A digital and intelligent management system and method for landscaping maintenance

Abstract

The application discloses a kind of digitalization management system and method for garden greening maintenance, it is related to data processing technical field, the system includes: data perception module, for obtaining M garden area information, obtains M garden state data stream;Parallel analysis module is used to carry out parallel analysis to M garden state data stream, and output M garden growth analysis result;Strategy analysis module is used to obtain garden greening maintenance strategy library, determines M garden greening maintenance strategy parameter;Digitalization control module is used to match activation, obtains garden linkage maintenance equipment set, executes greening maintenance remote control.This application solves the technical problems of resource waste and management lag caused by extensive area division and poor linkage of maintenance strategy in traditional garden management in the prior art, achieves precise intelligent control of garden maintenance by using big data collection and processing, and improves the technical effects of garden greening maintenance efficiency and real-time pest control.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a digital management system and method for garden greening maintenance. BACKGROUND

[0002] With the acceleration of urbanization, the contradiction between the expansion of garden greening scale and the fine management of ecology is increasingly prominent. Traditional garden maintenance mainly relies on manual patrol and experience-based decision-making, which has problems such as fuzzy regional division, delayed pest warning, and waste of water and fertilizer resources. In recent years, although technologies such as Internet of Things sensing and unmanned aerial vehicle inspection have been introduced, existing solutions mostly use a single data source or a fixed threshold analysis model, resulting in poor regional adaptability, fragmented processing of multi-dimensional environmental parameters (such as soil moisture, vegetation growth, and pest characteristics), and difficulty in achieving dynamic and accurate maintenance. For example, extensive zoning can cause uneven irrigation coverage, pest identification relies on manual photo uploading, resulting in delayed response, and the maintenance strategy is disconnected from the equipment execution link, further reducing resource utilization efficiency.

[0003] The existing technology has the technical problems of resource waste and management lag caused by extensive regional division and poor linkage of maintenance strategies in traditional garden management. SUMMARY

[0004] The present application provides a digital management system and method for garden greening maintenance, which is used to solve the technical problems of resource waste and management lag caused by extensive regional division and poor linkage of maintenance strategies in traditional garden management in the prior art.

[0005] In view of the above problems, the present application provides a digital management system and method for garden greening maintenance.

[0006] In a first aspect of the present application, a digital management system for garden greening maintenance is provided, which comprises:

[0007] The data perception module is used for regionally dividing and encoding a target garden to obtain M garden region information, and a perception network deployment collects and obtains M garden state data streams of the M garden region information; the parallel analysis module is used for constructing a garden analysis double channel, the garden analysis double channel including a garden greening prediction channel and a pest and disease identification channel, and the M garden state data streams are analyzed in parallel by using the garden analysis double channel to output M garden growth analysis results; the strategy analysis module is used for obtaining a garden greening maintenance strategy library, and the M garden growth analysis results are analyzed and associatedly corrected by using the garden greening maintenance strategy library to determine M garden greening maintenance strategy parameters; and the digitalization management module is used for matching and activating the M garden greening maintenance strategy parameters with a garden maintenance equipment list to obtain a garden linkage maintenance equipment set, and the target garden is executed for greening maintenance remote management by using the garden linkage maintenance equipment set.

[0008] In a second aspect of the present application, a digital management method for garden greening maintenance is provided, and the method comprises:

[0009] The target garden is regionally divided and encoded to obtain M garden region information, and a perception network deployment collects and obtains M garden state data streams of the M garden region information; a garden analysis double channel is constructed, the garden analysis double channel including a garden greening prediction channel and a pest and disease identification channel, and the M garden state data streams are analyzed in parallel by using the garden analysis double channel to output M garden growth analysis results; a garden greening maintenance strategy library is obtained, and the M garden growth analysis results are analyzed and associatedly corrected by using the garden greening maintenance strategy library to determine M garden greening maintenance strategy parameters; and the M garden greening maintenance strategy parameters are matched and activated with a garden maintenance equipment list to obtain a garden linkage maintenance equipment set, and the target garden is executed for greening maintenance remote management by using the garden linkage maintenance equipment set.

[0010] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0011] The data perception module is used for region division coding of a target garden to obtain M garden region information, and M garden state data streams of the M garden region information are collected and acquired; the parallel analysis module is used for constructing a garden analysis double channel, performing parallel analysis, and outputting M garden growth analysis results; the strategy analysis module is used for acquiring a garden greening maintenance strategy library, performing strategy analysis and correlation correction on the M garden growth analysis results, and determining M garden greening maintenance strategy parameters; and the digital intelligent control module is used for matching and activating based on the M garden greening maintenance strategy parameters and a garden maintenance equipment list, obtaining a garden linkage maintenance equipment set, and performing remote greening maintenance control on the target garden. The technical effect of realizing precise intelligent control of garden maintenance by using big data collection and processing, and improving garden greening maintenance efficiency and real-time pest control is achieved. BRIEF DESCRIPTION OF DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0013] Figure 1 A structural schematic diagram of a digital intelligent management system for garden greening maintenance provided by the embodiment of the present application;

[0014] Figure 2 A flowchart of a digital intelligent management method for garden greening maintenance provided by the embodiment of the present application.

[0015] The reference signs are explained as follows: data perception module 10, parallel analysis module 20, strategy analysis module 30, and digital intelligent control module 40. DETAILED DESCRIPTION

[0016] The present application provides a digital intelligent management system and method for garden greening maintenance, which is used to solve the technical problems of resource waste and lag in management and maintenance caused by extensive region division and poor linkage of maintenance strategies in traditional garden management in the prior art.

[0017] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0018] Embodiment one, as Figure 1As shown, the present application provides a digital management system for garden maintenance, the system comprises:

[0019] A data perception module 10 is used to divide and encode the target garden into M garden area information, and the M garden state data streams of the M garden area information are collected and obtained by the perception network deployment.

[0020] Specifically, first, according to various factors such as regional function, geographical features, and regional area, the target garden is comprehensively and meticulously divided into regions. These division factors each have their importance, and they are prioritized, and then combined with the design survey data of the target garden, the regional multi-level division work is carried out, and finally M different garden regions are obtained. In order to facilitate management and data correspondence, a special coding system is designed for each region, including coding content, order and identification, the coding annotation of the M garden regions is completed, thereby generating M garden area information. Then, the perception network is deployed according to these regional information, and real-time data of each region is collected by using various sensors, monitoring equipment, etc., forming M garden state data streams reflecting the growth trend of garden plants, soil environment, meteorological conditions and other conditions, providing strong data support for subsequent accurate garden maintenance analysis and decision-making.

[0021] A parallel analysis module 20 is used to construct a garden analysis double channel, which includes a garden greening prediction channel and a pest and disease identification channel, and the M garden state data streams are analyzed in parallel by using the garden analysis double channel to output M garden growth analysis results.

[0022] Specifically, the parallel analysis module 20 constructs an efficient garden analysis double channel system, which is composed of a garden greening prediction channel and a pest and disease identification channel. In the construction process, a large number of garden greening historical data sets and garden pest and disease historical data sets are collected, and through in-depth analysis of the garden greening prediction target and the pest and disease identification target, the corresponding key indicators are extracted to form a garden greening prediction index set and a pest and disease identification index set. These index sets are used to classify and identify the corresponding historical data sets for training, thereby obtaining the garden greening prediction channel and the pest and disease identification channel, and then the two are integrated in parallel. When receiving the M garden state data streams transmitted by the data perception module, the double channel simultaneously analyzes these data streams quickly and independently, starting from multiple dimensions such as plant growth trend prediction, early diagnosis of pests and diseases, etc., and outputs M accurate garden growth analysis results, providing a scientific basis for subsequent maintenance decisions.

[0023] A strategy analysis module 30 is used to obtain a garden greening maintenance strategy library, and the M garden growth analysis results are analyzed and associated corrected by using the garden greening maintenance strategy library to determine M garden greening maintenance strategy parameters.

[0024] Specifically, the strategy analysis module 30 first accesses and acquires a pre-constructed landscaping maintenance strategy library. This strategy library is a knowledge collection based on a large number of landscaping maintenance experiences, scientific researches, and industry standards, covering a wide variety of maintenance strategies for different landscaping growth conditions, plant species characteristics, seasonal changes, and other situations. The M landscaping growth analysis results output by the parallel analysis module are taken as input basis. For the growth analysis result of each landscaping area, a comprehensive and detailed search is performed in the strategy library to filter out the maintenance strategies that match the actual situation of the area. In this process, it comprehensively considers factors such as the growth stage of the plant, the type and severity of the disease and pest, the soil fertility condition, and the climate condition. After filtering out the initially applicable strategies, because the areas in the landscaping are not completely independent, there may be ecological relations, environmental influences, and other interactions between them, and the actual maintenance scene is complex and variable, a single strategy may not be able to fully meet the needs. Combining the characteristics of each area and the relationship between them, the initially filtered strategies are adjusted and optimized, and the priority and synergy between different strategies are weighed. After the above strategy analysis and correlation correction process, M landscaping maintenance strategy parameters are finally determined. These parameters accurately specify the detailed standards and execution plans for specific maintenance operations such as watering, fertilizing, pruning, and disease and pest control in each landscaping area, providing key decision support for the subsequent digital management and control module to drive the landscaping maintenance equipment to carry out precise maintenance work, ensuring that the entire landscaping maintenance work is scientifically and efficiently promoted.

[0025] The digital management and control module 40 is used to match and activate the M landscaping maintenance strategy parameters with a list of landscaping maintenance equipment to obtain a set of landscaping linked maintenance equipment, and perform remote management and control of landscaping maintenance on the target garden through the set of landscaping linked maintenance equipment.

[0026] Specifically, the digital management and control module 40 first receives the determined M landscaping maintenance strategy parameters, which specify the required maintenance operations and standards for each landscaping area in detail. At the same time, a list of landscaping maintenance equipment is obtained, which covers detailed information of various equipment such as irrigation equipment, pruning machinery, and disease and pest control devices.

[0027] The matching and activation work is started, and the maintenance strategy parameters of each landscaping area are accurately compared with the attributes such as device functions and application ranges in the equipment list. For example, for areas that require a lot of watering, irrigation equipment with flow and coverage that meet the requirements is selected; for areas with diseases and pests, corresponding disease and pest control equipment is matched. Through this precise matching, a series of applicable equipment is activated to form a set of landscaping linked maintenance equipment.

[0028] Finally, with the help of Internet of Things and automation control technology, the target garden is executed remote control of green maintenance through the collection of garden linkage maintenance equipment. In the process of remote control, accurate control instructions are sent to each device according to the maintenance strategy parameters, such as the opening time, watering time and water volume of the irrigation equipment, the pruning equipment according to the specific shape and size, and the pesticide spraying dose and frequency of the pest control equipment. In this way, efficient and accurate maintenance of the entire target garden is realized, ensuring the healthy growth of garden plants and improving the landscape effect of the garden.

[0029] In one possible implementation, the data perception module 10 further includes:

[0030] A garden area division factor acquisition unit is configured to acquire garden area division factors, including area function, geographical features, and area size.

[0031] A garden division factor priority sequence acquisition unit is configured to prioritize the garden area division factors according to the garden area division factors, and obtain a garden division factor priority sequence.

[0032] A garden division area acquisition unit is configured to divide the design survey data of the target garden into multiple levels based on the garden division factor priority sequence, and obtain M garden division areas.

[0033] A region coding system design unit is configured to design a region coding system according to the garden division factor priority sequence, including coding content, coding sequence, and coding identification.

[0034] A garden area information acquisition unit is configured to code and label the M garden division areas according to the region coding system, and obtain M garden area information.

[0035] Specifically, it is crucial to obtain comprehensive and accurate garden area division factors. Among them, the area function is an important division basis, and different functional areas have different planning and maintenance priorities. Geographical features cover topography, such as the drainage and irrigation methods of mountainous and plain areas, which affect the plant growth environment. Soil type also determines the types of plants suitable for planting, with acid soil suitable for some plants and alkaline soil suitable for others. The area size is related to the scale of maintenance resources, and large areas require more manpower, material resources and time for management, and the area factor needs to be considered when planning the irrigation system and pest control range. Combining these area function, geographical features and area size factors, targeted maintenance strategies are formulated for subsequent reasonable division of garden areas.

[0036] After obtaining the garden area division factors, these factors are prioritized to obtain a garden division factor priority sequence. First, each factor is studied in depth to consider its impact on the overall planning of the garden, plant growth, and maintenance management. For example, the area function determines the core purpose of the garden. If it is an ornamental area, the selection and layout of plants will focus on aesthetics. The geographical features fundamentally limit the growth conditions of plants, and different terrain and soil types will affect the survival and growth state of plants. The area size affects the scale and efficiency of resource allocation. Through a comprehensive evaluation of these factors, the weight of each factor is determined using the analytic hierarchy process. According to the weight, the garden area division factors are ordered to form a clear priority sequence.

[0037] After determining the garden division factor priority sequence, the target garden area is actually divided. Using the priority sequence as a guide, the design survey data of the target garden is processed. The design survey data contains rich information such as the topography of the garden, plant distribution, infrastructure layout, etc. According to the highest priority factor, such as the area function, the target garden is initially divided into large areas of different functional types, such as leisure areas, planting areas, and landscape areas. Then, based on the second-ranked geographical feature factor, the large areas are further subdivided. For example, in the planting area, it is divided into sub-areas suitable for the growth of different plants according to different soil types and terrain undulations. Finally, combined with the area size factor, these sub-areas are reasonably adjusted and optimized to ensure that each area meets the actual maintenance and management needs in terms of area. After this multi-level division process, M garden division areas with clear boundaries, different functions, and consistent with the actual situation of the garden are obtained, providing strong support for the subsequent precise management and maintenance of each area.

[0038] According to the determined priority sequence of the garden division factors, a regional coding system is designed. In the design of coding content, the factors with higher priority, such as regional function and geographical features, are fully combined to convert them into coding information with practical significance. For example, different regional functions are represented by numbers or letters, such as “A” for viewing area and “B” for leisure area; and geographical features are represented by specific code combinations, such as “01” for mountainous area and “02” for plain. The coding sequence is determined according to the priority of the garden division factors, first reflecting the regional function related coding, then the geographical feature coding, and finally integrating the regional area related information, to ensure that the coding sequence is logically clear and easy to identify and interpret. The coding mark, as a special symbol or sign to distinguish different regional codes, is set at a specific position of the code, such as adding a unique symbol at the beginning or end of the code, to quickly identify the coding type and the category it belongs to, and to enhance the identification and management efficiency of the code. By designing the coding content, reasonably arranging the coding sequence and clearly defining the coding mark, a complete and practical regional coding system is established, laying a solid foundation for the precise management and data connection of the garden area.

[0039] When the design of the regional coding system is completed, the M garden division regions divided previously are coded and marked to obtain M garden region information. According to the established regional coding system, each garden division region is given a corresponding code one by one. For each region, the rules of coding content, coding sequence and coding mark are followed. For example, a certain region belongs to a viewing area (corresponding to the symbol representing the viewing area in the coding content), is located in a plain (complies with the code of the plain in the geographical feature coding), and has an area within a certain range (corresponds to the coding part related to the area of the region), according to the coding sequence, these information are combined together, and a unique coding mark is added, to form a unique code for the region. In this way, the M garden division regions are coded and marked one by one, and each region has a specific code. These codes and the region attribute information they represent together constitute the M garden region information. These information become the basic data of the entire garden digital management system.

[0040] In one possible implementation manner, the garden division factor priority sequence acquisition unit further includes:

[0041] The garden division factor maintenance influence dataset acquisition unit is configured to perform maintenance influence data mining based on the garden region division factors, to obtain a garden division factor maintenance influence dataset.

[0042] The garden division factor principal component information acquisition unit is configured to perform standardization processing and principal component selection on the garden division factor maintenance influence dataset, to obtain garden division factor principal component information.

[0043] The garden division factor weight factor determination unit is configured to perform PCA weight distribution on each division factor in the garden region division factors based on the garden division factor principal component information, and determine a garden division factor weight factor.

[0044] The priority ranking unit is configured to perform priority ranking on the garden region division factors according to the garden division factor weight factor, and obtain a garden division factor priority sequence.

[0045] Specifically, based on the garden region division factors, maintenance influence data mining is performed, and a large amount of maintenance data related to the division factors such as region function, geographical feature, and region area is collected. These data cover the actual effects of different gardens under various maintenance measures, such as the growth status of plants in different functional regions, the occurrence frequency of diseases and pests in different geographical feature regions, and the consumption of maintenance resources in different area regions. By using data mining technology, maintenance influence data closely related to the garden division factors is extracted from the massive data, and a garden division factor maintenance influence data set is obtained.

[0046] The data set is subjected to standardization processing and principal component selection. Standardization processing can eliminate the influence of different data due to different dimensions and value ranges, so that the data are comparable. On this basis, the principal component analysis method is used to select the principal components that have the most critical influence on garden maintenance from the numerous influence data, so as to obtain garden division factor principal component information. These principal component information condenses the main features of the original data, and can more simply and effectively reflect the relationship between each division factor and garden maintenance.

[0047] Based on the garden division factor principal component information, the principal component analysis (PCA) algorithm is used to perform weight distribution on each division factor in the garden region division factors. The PCA algorithm determines the contribution degree of each division factor in the principal component by calculating the correlation between each principal component and the original division factor, and then obtains a garden division factor weight factor. These weight factors intuitively reflect the relative importance of each division factor to garden maintenance.

[0048] Each division factor in the garden region division factors is subjected to priority ranking according to the garden division factor weight factor. The greater the weight factor, the higher the importance of the corresponding division factor in garden maintenance, and the higher the priority. After such ranking, a garden division factor priority sequence is finally obtained. This sequence provides an important basis for subsequent scientific and reasonable division of garden regions and formulation of targeted maintenance strategies, and helps to improve the efficiency and quality of garden maintenance work.

[0049] In one possible implementation manner, the parallel analysis module 20 further includes:

[0050] A historical dataset collection unit is configured to collect a landscaping historical dataset and a landscaping pest and disease historical dataset.

[0051] An index extraction unit is configured to obtain a landscaping prediction target and a pest and disease identification target, and extract indexes from the landscaping prediction target and the pest and disease identification target to obtain a landscaping prediction index set and a pest and disease identification index set.

[0052] A classification identification training unit is configured to perform classification identification training on the landscaping historical dataset and the landscaping pest and disease historical dataset respectively by using the landscaping prediction index set and the pest and disease identification index set to obtain a landscaping prediction channel and a pest and disease identification channel.

[0053] A landscaping analysis double-channel construction unit is configured to parallelly integrate the landscaping prediction channel and the pest and disease identification channel to form the landscaping analysis double-channel.

[0054] Specifically, a landscaping historical dataset and a landscaping pest and disease historical dataset are widely collected. These datasets are widely sourced and cover landscaping plant growth data, maintenance measures and effect records, and detailed information such as the types, time, and damage degree of pest and disease occurrence under different seasons, different years, and different climate conditions.

[0055] The landscaping prediction target is determined, including predicting the growth rate of plants, the flowering and fruiting time, and the health degree of plants. Meanwhile, the pest and disease identification target is determined, such as identifying the types of pest and disease, judging the infection range and severity of pest and disease, etc. After the target is determined, indexes are extracted from multiple aspects. For landscaping prediction, physiological data of plants are collected, such as the height of plants, the number of leaves, the chlorophyll content, etc. Environmental data, such as temperature, humidity, light duration and intensity, etc. And maintenance data, including watering frequency, fertilizer type and dosage, etc. These data comprehensively constitute the landscaping prediction index set. For the pest and disease identification target, the extracted indexes cover the morphological characteristics of pest and disease, such as the size, color, and pattern of pests, and the colony morphology of bacteria. The symptom performance of plant invasion, such as leaf discoloration, perforation, and wilting, plant deformation, etc. And environmental factors of disease occurrence, such as surrounding vegetation conditions, soil pH, etc. These indexes together constitute the pest and disease identification index set, which provides a core basis for subsequent classification identification training and channel construction.

[0056] For the construction of the garden greening prediction channel, the garden greening prediction index set is used to carry out classification identification training on the garden greening historical data set, thereby obtaining a plurality of branch garden greening index predictors. These predictors predict the growth status of garden plants from different angles. Then, the branch predictors are fused by equal weight to make them work together, and then the fusion result is tested and adjusted by using historical data and actual situation to improve the accuracy and reliability of the prediction, and then the garden greening prediction channel is generated. In constructing the pest and disease identification channel, the convolutional neural network technology is used, and the pest and disease identification index set is used to carry out classification identification training on the garden pest and disease historical data set, to obtain a plurality of branch pest and disease index identifiers. Then, the branch identifiers are verified and evaluated to judge their accuracy in identifying pests and diseases, and a fusion coefficient is calculated. The fusion coefficient is determined according to the decisive factor of the pest and disease index and the identification accuracy, and through it, the branch pest and disease index identifiers are fused by weight, and the advantages of multiple identifiers are integrated to finally obtain the pest and disease identification channel.

[0057] From the perspective of architecture design, a parallel processing architecture mode is adopted to build a unified data input interface for the two channels, ensuring that the garden state data stream from the data perception module can be transmitted to the two channels simultaneously and without conflict. Then, the output results of the two channels are uniformly planned, and a comprehensive output interface is designed to integrate the plant growth trend, health status prediction data output by the garden greening prediction channel and the pest and disease identification results such as pest and disease species, infection range and severity output by the pest and disease identification channel. Through such integration, comprehensive garden growth analysis results can be obtained at one time, including both plant growth trend prediction and pest and disease related information, providing rich and comprehensive data support for the subsequent strategy analysis module to formulate accurate maintenance strategies, greatly improving the efficiency and scientificity of garden management, and ensuring the healthy and stable development of the garden ecological system.

[0058] In one possible implementation, the classification identification training unit further includes:

[0059] The branch garden greening index predictor acquisition unit is configured to use the garden greening prediction index set to carry out classification identification training on the garden greening historical data set, thereby obtaining a plurality of branch garden greening index predictors.

[0060] The garden greening prediction channel generation unit is configured to fuse the branch garden greening index predictors by equal weight and perform verification and evaluation optimization, thereby generating a garden greening prediction channel.

[0061] The branch pest and disease index identifier acquisition unit is configured to use the convolutional neural network to carry out classification identification training on the garden pest and disease historical data set based on the pest and disease identification index set, thereby obtaining a plurality of branch pest and disease index identifiers.

[0062] The pest and disease identification channel acquisition unit is configured to verify and evaluate the branch pest and disease index identifier and calculate a fusion coefficient, acquire a branch identifier fusion coefficient, and perform weighted fusion on the branch pest and disease index identifier based on the branch identifier fusion coefficient to obtain a pest and disease identification channel.

[0063] Specifically, a decision tree algorithm is used to classify and identify the landscaping historical data set for training, thereby obtaining a branch landscaping index predictor. First, the landscaping prediction index set is taken as the feature attribute of the decision tree, and the actual plant growth result corresponding to each sample in the landscaping historical data set is taken as the class label. The decision tree algorithm starts from the root node, selects an optimal index from the numerous prediction indexes based on criteria such as information gain or Gini index, and divides the data. For example, taking the plant height as an example, if the data is divided according to a certain height threshold, the class purity of the data is maximally improved, the index and threshold are selected for division to generate left and right child nodes. Then, the above process is recursively repeated on the child nodes to continuously subdivide the data until a predetermined stop condition is met, such as too few samples in the node or the class purity reaching a high level. In this way, a decision tree is constructed, which is a branch landscaping index predictor that can predict the growth state of plants, such as growth trend and health level, according to the input landscaping prediction index data.

[0064] A plurality of branch landscaping index predictors are obtained, which are obtained by classifying and identifying the landscaping historical data set for training, and each predicts the growth of landscaping plants from a different dimension. Then, these predictors are fused in a uniform weight manner, each branch predictor has the same influence in the new comprehensive model, and their prediction results are equally treated and integrated. After fusion, to ensure the accuracy and reliability of the generated prediction channel, verification and evaluation optimization work will be carried out. A large amount of actual landscaping data not involved in training is used as a verification set, the output results of the prediction channel are compared and analyzed with the actual plant growth in the verification set. According to the difference between the two, an optimization algorithm is used to adjust the parameters of the prediction channel, such as adjusting the correlation weight between the predictors and optimizing the internal algorithm logic. After multiple rounds of verification and evaluation and parameter adjustment, the accuracy and stability of the prediction channel are gradually improved, and finally a landscaping prediction channel that can accurately predict the growth trend of landscaping plants and provide strong support for landscaping maintenance decision-making is generated.

[0065] The convolutional neural network is used to train the historical data set of the garden pests and diseases based on the pest and disease identification index set, so as to obtain a branch pest and disease index identifier. First, the pest and disease identification index set is converted into an input data format that can be processed by the convolutional neural network, for example, the morphological characteristics of the pest and disease and the symptoms of the plant affected are quantified into a numerical matrix. Then, a convolutional neural network model is built, which includes multiple convolutional layers, pooling layers and fully connected layers. In the training stage, the sample data with clear pest and disease category labels in the historical data set of the garden pests and diseases are input into the convolutional neural network in turn. The convolutional layer extracts local features in the data by sliding the convolution kernel on the data, and the pooling layer down-samples the feature map to reduce the data volume and preserve the key features. As the data passes through the network layer by layer, the model continuously learns the correlation pattern between the pest and disease characteristics and the categories. After multiple rounds of training, the weights and bias parameters of the network are adjusted, so that the model can accurately output the corresponding pest and disease category prediction result according to the input pest and disease identification index data. At this time, the trained model becomes a branch pest and disease index identifier, which can be used to preliminarily identify the type and related information of the pests and diseases in the garden.

[0066] Validation and evaluation are carried out, and a validation data set independent of the training set is used. The data is input into each branch identifier, and the pest and disease identification results output by the branch identifier are compared with the true labels in the validation data set. Accuracy, recall rate, F1 value and other indicators are used to comprehensively evaluate the performance of each branch identifier. Then, the fusion coefficient calculation is carried out. According to the pest and disease index decisive factor and the identification accuracy obtained by the validation and evaluation, the fusion coefficient of each branch identifier is determined. For the branch identifier that plays a key role in pest and disease identification and has high identification accuracy, a larger fusion coefficient is given; otherwise, a smaller coefficient is given. Finally, based on the calculated fusion coefficients of the branch identifiers, all the branch identifiers are weighted and fused. When new input data comes, each branch identifier first gives its own identification result, and then the results are weighted and combined according to their respective fusion coefficients, so as to obtain the final pest and disease identification result, thus constructing a pest and disease identification channel with better comprehensive performance and more accurate identification.

[0067] In one possible implementation manner, the pest and disease identification channel acquisition unit further includes:

[0068] A pest and disease index decisive factor determination unit is configured to evaluate the importance of the pest and disease identification index set based on the historical data set of the garden pests and diseases, and determine a pest and disease index decisive factor.

[0069] A fusion coefficient calculation formula construction unit is configured to construct a fusion coefficient calculation formula: Wherein, D i represents the fusion coefficient of the i th branch identifier, identify the i-th branch identifier, β i identify the i-th branch identifier, D i With the increase of , the fusion coefficient of the i-th branch identifier increases.

[0070] A fusion coefficient calculation unit is configured to perform performance verification evaluation and fusion coefficient calculation on the branch pest index identifiers based on the fusion coefficient calculation formula, and obtain the branch identifier fusion coefficients.

[0071] Specifically, information is extracted from the historical garden pest data set, which covers various types of pest cases occurring at different times and places, as well as multi-dimensional data such as corresponding environmental conditions and plant varieties. Then, statistical analysis methods such as correlation analysis are used to calculate the correlation coefficient between each pest index and the actual occurrence of pests (such as pest type, damage degree, etc.). The higher the correlation coefficient, the closer the association between the index and the pest, and the more important it may be in the identification process. At the same time, the feature importance evaluation function of the random forest algorithm is used, which calculates the contribution of each index to reducing the uncertainty of sample labels in the process of building decision trees to determine the importance of the index. By combining the results of these statistical analysis and machine learning methods, the importance of each index in the pest index set is ranked, and those that play a key role in pest identification and have a large impact are selected as the pest index decisive factor.

[0072] A special fusion coefficient calculation formula is constructed, i.e. In this formula, D i represents the fusion coefficient of the i-th branch identifier, which directly affects the weight of the branch identifier in the final pest identification result; β represents the pest index decisive factor of the i-th branch identifier, reflecting the importance of the key index relied on by the branch identifier; β i D is the identification accuracy of the i-th branch identifier, reflecting the performance of the branch identifier. Moreover, D i increases with the increase of , which means that the more critical the pest index decisive factor and the higher the identification accuracy of the branch identifier, the larger the fusion coefficient, and the higher the weight in the final identification result.

[0073] According to the constructed fusion coefficient calculation formula, the performance verification evaluation and fusion coefficient calculation are performed on each branch disease and pest index identifier. In the performance verification evaluation process, a large amount of test data is used to test the actual performance of each branch identifier to ensure its reliability. Then, the fusion coefficient corresponding to each branch identifier is calculated by the formula, which will be used for subsequent weighted fusion of the branch disease and pest index identifier, so as to obtain a more accurate disease and pest identification channel, and provide strong support for garden disease and pest control.

[0074] In a possible implementation manner, the policy analysis module 30 further includes:

[0075] The garden greening maintenance label policy library acquisition unit is configured to classify and label the garden greening maintenance policy library according to the garden greening maintenance element information, and obtain a garden greening maintenance label policy library.

[0076] The greening maintenance policy parameter threshold acquisition unit is configured to perform policy retrieval and analysis in the garden greening maintenance label policy library based on the M garden growth analysis results, and obtain M greening maintenance policy parameter thresholds.

[0077] The target greening maintenance policy parameter acquisition unit is configured to perform global optimization in the M greening maintenance policy parameter thresholds, and obtain M target greening maintenance policy parameters.

[0078] The garden greening maintenance policy parameter determination unit is configured to perform correlation analysis and balanced correction on the M target greening maintenance policy parameters, and determine M garden greening maintenance policy parameters.

[0079] Specifically, according to the garden greening maintenance element information, the garden greening maintenance policy library is classified and labeled. Each type of strategy in the maintenance policy library is classified in detail according to the plant maintenance type (such as watering, fertilizing, pruning, etc.), the garden area function (such as ornamental area, leisure area, etc.), the seasonal characteristics and other maintenance elements, and the corresponding label is added to each type of strategy, so as to construct a clear and orderly garden greening maintenance label policy library.

[0080] In determining the M green maintenance strategy parameter thresholds, the M garden growth analysis results are used as retrieval basis. These analysis results contain rich information such as the growth trend of garden plants, the occurrence of plant diseases and insect pests, etc. Each garden growth analysis result is compared with the labels in the garden greening maintenance label strategy library. For example, if the growth analysis result of a certain garden area shows that the plants have signs of water shortage and have mild plant diseases and insect pests, the maintenance strategies with "plant water shortage" and "mild plant diseases and insect pests" related labels in the strategy library are retrieved. For each strategy retrieved, the content is analyzed in depth, and the key parameters related to maintenance operations are extracted, such as the amount and frequency of watering, the concentration and dosage of plant disease control agents, etc. Since the same maintenance scenario corresponds to multiple different strategies, the parameter values in these strategies will differ. The parameter value ranges are integrated to determine a reasonable parameter threshold interval for each garden area, and finally M green maintenance strategy parameter thresholds are obtained

[0081] According to the M green maintenance strategy parameter thresholds, the M green maintenance strategy particle swarm space is initialized. In this space, each particle represents a potential green maintenance strategy parameter combination, which is randomly generated within the previously determined parameter threshold range, providing diverse initial solutions for subsequent optimization. The garden greening maintenance goals are defined, including promoting plant healthy growth, controlling plant diseases and insect pests, and improving landscape effect, etc. These goals are analyzed in depth to convert them into quantifiable indicators, such as plant growth height, plant disease and insect pest occurrence rate, and landscape aesthetic score, etc. Then, through data correlation fitting, the relationship between different maintenance strategy parameters and these quantifiable indicators is analyzed, and a garden greening maintenance effect fitness function is constructed. Finally, the constructed garden greening maintenance effect fitness function is used to perform iterative global optimization in the M green maintenance strategy particle swarm space. In each iteration, the particle adjusts its position according to its historical optimal position and the global optimal position of the entire particle swarm, i.e. changes the maintenance strategy parameter combination. By continuously calculating the fitness values of each particle and comparing their sizes, the particle with the maximum fitness value is gradually selected. After multiple iterations, the M target green maintenance strategy parameters with the maximum particle fitness are finally determined, which are the strategy parameter combinations that are most beneficial to achieving the garden greening maintenance goals within the given parameter threshold range.

[0082] In combination with the M garden area information, such as the function, geographical features, plant species distribution, etc. of each area, the correlation influence analysis is performed on the M target green maintenance strategy parameters. This analysis process will further explore the unique attributes of each garden area on the maintenance strategy parameters, for example, in the plant-dense ornamental area, the fertilization strategy may be affected by the space limitation and landscape demand, and cannot be simply executed according to the general standard. Through this analysis, the influence degree of each area attribute on the maintenance strategy parameter is quantified, and then the M green maintenance influence coefficients are obtained. These coefficients reflect the correlation between different garden areas and the maintenance strategy parameters. Based on the obtained M green maintenance influence coefficients, the Nash equilibrium theory is used to correct the M target green maintenance strategy parameters, so that the adjustment of the self-strategy of each participant will not bring better results when the strategies of other participants remain unchanged. In the garden maintenance scenario, each garden area can be regarded as a participant, and the maintenance strategy parameter is the decision variable. According to the green maintenance influence coefficient, the target green maintenance strategy parameter is adjusted, the mutual influence between different areas is comprehensively considered, the over-optimization of a certain area affecting the maintenance effect of other areas is avoided, and a balanced maintenance strategy parameter combination is sought among the areas. After such Nash equilibrium correction, the M garden green maintenance strategy parameters are finally determined, thereby providing a scientific and reasonable basis for the precise maintenance of the entire garden, and ensuring that each area of the garden reaches the best state under unified maintenance management.

[0083] In one possible implementation manner, the target green maintenance strategy parameter acquisition unit further includes:

[0084] The green maintenance strategy particle swarm space initialization unit is configured to initialize M green maintenance strategy particle swarm spaces according to the M green maintenance strategy parameter thresholds.

[0085] The fitness function construction unit is configured to acquire a garden green maintenance target, perform effect index analysis and data correlation fitting on the garden green maintenance target, and construct a garden green maintenance effect fitness function.

[0086] The iterative global optimization unit is configured to perform iterative global optimization in the M green maintenance strategy particle swarm spaces by using the garden green maintenance effect fitness function, and determine the M target green maintenance strategy parameters with the maximum particle fitness.

[0087] Specifically, the M greening maintenance strategy parameter thresholds are determined, and for each garden area (a total of M areas), a corresponding particle swarm space is constructed. In each particle swarm space, a large number of particles are randomly generated, and each particle represents a unique combination of greening maintenance strategy parameters. The parameter values in these combinations are strictly within the corresponding parameter threshold range. For example, if the fertilizer application rate threshold for a certain area is 0.5-2 kg per square meter, and the watering frequency threshold is 3-7 times per week, then the particle's fertilizer application rate parameter will be randomly selected between 0.5-2 kg, and the watering frequency parameter will be randomly determined between 3-7 times per week, thereby forming a diverse initial parameter combination. These particles collectively form the greening maintenance strategy particle swarm space.

[0088] In constructing the garden greening maintenance effect fitness function, the garden greening maintenance goals are first determined, mainly covering promoting plant growth, improving landscape effect, and preventing and controlling pests and diseases. Effect index analysis is carried out for these goals. For example, select plant growth index G (a comprehensive value can be calculated by plant height, leaf number, etc. to measure), landscape effect index L (quantified by landscape aesthetics score), and pest and disease control index P (represented by the reciprocal of the infection rate, the lower the infection rate, the larger the value). Then, data correlation fitting is carried out. Through analysis of historical maintenance data, it is found that plant growth is greatly affected by fertilizer application rate and watering amount, landscape effect is related to plant layout and pruning frequency, and pest and disease control is closely related to pesticide use amount. These relationships are fitted by linear relationship. The relationship between plant growth and maintenance factors is G = a1x1 + a2x2 (x1 represents fertilizer application rate, x2 represents watering amount, a1 and a2 are the corresponding influence coefficients), the relationship between landscape effect and maintenance factors is L = b1y1 + b2y2 (y1 represents plant layout adjustment parameter, y2 represents pruning frequency, b1 and b2 are influence coefficients), and the relationship between pest and disease control and maintenance factors is P = c1z1 (z1 represents pesticide use amount, c1 is the influence coefficient). Based on the above analysis, a simple garden greening maintenance effect fitness function F = w1G + w2L + w3P is constructed, where w1, w2, and w3 are the weights of plant growth, landscape effect, and pest and disease control, respectively, and w1 + w2 + w3 = 1. The weights are determined according to the emphasis of different gardens, such as a garden emphasizing plant growth, w1 is relatively large. Through this function, the maintenance effect under different maintenance strategy combinations can be quantified, making it easy to select the most suitable maintenance strategy for the maintenance goal.

[0089] The constructed garden greening maintenance effect fitness function is used to carry out iterative global optimization in M greening maintenance strategy particle swarm spaces, and then M target greening maintenance strategy parameters are determined. In each greening maintenance strategy particle swarm space, a large number of particles each represent different combinations of greening maintenance strategy parameters. In the initial state, these particles are randomly distributed within a given parameter threshold range. Taking a particle swarm as an example, one particle may represent a specific combination of fertilizer application rate, watering frequency and pruning cycle. When the iterative optimization begins, first, the parameter combination represented by each particle is substituted into the garden greening maintenance effect fitness function for calculation to obtain a fitness value, which reflects the degree of fit of the maintenance strategy to the maintenance goal. In each iteration process, the particle adjusts its position (i.e., the best maintenance strategy parameter combination ever exhibited) by referring to the optimal fitness value obtained in its history and the global optimal position (the parameter combination with the maximum fitness value among all particles) found by the entire particle swarm, that is, by changing the maintenance strategy parameter combination it represents. For example, if a particle finds that moving towards the global optimal particle can improve its fitness value, it will adjust parameters such as fertilizer application rate, watering frequency, etc. With each iteration, the particle swarm as a whole gradually moves in a better direction, and the fitness value continuously improves. After multiple iterations, in each particle swarm, the parameter combination represented by the particle with the maximum fitness value gradually emerges. The parameter combinations corresponding to these particles with the maximum fitness value selected in M particle swarms are the final M target greening maintenance strategy parameters. These parameter combinations, after considering factors such as plant growth, landscape effect and pest control, can best meet the garden greening maintenance goals and provide scientific and effective guidance for fine maintenance of gardens.

[0090] In one possible implementation, the determining of the M garden greening maintenance strategy parameters further includes:

[0091] A greening maintenance influence coefficient acquisition unit is configured to perform associated influence analysis on the M target greening maintenance strategy parameters according to the M garden area information, and obtain M greening maintenance influence coefficients.

[0092] A Nash equilibrium correction unit is configured to perform Nash equilibrium correction on the M target greening maintenance strategy parameters based on the M greening maintenance influence coefficients, and determine the M garden greening maintenance strategy parameters.

[0093] Specifically, the M garden area information is comprehensively combed, which covers the geographical location, topography, plant species distribution, regional function positioning (such as leisure area, ornamental area, etc.) and past maintenance history of the area and other aspects. For each garden area, its unique information characteristics and corresponding target green maintenance strategy parameters are analyzed one by one. Taking a garden area as an example, if the area is located at the wind outlet position and is planted with shallow-rooted plants, the windproof measures (such as the height and material of the fence parameters) will be affected by the geographical location and plant species when analyzing the target green maintenance strategy parameters. From the perspective of geographical location, the wind power at the wind outlet is large, and the height of the fence may need to be increased; from the perspective of plant species, shallow-rooted plants have weak wind resistance, and the stability and protection range of the fence are required to be higher. Through in-depth study of this influence relationship, a quantitative analysis method is adopted to establish a regression model, taking each factor in the area information as the independent variable and the target green maintenance strategy parameter as the dependent variable, and calculating the influence degree of each factor on the parameter. After considering all relevant factors, a value that can quantify the influence degree of the area information on the target green maintenance strategy parameter is obtained, which is a green maintenance influence coefficient. In this way, the correlation and influence analysis is carried out for the M garden areas one by one, and finally M green maintenance influence coefficients are obtained, which provide a strong basis for subsequent precise adjustment of maintenance strategy parameters.

[0094] The coefficients are used to correct the Nash equilibrium of M target green maintenance strategy parameters, and then determine the key stage of the final M garden green maintenance strategy parameters. The core of Nash equilibrium theory is that in a system involving multiple decision-making subjects and mutual influence, the decision of each subject not only considers its own interests, but also takes into account the decisions of other subjects, so as to achieve a balanced state. At this time, any unilateral change in decision-making cannot achieve better results. In the context of garden maintenance, M garden areas are M decision-making subjects, and the target green maintenance strategy parameters of each area are their respective decision variables. Based on the obtained M green maintenance influence coefficients, the mutual relationship and influence between regions are analyzed. For example, increasing the watering frequency in area A may cause the soil humidity in surrounding area B to rise, affecting the growth environment of plants in area B, and then changing the originally set watering and fertilization strategy parameters in area B. At this time, according to the Nash equilibrium principle, the interests of each region and the overall garden maintenance goal need to be considered. Adjust the target green maintenance strategy parameters of each region to meet the basic maintenance needs of each region while ensuring the balance and stability of the entire garden system. In the adjustment process, different parameter combinations are constantly tried, the maintenance effect of each region under each combination is calculated, and the weight distribution is made according to the green maintenance influence coefficient. Through multiple iterations and optimization, a set of maintenance strategy parameters that achieve Nash equilibrium between regions is finally determined, that is, M garden green maintenance strategy parameters. These parameters not only consider the unique needs of each region, but also take into account the mutual influence between regions, and can achieve scientific and efficient maintenance of the entire garden.

[0095] In the second embodiment, based on the same inventive concept as the digital management system for garden green maintenance in the preceding embodiments, as shown in Figure 2 The present application provides a digital management method for garden green maintenance. The method and system embodiments in the present application are based on the same inventive concept. The method comprises:

[0096] Step S100: Divide and encode the target garden into M garden area information, and collect M garden state data streams of the M garden area information by the perception network deployment.

[0097] Step S200: Construct a garden analysis double-channel, which includes a garden green prediction channel and a pest and disease identification channel. The M garden state data streams are analyzed in parallel by using the garden analysis double-channel, and M garden growth analysis results are output.

[0098] Step S300: Obtain a garden green maintenance strategy library, and use the garden green maintenance strategy library to analyze and correct the M garden growth analysis results, and determine M garden green maintenance strategy parameters.

[0099] Step S400: based on the matching activation of the M garden maintenance strategy parameters and the garden maintenance equipment list, a garden linkage maintenance equipment set is obtained, and the target garden is executed for green maintenance remote control through the garden linkage maintenance equipment set.

[0100] Further, step S100 further includes:

[0101] Step S110: obtain garden area division factors, including regional function, geographical features, and regional area.

[0102] Step S120: prioritize the garden area division factors according to the garden area division factors, and obtain a garden division factor priority sequence.

[0103] Step S130: based on the garden division factor priority sequence, the design survey data of the target garden is regionally multi-divided, and M garden division regions are obtained.

[0104] Step S140: according to the garden division factor priority sequence, design a regional coding system, including coding content, coding sequence, and coding identification.

[0105] Step S150: according to the regional coding system, code and label the M garden division regions, and obtain the M garden region information.

[0106] Further, step S120 further includes:

[0107] Step S121: based on the garden area division factors, maintenance influence data mining is performed to obtain a garden division factor maintenance influence dataset.

[0108] Step S122: standardize the garden division factor maintenance influence dataset and select principal components to obtain garden division factor principal component information.

[0109] Step S123: based on the garden division factor principal component information, PCA weight distribution is performed on each division factor in the garden area division factors to determine a garden division factor weight factor.

[0110] Step S124: prioritize the garden area division factors according to the garden division factor weight factor to obtain the garden division factor priority sequence.

[0111] Further, step S200 further includes:

[0112] Step S210: collect garden greening historical data set and garden pest historical data set.

[0113] Step S220: Obtain a landscaping prediction target and a pest and disease identification target, perform index extraction on the landscaping prediction target and the pest and disease identification target, and obtain a landscaping prediction index set and a pest and disease identification index set.

[0114] Step S230: Perform classification and identification training on the landscaping historical data set and the pest and disease historical data set using the landscaping prediction index set and the pest and disease identification index set, respectively, to obtain a landscaping prediction channel and a pest and disease identification channel.

[0115] Step S240: Parallelly integrate the landscaping prediction channel and the pest and disease identification channel to form the garden analysis double channel.

[0116] Further, step S230 further includes:

[0117] Step S231: Perform classification and identification training on the landscaping historical data set using the landscaping prediction index set to obtain a branch landscaping index predictor.

[0118] Step S232: Perform equal-weight fusion and verification evaluation optimization on the branch landscaping index predictor to generate a landscaping prediction channel.

[0119] Step S233: Perform classification and identification training on the pest and disease historical data set using a convolutional neural network based on the pest and disease identification index set to obtain a branch pest and disease index identifier.

[0120] Step S234: Perform verification evaluation and fusion coefficient calculation on the branch pest and disease index identifier to obtain a branch identifier fusion coefficient, and perform weighted fusion on the branch pest and disease index identifier based on the branch identifier fusion coefficient to obtain a pest and disease identification channel.

[0121] Further, step S234 further includes:

[0122] Step S2341: Perform importance evaluation on the pest and disease identification index set based on the pest and disease historical data set to determine a pest and disease index decisive factor.

[0123] Step S2342: Construct a fusion coefficient calculation formula: wherein, D i represents the fusion coefficient of the i-th branch identifier, represents the pest and disease index decisive factor of the i-th branch identifier, β i represents the identification accuracy of the i-th branch identifier, D i increases with the increase of .

[0124] Step S2343: performance verification evaluation and fusion coefficient calculation are performed on the branch pest and disease index identifier based on the fusion coefficient calculation formula, and the branch identifier fusion coefficient is obtained.

[0125] Further, step S300 further includes:

[0126] Step S310: the garden greening maintenance strategy library is classified and labeled according to the garden greening maintenance element information, and a garden greening maintenance label strategy library is obtained.

[0127] Step S320: strategy retrieval and analysis are performed in the garden greening maintenance label strategy library based on the M garden growth analysis results, and M greening maintenance strategy parameter thresholds are obtained.

[0128] Step S330: global optimization is performed in the M greening maintenance strategy parameter thresholds, and M target greening maintenance strategy parameters are obtained.

[0129] Step S340: correlation analysis and balanced correction are performed on the M target greening maintenance strategy parameters, and M garden greening maintenance strategy parameters are determined.

[0130] Further, step S330 further includes:

[0131] Step S331: M greening maintenance strategy particle swarm spaces are initialized according to the M greening maintenance strategy parameter thresholds.

[0132] Step S332: a garden greening maintenance target is obtained, effect index analysis and data correlation fitting are performed on the garden greening maintenance target, and a garden greening maintenance effect fitness function is constructed.

[0133] Step S333: the garden greening maintenance effect fitness function is used to perform iterative global optimization in the M greening maintenance strategy particle swarm spaces, and M target greening maintenance strategy parameters with the maximum particle fitness are determined.

[0134] Further, step S340 further includes:

[0135] Step S341: correlation influence analysis is performed on the M target greening maintenance strategy parameters according to the M garden area information, and M greening maintenance influence coefficients are obtained.

[0136] Step S342: Nash equilibrium correction is performed on the M target greening maintenance strategy parameters based on the M greening maintenance influence coefficients, and the M garden greening maintenance strategy parameters are determined.

[0137] It should be noted that the above-mentioned embodiment sequences of the present application are merely for description only, but not for representing the advantages and disadvantages of the embodiments. And the above-mentioned embodiments of the present specification have been described. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0138] The above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0139] The specification and drawings are merely exemplary of the present application, and any and all modifications, variations, combinations or equivalents that are within the scope of the present application should be included. Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A digital and intelligent management system for landscaping maintenance, characterized in that, The system comprises: a data perception module for regionally dividing and encoding a target garden to obtain M garden region information, and for collecting and obtaining M garden state data streams of the M garden region information by network deployment; a parallel analysis module for constructing a garden analysis double channel, the garden analysis double channel comprising a garden greening prediction channel and a pest and disease identification channel, and for performing parallel analysis on the M garden state data streams by using the garden analysis double channel to output M garden growth analysis results; a strategy analysis module for obtaining a garden greening maintenance strategy library, and for performing strategy analysis and correlation correction on the M garden growth analysis results by using the garden greening maintenance strategy library to determine M garden greening maintenance strategy parameters; a digitalized management and control module for matching and activating the M garden greening maintenance strategy parameters with a garden maintenance equipment list to obtain a garden linkage maintenance equipment set, and for performing remote management and control of garden greening maintenance on the target garden by using the garden linkage maintenance equipment set; wherein the parallel analysis module further comprises: a historical data set collection unit for collecting and obtaining a garden greening historical data set and a garden pest and disease historical data set; an index extraction unit for obtaining a garden greening prediction target and a pest and disease identification target, and for performing index extraction on the garden greening prediction target and the pest and disease identification target to obtain a garden greening prediction index set and a pest and disease identification index set; a classification identification training unit for performing classification identification training on the garden greening historical data set and the garden pest and disease historical data set by using the garden greening prediction index set and the pest and disease identification index set respectively to obtain a garden greening prediction channel and a pest and disease identification channel; a garden analysis double channel construction unit for parallel integration of the garden greening prediction channel and the pest and disease identification channel to form the garden analysis double channel; the classification identification training unit further comprises: a branch garden greening index predictor acquisition unit for performing classification identification training on the garden greening historical data set by using the garden greening prediction index set to obtain a branch garden greening index predictor; a garden greening prediction channel generation unit for performing equal weight fusion and verification evaluation optimization on the branch garden greening index predictor to generate a garden greening prediction channel; a branch pest and disease index identifier acquisition unit for performing classification identification training on the garden pest and disease historical data set by using a convolutional neural network based on the pest and disease identification index set to obtain a branch pest and disease index identifier; a pest and disease identification channel acquisition unit for performing verification evaluation and fusion coefficient calculation on the branch pest and disease index identifier to obtain a branch identifier fusion coefficient, and for performing weighted fusion on the branch pest and disease index identifier based on the branch identifier fusion coefficient to obtain a pest and disease identification channel; the pest and disease identification channel acquisition unit further comprises: a pest and disease index decisive factor determination unit for performing importance evaluation on the pest and disease identification index set based on the garden pest and disease historical data set to determine a pest and disease index decisive factor; A fusion coefficient calculation formula construction unit is used to construct the fusion coefficient calculation formula: Among them, D i The fusion coefficients representing the i-th branch recognizer The determining factor for the pest and disease index of the i-th branch identifier, β i D represents the recognition accuracy of the i-th branch recognizer. i along with It increases with the increase of; The fusion coefficient calculation unit is configured to perform performance verification evaluation and fusion coefficient calculation on the branch pest index identifiers based on the fusion coefficient calculation formula, and obtain the branch identifier fusion coefficients.

2. The digital intelligent management system for garden maintenance according to claim 1, wherein The data perception module further comprises: The garden area division factor obtaining unit is configured to obtain garden area division factors, which include area functions, geographical features, and area sizes. The garden division factor priority sequence obtaining unit is configured to perform priority sequencing on the garden area division factors according to the garden area division factors, and obtain a garden division factor priority sequence. The garden division area obtaining unit is configured to perform multi-level division on the design survey data of the target garden based on the garden division factor priority sequence, and obtain M garden division areas. The area coding system design unit is configured to design an area coding system according to the garden division factor priority sequence, which includes coding content, coding sequence, and coding identification. The garden area information obtaining unit is configured to code and label the M garden division areas according to the area coding system, and obtain M garden area information.

3. The digitalized management system for garden maintenance according to claim 2, wherein, The garden division factor priority sequence obtaining unit further comprises: The garden division factor maintenance influence dataset obtaining unit is configured to perform maintenance influence data mining based on the garden area division factors, and obtain a garden division factor maintenance influence dataset. The garden division factor principal component information obtaining unit is configured to perform standardization processing and principal component selection on the garden division factor maintenance influence dataset, and obtain garden division factor principal component information. The garden division factor weight factor determination unit is configured to perform PCA weight allocation on each division factor in the garden area division factors based on the garden division factor principal component information, and determine garden division factor weight factors. The priority sequencing unit is configured to perform priority sequencing on the garden area division factors according to the garden division factor weight factors, and obtain the garden division factor priority sequence.

4. The digitalized management system for garden maintenance according to claim 1, wherein, The strategy analysis module further comprises: The garden greening maintenance label strategy library obtaining unit is configured to classify and label the garden greening maintenance strategy library according to garden greening maintenance element information, and obtain a garden greening maintenance label strategy library. The greening maintenance strategy parameter threshold obtaining unit is configured to perform strategy retrieval analysis in the garden greening maintenance label strategy library based on the M garden growth analysis results, and obtain M greening maintenance strategy parameter thresholds. The target greening maintenance strategy parameter obtaining unit is configured to perform global optimization in the M greening maintenance strategy parameter thresholds, and obtain M target greening maintenance strategy parameters. The garden greening maintenance strategy parameter determination unit is configured to perform correlation analysis and balanced correction on the M target greening maintenance strategy parameters, and determine M garden greening maintenance strategy parameters.

5. The digitalized management system for garden maintenance according to claim 4, wherein, The target greening maintenance strategy parameter obtaining unit further comprises: The space initialization unit is configured to initialize M greening maintenance strategy particle swarm spaces according to the M greening maintenance strategy parameter thresholds. The fitness function construction unit is configured to obtain a landscaping maintenance target, perform effect index analysis and data correlation fitting on the landscaping maintenance target, and construct a landscaping maintenance effect fitness function. The iterative global optimization unit is configured to perform iterative global optimization in the M landscaping maintenance strategy particle swarm space using the landscaping maintenance effect fitness function, and determine M target landscaping maintenance strategy parameters with the maximum particle fitness.

6. The digitalized management system for garden maintenance according to claim 5, wherein, The landscaping maintenance strategy parameter determination unit further includes: The landscaping maintenance influence coefficient acquisition unit is configured to perform correlation influence analysis on the M target landscaping maintenance strategy parameters according to the M landscaping area information, and obtain M landscaping maintenance influence coefficients. The Nash equilibrium correction unit is configured to perform Nash equilibrium correction on the M target landscaping maintenance strategy parameters based on the M landscaping maintenance influence coefficients, and determine the M landscaping maintenance strategy parameters.

7. A digital and intelligent management method for landscaping maintenance, characterized in that, The method is implemented by the digital and intelligent management system for landscaping maintenance according to any one of claims 1-6, and the method includes: The target garden is divided into regions and coded to obtain M landscaping area information, and M garden state data streams of the M landscaping area information are collected by a perception network deployment; A garden analysis double channel is constructed, including a landscaping prediction channel and a pest and disease identification channel, the M garden state data streams are analyzed in parallel by the garden analysis double channel, and M garden growth analysis results are output; A landscaping maintenance strategy library is obtained, the M garden growth analysis results are analyzed and corrected by the landscaping maintenance strategy library, and M landscaping maintenance strategy parameters are determined; The M landscaping maintenance strategy parameters are matched and activated based on a landscaping maintenance equipment list to obtain a landscaping linkage maintenance equipment set, and the target garden is executed for landscaping maintenance remote control by the landscaping linkage maintenance equipment set.

Images