptation and functional development were completed for three types of core terminals: Garden Management Terminal: Deployed in the management center, configured with a visual management interface, supporting the receipt, review, and modification of intelligent decision-making schemes and the batch issuance of maintenance instructions, and possessing data statistical report generation and anomaly warning functions; Mobile Maintenance Terminal: Developed with a dedicated APP based on Android and HarmonyOS systems, supporting maintenance personnel to upload maintenance execution photos, operation records, and on-site problem feedback in real time, integrating GPS positioning function to achieve maintenance trajectory tracking, and offline caching function to ensure data storage and subsequent synchronization in environments without network access; Public Inquiry Terminal: Deployed at garden entrances and core areas, using touchscreen interaction, supporting inquiries about non-sensitive data such as garden opening hours, plant distribution, and popular science knowledge, and possessing information push function to publish temporary control or activity notices; The TLS 1.3 encryption protocol is used to achieve end-to-end encrypted data transmission between the cloud and the terminal. The transmitted data is segmented and verified for integrity to prevent data tampering or loss. Each terminal is assigned a unique device identifier and dynamic key, and the device authentication mechanism prevents unauthorized terminal access. The mobile maintenance terminal adopts dual authentication of SIM card encryption and device binding to ensure the legitimacy of the data interaction subject.

[0028] See Figure 7 As shown, real-time monitoring of the hybrid cloud platform and intelligent analysis model is implemented. Data encryption technology is used to protect sensitive data throughout its entire lifecycle, and different operation permissions are assigned to different users. Specifically, this includes: Real-time monitoring of hybrid cloud platforms and intelligent analysis models; collection of server CPU utilization, memory usage, storage capacity hardware parameters, model inference latency, data transmission rate performance indicators. During the data transmission phase, SSL / TLS encryption protocol is used; during the storage phase, AES-256 symmetric encryption algorithm is used for the core data of the private cloud; and during the data usage phase, decryption and access are achieved through an encryption sandbox. By dividing users into administrators, maintenance supervisors, frontline maintenance personnel, and public users, differentiated operation permissions are assigned to each role, and all users' login, data access, and operation behaviors are recorded.

[0029] Specifically, a distributed monitoring architecture is adopted, deploying monitoring agents on private cloud, public cloud nodes, and cloud gateways to collect core operational metrics in real time. The monitoring scope covers three dimensions: infrastructure monitoring, including server CPU utilization, memory usage, disk I / O, and network bandwidth, ensuring hardware stability through threshold alarm mechanisms; data layer monitoring, real-time monitoring of hybrid cloud storage capacity, data read / write speed, and index integrity, triggering alerts for anomalies such as insufficient storage capacity and data read / write latency exceeding 1 second; and application layer monitoring, tracking the operating efficiency of intelligent analysis models, interface response time, and monitoring the data transmission status of multi-terminal interactions. The formula for predicting hybrid cloud storage capacity is:

[0030] in, This represents the expected storage capacity requirement after time t. This represents the total storage capacity currently used by the hybrid cloud. This represents the average daily data growth rate. To predict time periods, For the planned new storage demand within the next t time period; A tiered encryption strategy is implemented, with sensitive data stored and encrypted using the AES-256 encryption algorithm, and transmitted using the TLS 1.3 protocol. Non-sensitive data is encrypted using a lightweight encryption algorithm to ensure transmission security. A role-based access control system is constructed, dividing access into four levels: administrator, operations and maintenance personnel, maintenance personnel, and the public. Each role is assigned exclusive operation permissions and data access scope, and identity authentication is strengthened through dynamic keys and regular password update mechanisms.

[0031] See Figure 8 As shown, a data traceability log is generated based on the entire process of information collection, preprocessing, storage, analysis, and application. Historical data is archived and cleaned regularly, and iterative optimization is carried out in conjunction with the running effect of the intelligent analysis model. Specifically, this includes: Based on information from collection, preprocessing, storage, analysis to application, standardized traceability logs are generated. The logs include the operation time, operation subject, data source, data status and change records for each stage. Inactive historical data exceeding 3 years is categorized and archived to private cloud offline storage nodes, and compressed storage technology is used to save space. Based on the feedback from the intelligent analysis model, the completeness and accuracy of the existing collected data are evaluated, and the deployment location and collection frequency of the distributed collection terminals are adjusted. The training set of the model is expanded based on archived historical data, and the parameters of the intelligent analysis model are iteratively adjusted by combining maintenance effect data and plant growth status data.

[0032] Specifically, an immutable data traceability module is constructed using distributed ledger technology and deployed on a core node of a private cloud to record data from collection, preprocessing, storage, analysis to application, and all related process nodes. The traceability log includes data source, processing personnel, timestamp, data flow path, and operation content. A traceability index is established by associating spatiotemporal tags and category tags with the data, supporting precise retrieval of traceability information by data type, time range, collection area, and other dimensions. Based on the traceability analysis results of the whole life cycle data, the effectiveness of the existing data collection system is evaluated. If the accuracy of environmental data collection in a certain area is insufficient or the missing rate of plant status data exceeds 5%, the deployment density of collection terminals, sensor parameters or image collection angle in that area are adjusted. For the intelligent analysis model, incremental training is carried out every six months by calling archived historical data and newly added real-time data. The model parameters are corrected by combining maintenance effect feedback data, the feature extraction dimension is optimized, and the model's adaptability in complex scenarios is improved.

[0033] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0034] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0035] 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 principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A cloud-based intelligent management method for garden data, characterized in that: include: Multi-dimensional data of the garden is collected through distributed acquisition terminals and preprocessed. Based on the garden data classification rules, category labels and spatiotemporal labels are added to different types of data. A garden data storage and collaboration platform is built based on a hybrid cloud architecture, which includes private cloud and public cloud. Data interaction and access control between private cloud and public cloud are realized through cloud gateway. Preprocessed data is classified and stored, and data index is established. A deep learning-based intelligent analysis model for garden data is constructed. The intelligent analysis model includes a plant growth status assessment sub-model, an early warning sub-model for pests and diseases, and a maintenance demand prediction sub-model. The intelligent analysis model is trained and optimized using historical garden data. Based on the output of the intelligent analysis model and combined with the garden management standards, a personalized maintenance plan is generated. The maintenance effect data is collected in real time for feedback, and the maintenance plan is dynamically adjusted. By constructing a multi-terminal collaborative interaction module, the hybrid cloud platform enables bidirectional data interaction with garden management terminals, mobile maintenance terminals, and public inquiry terminals, and protects the interactive data through an encrypted transmission protocol; Real-time monitoring of the hybrid cloud platform and intelligent analysis model; data encryption technology is used to encrypt and protect sensitive data throughout its entire lifecycle; and different operation permissions are assigned to different users. Data traceability logs are generated based on the entire process of information collection, preprocessing, storage, analysis and application. Historical data is archived and cleaned regularly, and iterative optimization is carried out in combination with the running effect of intelligent analysis models.

2. The cloud computing-based intelligent management method for garden data according to claim 1, characterized in that, The process of collecting multi-dimensional garden data through distributed acquisition terminals and performing data preprocessing, and adding category labels and spatiotemporal labels to different types of data based on garden data classification rules, specifically includes: The distributed data acquisition terminal includes an environmental sensing terminal, an image acquisition terminal, a manual data entry terminal, and an equipment operation terminal. The types of data collected cover environmental perception data, plant status data, maintenance operation data, and equipment operation data. Data cleaning algorithms are used to remove abnormal data and missing values. Image data, sensor data, and text data are uniformly converted into structured data format through format conversion. Based on the garden data classification rules, category labels and spatiotemporal labels are added to different types of data. The spatiotemporal labels include the collection timestamp and collection location coordinates.

3. The cloud computing-based intelligent management method for garden data according to claim 1, characterized in that, The aforementioned garden data storage and collaboration platform, built on a hybrid cloud architecture, includes both private and public clouds. A cloud gateway enables data interaction and access control between the private and public clouds. Preprocessed data is categorized and stored, and a data index is established. Specifically, this includes: A garden data storage and collaboration platform is built based on a hybrid cloud architecture, which includes a private cloud and a public cloud. The private cloud stores sensitive maintenance data, plant germplasm resource data and core management data, while the public cloud stores massive amounts of environmental monitoring data, image data and non-sensitive shared data. Deploy a cloud gateway at the boundary between private and public clouds to perform protocol adaptation and data format conversion. By configuring access permissions, only authorized users can achieve cross-cloud data interaction through the cloud gateway. Create structured data partitions and unstructured data partitions, and store the preprocessed structured data and unstructured data into the corresponding partitions; An index mapping table is built based on the category labels and spatiotemporal labels of the data to enable rapid data location and retrieval.

4. The cloud computing-based intelligent management method for garden data according to claim 1, characterized in that, The construction of a deep learning-based intelligent analysis model for garden data includes a plant growth status assessment sub-model, an early warning sub-model for pests and diseases, and a maintenance demand prediction sub-model. The training and optimization of the intelligent analysis model using historical garden data specifically includes: The plant growth status assessment sub-model is based on a convolutional neural network to fuse and analyze plant image data and environmental perception data, and outputs a plant growth health score. The early warning sub-model for pests and diseases is based on a recurrent neural network to perform time-series analysis on historical pest and disease data, environmental data and plant status data, identify potential risks of pest and disease occurrence and output warning levels. The maintenance demand prediction sub-model uses the gradient boosting tree algorithm to perform correlation analysis on plant growth cycle data, environmental data and maintenance history data to predict maintenance projects and maintenance timing within a preset time period in the future. Historical garden data was retrieved from a private cloud to train three sub-models. The hyperparameters of the models were iteratively adjusted using cross-validation. The trained three sub-models were then integrated, and their overall performance was verified using a validation set.

5. The cloud computing-based intelligent management method for garden data according to claim 1, characterized in that, The output of the intelligent analysis model, combined with garden management standards, generates personalized maintenance plans. Feedback is provided through real-time collection of maintenance effect data, allowing for dynamic adjustments to the maintenance plans. Specifically, this includes: Based on the growth health score and pest and disease early warning level output by the intelligent analysis model, combined with garden management standards, human and equipment resource information, a personalized maintenance plan is generated. The personalized maintenance plan includes watering frequency, fertilizer type and amount, pruning time, pest and disease control measures and equipment scheduling plan. The system collects real-time data on changes in plant growth status and the execution of maintenance operations, adjusts the parameters of the intelligent analysis model, and iteratively updates the maintenance plan.

6. The cloud computing-based intelligent management method for garden data according to claim 1, characterized in that, The construction of a multi-terminal collaborative interaction module enables bidirectional data interaction between the hybrid cloud platform and garden management terminals, mobile maintenance terminals, and public inquiry terminals, and protects the interactive data through an encrypted transmission protocol. Specifically, this includes: By constructing a multi-terminal collaborative interaction module, a two-way data flow channel between the platform and the terminal is established. When the platform pushes decision-making schemes and maintenance instructions to the terminal, a push service protocol is used, and when the terminal uploads data to the platform, an incremental upload mechanism is used. An identity authentication mechanism is used for all terminal access, which uses dual verification through the device's unique identifier and the user's account; data transmission uses the SSL / TLS encryption protocol and performs CRC verification on the transmitted data; The terminals include garden management terminals, mobile maintenance terminals, and public inquiry terminals.

7. The cloud computing-based intelligent management method for garden data according to claim 1, characterized in that, The real-time monitoring of the hybrid cloud platform and intelligent analysis model, the use of data encryption technology to protect sensitive data throughout its entire lifecycle, and the allocation of different operating permissions to different users specifically include: Real-time monitoring of hybrid cloud platforms and intelligent analysis models; collection of server CPU utilization, memory usage, storage capacity hardware parameters, model inference latency, data transmission rate performance indicators. During the data transmission phase, SSL / TLS encryption protocol is used; during the storage phase, AES-256 symmetric encryption algorithm is used for the core data of the private cloud; and during the data usage phase, decryption and access are achieved through an encryption sandbox. By dividing users into administrators, maintenance supervisors, frontline maintenance personnel, and public users, differentiated operation permissions are assigned to each role, and all users' login, data access, and operation behaviors are recorded.

8. The cloud computing-based intelligent management method for garden data according to claim 1, characterized in that, The process of generating data traceability logs based on the entire process of information collection, preprocessing, storage, analysis, and application, regularly archiving and cleaning historical data, and iteratively optimizing the data based on the performance of the intelligent analysis model specifically includes: Based on information from collection, preprocessing, storage, analysis to application, standardized traceability logs are generated. The logs include the operation time, operation subject, data source, data status and change records for each stage. Inactive historical data exceeding 3 years is categorized and archived to private cloud offline storage nodes, and compressed storage technology is used to save space. Based on the feedback from the intelligent analysis model, the completeness and accuracy of the existing collected data are evaluated, and the deployment location and collection frequency of the distributed collection terminals are adjusted. The training set of the model is expanded based on archived historical data, and the parameters of the intelligent analysis model are iteratively adjusted by combining maintenance effect data and plant growth status data.

9. A cloud-based intelligent management system for garden data, used to implement the cloud-based intelligent management method for garden data as described in any one of claims 1-8, characterized in that, include: Hybrid cloud storage and collaborative management module: Core sensitive data is stored in the private cloud, non-sensitive data is processed in the public cloud, and cross-cloud secure interaction and classification index construction are achieved using the cloud gateway; Intelligent analysis model module: Deploys three AI sub-models for plant growth assessment, pest and disease early warning, and maintenance prediction, which are trained and optimized based on historical data; Maintenance plan generation module: Combines model output with landscape standards to generate personalized maintenance plans, and dynamically adjusts plan parameters through real-time feedback data; Multi-terminal collaborative interaction module: Supports two-way data interaction between management terminal, maintenance terminal and public terminal, and uses encrypted transmission and identity authentication to ensure communication security; Monitoring and Access Control Module: Monitors platform operation status in real time, implements full lifecycle data encryption, and assigns operation permissions based on roles; Data traceability and optimization module: records the entire process operation log, archives cold data regularly, and optimizes models and collection strategies using historical data; Processor: The processor is used to handle the calculation process of each formula and the construction calculation process of each model.

Images