A method and system for generating management recommendations for a long-lasting, low-maintenance mixed flower bed

Abstract

This invention relates to the field of management technology and discloses a method and system for generating management suggestions for long-term, low-maintenance mixed flower borders. The method includes: acquiring plant growth data and community succession data based on the spatial distribution characteristics of the sample plot, plant growth morphology characteristics, and environmental characteristics during the flower border's growth cycle; extracting characteristic data on plant morphology, growth rhythm, and tolerance performance; classifying all plant feature vectors based on a decision tree model; determining distance parameters between different plant combinations based on the K-nearest neighbor algorithm and judging the stability trend and substitution relationship of each plant combination; analyzing the vertical hierarchy, coverage relationship, and color coordination relationship of plants based on the plant combination structure; determining the association results between the community discrimination results and the flower border structure layout based on an association rule algorithm; and determining management suggestions based on the association results and a planning suggestion model. This invention ensures the reliability of mixed flower border management.

Description

Technical Field

[0001] This invention relates to the field of management technology, and more specifically, to a method and system for generating management suggestions for long-term, low-maintenance mixed flower borders. Background Technology

[0002] With the advancement of ecological landscaping construction, mixed flower borders, as a form of landscaping that integrates multiple plant types and combines aesthetic appeal with ecological stability, are widely used in urban green spaces, park landscapes, residential greening, and ecological restoration projects. However, the design and management of existing mixed flower borders largely rely on the experience of landscaping professionals, leading to insufficient landscape stability in the long term. Firstly, the configuration of plant combinations lacks data support, neglecting morphological changes, community succession patterns, and differences in environmental adaptability during the plant's growth cycle. The design process fails to collect and integrate plant growth data, community succession data, and environmental characteristics, making it impossible to accurately predict the long-term coexistence stability of different plant combinations. This can easily lead to risks such as overexpansion of dominant species, extinction of weaker species, and disordered landscape hierarchy. Secondly, current technologies for assessing flower border communities typically remain at the qualitative descriptive level, making it difficult to analyze the interactions between plant combinations through quantitative indicators. This makes it impossible to identify the risk of community degradation, resulting in difficulties in ensuring the long-term effectiveness of the flower border landscape. Third, there is insufficient correlation between the structural layout of flower borders and community characteristics. Existing flower border management is based on a single dimension of landscape needs or plant growth status, without linking the structural layout with community stability and succession patterns. Management recommendations rely on experience and cannot output precise control recommendations based on the dynamic changes of flower borders, resulting in blind maintenance work. Flower border management cannot achieve the full chain transformation from data to plant combination structure and from community identification results to management recommendations, thus restricting the level of intelligence and precision in flower border management.

[0003] Therefore, it is necessary to design a long-term, low-maintenance management suggestion generation method and system to solve the problems existing in the current technology. Summary of the Invention

[0004] In view of this, the present invention proposes a method and system for generating management suggestions for long-term, low-maintenance mixed flower borders, aiming to solve the above problems.

[0005] In one aspect, this invention proposes a method for generating management recommendations for long-lasting, low-maintenance mixed flower borders, comprising: During the growth cycle of the flower border, plant growth data and community succession data were obtained based on the spatial distribution characteristics, plant growth morphology characteristics and environmental characteristics of the sample plots, and the plant growth data and community succession data were constructed into a flower border data table. The parameters of the flower border data table are adjusted, the characteristic data of plant morphology, growth rhythm and tolerance are extracted, the plant feature vectors are determined, all plant feature vectors are classified based on the decision tree model, and the plant combination structure is determined according to the split node characteristics. The distance parameters between different plant combinations are determined based on the K-nearest neighbor algorithm. The stability trend and substitution relationship of each plant combination are judged based on the distance parameters. The community discrimination result is determined based on the judgment result and the plant combination structure. Based on the plant combination structure, the vertical hierarchy, coverage relationship and color coordination relationship of the plants are analyzed to determine the hierarchical flower border structure layout. Based on the association rule algorithm, the association result between the community discrimination result and the flower border structure layout is determined. Based on the association result and the planning suggestion model, management suggestions are determined.

[0006] Furthermore, during the flower border growth cycle, plant growth data and community succession data are obtained based on the spatial distribution characteristics, plant growth morphology characteristics, and environmental characteristics of the sample plots. When constructing the plant growth data and community succession data into a flower border data table, the following steps are included: Spatial reference points were set up within the sample plot and several planar coordinate grids were established. The position of each individual plant was marked with the planar coordinate grids, and the coordinate number of each plant was determined. At each observation time, the growth morphology indicators and environmental monitoring indicators of the marked plants are collected, and the plant distance between adjacent plants is recorded. Based on the plant distance and growth morphology indicators, plant growth records are determined. The growth morphology indicators include plant height, crown width, leaf area, flowering status and wilting status. The environmental monitoring indicators include light intensity, air temperature, soil moisture content and wind speed. By comparing the appearance or disappearance of plants, changes in canopy coverage, and changes in species composition between adjacent observation times, community succession events are determined. These community succession events include the addition of new vegetation, the withdrawal of existing vegetation, and the transfer of dominant vegetation. Plant growth records, environmental monitoring indicators, and community succession events at each observation time were integrated in chronological order and by coordinate number to construct a flower border data table with time as the index and coordinate number as the field.

[0007] Furthermore, when performing parameter normalization on the flower border data table, extracting characteristic data on plant morphology, growth rhythm, and tolerance performance, and determining plant feature vectors, the process includes: Based on the time and coordinate number in the flower border data table, plant growth records appearing at consecutive observation times within the same plane coordinate grid are merged to determine the growth sequence of a single plant or plant combination. In each growth sequence, structural morphological parameters are determined based on plant height, crown width, leaf area, and clump width. The structural morphological parameters include average height, maximum height, crown width expansion, and leaf area index. Based on the changes in plant height, the rate of crown expansion, and the timing of flowering start, full bloom, and fruiting at adjacent observation times, growth rhythm parameters are determined. These growth rhythm parameters include growth rate, length of seasonal change cycle, and seasonal transition time interval. The tolerance performance parameters during the injury period were statistically analyzed, including the number of surviving plants, the degree of leaf damage, and the rate of decline in growth rate. The structural morphology parameters, growth rhythm parameters, and tolerance performance parameters are normalized to determine the plant feature vectors of individual plants or plant combinations.

[0008] Furthermore, when classifying all plant feature vectors based on a decision tree model and determining the plant combination structure based on the features of split nodes, the process includes: All plant feature vectors are constructed into a plant feature vector set. In the plant feature vector set, structural morphology parameters, growth rhythm parameters, and tolerance performance parameters are used as candidate splitting attributes. The candidate splitting attributes are used as the splitting basis. The plant feature vector set is divided into several attribute subsets according to the differences of different attributes. Based on the distribution changes of structural morphology differences, growth rhythm differences, and tolerance characteristic differences within each attribute subset, the splitting nodes and node splitting paths are determined. Based on the hierarchical order of the node splitting path, each plant feature vector is assigned to the corresponding splitting node level by level to determine the plant category tree. In the plant category tree, the set of plant feature vectors located at the same end node is determined, and a plant combination group with the end node as the category boundary is constructed. The distribution of the plant combination group is compared with the distribution in the entire category tree to determine the similarity. Combination groups with similarity greater than or equal to the similarity threshold are merged to determine the plant combination structure.

[0009] Furthermore, when determining the distance parameters between different plant combinations based on the K-nearest neighbor algorithm, judging the stability trend and substitution relationship of each plant combination based on the distance parameters, and determining the community discrimination result based on the judgment result and the plant combination structure, the process includes: The plant feature vectors of the same category nodes in the plant combination structure are merged to determine the combined feature vector. The vector difference is calculated for every two combined feature vectors in the feature space, and the distance parameter is determined based on the vector difference. The distance parameters in the feature space are arranged in ascending order to determine several neighboring combinations. The neighboring combinations of the same plant combination are compared to determine the persistence of the neighboring relationship of the plant combination. The number of differences in neighboring combinations at different stages is counted to determine the volatility of the neighboring relationship of the plant combination. Based on the aforementioned persistence and volatility, the stability trend of each plant combination is determined. According to the distance variation between adjacent plant combinations within a neighboring combination, the substitution relationship of each plant combination in the structural hierarchy is determined, and the substitution order is determined according to the continuity interval. The community discrimination result is determined based on the stability trend, substitution relationship, and substitution order.

[0010] Furthermore, when analyzing the vertical hierarchy, coverage, and color coordination of plants based on the aforementioned plant combination structure to determine the hierarchical flower border structure layout, the following steps are included: Extract the height parameters from the feature vectors of each plant, and arrange the height parameters in ascending order to determine several height intervals; Several height intervals are mapped to the planar coordinate grid of the sample plot to determine the spatial hierarchy of each plant combination. Based on the coverage range and crown extension of the plant combinations at adjacent levels in the spatial hierarchy, the coverage overlap ratio between each plant combination is determined. Based on the coverage overlap ratio, the dominant and subordinate combinations are divided, and the coverage relationship structure diagram is determined.

[0011] Furthermore, when analyzing the vertical hierarchy, coverage, and color coordination of plants based on the aforementioned plant combination structure to determine the hierarchical flower border structure layout, the method also includes: Extract the color attributes corresponding to each plant combination, determine the color distribution points based on hue, brightness and saturation, determine the color transition sequence based on the distance between the color distribution points, and determine the color arrangement order between each plant combination based on the color transition sequence; The color arrangement order is superimposed on the coverage relationship structure diagram. Based on the color distance between each dominant coverage combination and the subordinate coverage combination, the coverage-color correspondence structure is determined. The coverage-color correspondence structure is then superimposed on the spatial hierarchy diagram to determine the hierarchical position of different height intervals in the superimposed space.

[0012] Furthermore, when determining the association results between the community discrimination result and the flower border structure layout based on the association rule algorithm, the following steps are included: The Eclat algorithm is used to generate several candidate item sets based on the hierarchical location and community discrimination results. Frequent itemsets are determined based on the support of the candidate item sets. Association rules are determined based on the frequent itemsets. Association results are determined based on the association rules.

[0013] Furthermore, when determining management recommendations based on the aforementioned correlation results and planning recommendation model, the following steps are included: Obtain the flower border dataset and divide the flower border dataset into a training set and a test set; The model establishes parameters and builds a decision model based on grid search. The decision model is trained on the training set and tested on the test set. Finally, the input is determined to be the association result, and the output is a planning suggestion model for management suggestions.

[0014] Compared with existing technologies, the advantages of this invention are as follows: By collecting spatial distribution, plant growth morphology, and environmental characteristic data and constructing a flower border data table, it avoids the traditional experience-based approach and provides multi-dimensional data support for plant combination configuration. This allows for accurate prediction of the long-term coexistence stability of plant combinations, avoiding risks such as overexpansion of dominant species, extinction of weak species, and disordered landscape hierarchy, thus ensuring the integrity of the flower border community structure and landscape stability. Based on the classification of plant feature vectors using a decision tree model and the quantification of plant combination distance parameters using the K-nearest neighbor algorithm, it achieves a leap from qualitative description to quantitative analysis in flower border community assessment, accurately determining the trend of combination stability and substitution relationships, thereby identifying community degradation risks in advance and providing a basis for the long-term maintenance of flower borders. By analyzing the vertical layers, coverage, and color relationships of plants to determine the hierarchical layout, and combining the association rule algorithm to establish the relationship between community discrimination results and structural layout, a full-chain transformation from data to plant combination structure and community discrimination results to management suggestions is realized. This avoids single-dimensional experience-based management, and the output control suggestions can reduce the blindness of maintenance, thereby reducing the input of human and material resources and improving the long-term effectiveness of flower border landscape and the level of intelligent management.

[0015] On the other hand, this application also provides a management suggestion generation system for long-lasting, low-maintenance mixed flower borders, used to apply the above-mentioned management suggestion generation method for long-lasting, low-maintenance mixed flower borders, including: The data acquisition unit is configured to acquire plant growth data and community succession data based on the spatial distribution characteristics, plant growth morphology characteristics and environmental characteristics of the sample plot during the flower border growth cycle, and to construct the plant growth data and community succession data into a flower border data table. The landscape structure unit is configured to perform parameter adjustment on the flower border data table, extract feature data of plant morphology, growth rhythm and tolerance performance, determine plant feature vectors, classify all plant feature vectors based on the decision tree model, and determine the plant combination structure according to the split node features. The community discrimination unit is configured to determine the distance parameters between different plant combinations based on the K-nearest neighbor algorithm, determine the stability trend and substitution relationship of each plant combination based on the distance parameters, and determine the community discrimination result based on the judgment result and the plant combination structure. The management suggestion unit is configured to analyze the vertical hierarchy, coverage relationship and color coordination relationship of the plants based on the plant combination structure, determine the hierarchical flower border structure layout, determine the association result between the community discrimination result and the flower border structure layout based on the association rule algorithm, and determine management suggestions based on the association result and the planning suggestion model.

[0016] It is understandable that the above-mentioned method and system for generating management suggestions for long-term, low-maintenance mixed flower borders have the same beneficial effects, and will not be elaborated further here. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating a method for generating management recommendations for a long-lasting, low-maintenance mixed flower border, as provided in an embodiment of the present invention; Figure 2 This is a functional block diagram of a long-lasting, low-maintenance mixed flower border management suggestion generation system provided in an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] See Figure 1 As shown in some embodiments of this application, a method for generating management recommendations for long-lasting, low-maintenance mixed flower borders includes: S100: During the growth cycle of the flower border, plant growth morphology and environmental characteristics of the sample plots are used to obtain plant growth data and community succession data, and the plant growth data and community succession data are used to construct a flower border data table.

[0022] S200: Perform parameter adjustment on the flower border data table, extract feature data on plant morphology, growth rhythm and tolerance performance, determine plant feature vectors, classify all plant feature vectors based on the decision tree model, and determine the plant combination structure based on the split node features.

[0023] S300: Based on the K-nearest neighbor algorithm, the distance parameters between different plant combinations are determined. Based on the distance parameters, the stability trend and substitution relationship of each plant combination are judged. Based on the judgment results and the plant combination structure, the community discrimination result is determined.

[0024] S400: Based on the plant combination structure, the vertical hierarchy, coverage relationship and color coordination relationship of the plants are analyzed to determine the hierarchical flower border structure layout. Based on the association rule algorithm, the association results of community discrimination and flower border structure layout are determined. Based on the association results and planning suggestion model, management suggestions are determined.

[0025] Specifically, a sample plot is a designated, specific mixed flower border area used for continuous observation, such as a mixed flower border plot with fixed boundaries in an urban green space. Spatial distribution characteristics refer to the location, spacing, and community aggregation of all plants within the sample plot, such as the difference between some plants being distributed in clumps and others in scattered distribution. Plant growth morphology characteristics represent the external physical state exhibited by plants during growth, encompassing visually observable features such as plant height, crown size, and leaf morphology. Environmental characteristics are the external objective conditions affecting plant growth within the sample plot, including light intensity, air temperature, and soil moisture. Plant growth data are continuously recorded information on changes in the growth status of individual plants or small groups of plants, such as the change in plant height and leaf spread of a shrub from budding to full bloom. Community succession data is information on changes in species composition and dominance of the flower border community over time, such as the gradual spread and coverage of a certain type of ground cover plant, the gradual decline of existing shrubs due to insufficient light, or the invasion of new weeds. The flower border data sheet integrates plant growth data with community succession data into a record-keeping medium. It uses time as a distinguishing element to archive growth data, succession status, and environmental conditions at different points in time. Within the complete growth cycle of the flower border—from budding, growth, flowering to withering—based on designated plots, it combines spatial distribution characteristics of plants within the plots (such as clump / scattered arrangement, plant spacing), plant growth morphology characteristics (such as plant height, crown width, and flowering status), and environmental characteristics (such as light, soil moisture, and temperature). On one hand, it collects plant growth data, continuously tracking changes in the growth status of each plant or group of plants. For example, it tracks the changes in plant height and leaf expansion of a group of roses from budding to full bloom. On the other hand, it collects community succession data, recording real-time species changes within the flower border community. For example, it tracks succession phenomena such as the gradual spread of liriope muscari covering surrounding weeds, the gradual decline of existing roses due to insufficient light, and the invasion of new dandelions. The data was then organized chronologically and by plant location to create a border data table, ensuring a clear correspondence between plant growth status, environmental conditions, and succession at each time point and location. Parameter normalization standardized and streamlined the disorganized and inconsistent raw information in the border data table. Plant morphological characteristic data reflects information about the static structural attributes of plants, such as the extent of crown expansion and leaf density—characteristics that are fixed or slowly changing. Growth rhythm characteristic data reflects information about the plant's growth patterns over time, such as the annual budding time, flowering cycle length, and seasonal differences in growth rate (e.g., rapid growth in spring, stagnation in summer). Tolerance characteristic data reflects the plant's adaptability to adverse environments (e.g., high temperature, drought, pests and diseases), such as whether the plant loses leaves or stagnates during drought periods. Plant feature vectors integrate the morphological, growth rhythm, and tolerance characteristic data of individual plants or small groups of plants, comprehensively reflecting their growth and adaptability.Splitting node features are used in decision tree models to classify plant categories. Each node corresponds to a selection criterion; for example, "drought tolerance," "growth rate," and "canopy expansion" can be used as splitting nodes at different levels. Plant combination structures group plants with matching characteristics that can coexist long-term into combinations. For example, shade-tolerant, slow-growing, and small-canopied plants like Hosta and Liriope are grouped together, while light-loving, fast-growing, and large-canopied plants like Rose and Rosa rugosa are grouped together. The flower border data table is parameterized, and feature data on plant morphology, growth rhythm, and tolerance are extracted to form plant feature vectors for each plant or group. A decision tree model is used to classify all plant feature vectors, using plant tolerance, growth rhythm, and morphological characteristics as splitting nodes at different levels. Based on node splitting, plant feature vectors with similar characteristics are grouped together, ultimately determining the plant combination structure to ensure that plants within each combination have matching characteristics and can coexist long-term.

[0026] Understandably, the K-Nearest Neighbors algorithm is an analytical method based on the logic that objects with similar characteristics tend to be close together. By judging the degree of similarity (distance) between different objects, it determines the association and development trend of the objects. The distance parameter reflects the index of the similarity of characteristics between different plant combinations. The smaller the distance parameter, the more similar the growth rhythm, tolerance, and morphological characteristics of the two groups of plants are, and the easier it is for them to coexist. The larger the distance parameter, the worse the coexistence compatibility. The stability trend reflects the ability of a plant combination to maintain its own structure and coexist harmoniously with surrounding combinations during long-term growth. The substitution relationship represents the complementary substitution potential between different plant combinations. When a combination cannot be maintained due to environmental changes or its own decline, another combination with similar characteristics and adapted to the environment of the area can replace its position. For example, after the decline of sun-tolerant combination A, combination B, which is sun-tolerant and has a similar growth rhythm, can fill its ecological position. Vertical hierarchy refers to the layered structure formed by plants of different heights within a flower field, typically divided into ground cover (below 30cm, such as daisies), middle layer (30cm-2m, such as crape myrtle), and upper layer (above 2m, such as cherry blossoms). Coverage relationship indicates the overlapping and shading of the canopies of different plant combinations, including complete coverage, partial coverage, and no coverage. For example, a shrub combination partially covering a ground cover combination can suppress weed growth. Color harmony refers to the visual compatibility of the flower and leaf colors of different plants. For example, warm-colored roses paired with cool-colored sage create clear visual hierarchy, while irises of similar colors paired with daylilies ensure stylistic unity. The distance parameters between plant combinations are calculated using the K-nearest neighbor algorithm. These parameters are then used to determine the stability trend and replacement relationships of each plant combination. Combinations with small distance parameters and long-term stability among neighboring combinations indicate strong compatibility with the surrounding environment and other combinations, showing good stability. Conversely, combinations with large fluctuations in distance parameters and frequent changes among neighboring combinations indicate relatively weak stability and a risk of decline. Furthermore, the replacement relationships and order between combinations are determined based on the magnitude of distance parameter changes. For example, if combination A declines, the nearest combination B will replace it first. Based on the determined plant combination structure, a comprehensive community identification result is obtained to clarify the overall stability, risk areas, and succession direction of the flower border community. Based on the plant combination structure, the vertical hierarchy, coverage relationship, and color coordination relationship of each plant are analyzed: the plant is divided into three levels—ground cover, middle layer, and upper layer—according to plant height. The hierarchical affiliation of each combination is clarified, and the results of vertical hierarchy, coverage relationship, and color coordination relationship are integrated to form a hierarchically related flower border structure layout, ensuring both ecological adaptability and landscape effect.Based on association rule algorithms, this study mines the correlation between community identification results and flower border structure layout, clarifying which layouts correspond to stable communities and which correspond to degradation risks. The correlation results are input into a planning suggestion model, which combines learned data patterns with the current state of the flower border to output targeted management suggestions. For example, for edge combinations with weak stability, it suggests replacing them with alternative combinations; for areas with insufficient coverage, it suggests adding mid-layer plants; and for areas with chaotic colors, it suggests adjusting flower color combinations. The K-nearest neighbor algorithm is used to accurately determine the stability and substitution relationships of plant combinations, solving the problems of difficulty in quantifying and assessing community status and the inability to identify degradation risks in advance. This allows for the early prediction of potential problems such as overexpansion of dominant species, extinction of weak species, and disordered landscape layers, effectively ensuring the long-term sustainability of the flower border landscape. Meanwhile, a hierarchical layout is formed through multi-dimensional analysis, taking into account both ecological stability and landscape aesthetics. By using association rule algorithms to establish a deep correlation between the layout and community status, the limitations of single-dimensional management are broken, and the precise transformation of the entire chain from community identification results to management suggestions is achieved. Based on reinforcement learning of decision tree model, K-nearest neighbor algorithm and planning suggestion model, the experience-based management mode is replaced and the blindness of maintenance work is avoided.

[0027] In some embodiments of this application, when acquiring plant growth data and community succession data based on the spatial distribution characteristics, plant growth morphology characteristics, and environmental characteristics of the sample plot during the flower border growth cycle, and constructing the plant growth data and community succession data into a flower border data table, the following steps are included: setting up spatial reference points and establishing several planar coordinate grids within the sample plot; marking the position of each individual plant against the planar coordinate grids; determining the coordinate number of each plant; collecting the growth morphology indicators and environmental monitoring indicators of the marked plants at each observation time; recording the plant distance between adjacent plants; and determining the plant distance and growth morphology indicators based on the plant distance and growth morphology indicators. Plant growth records were established, including morphological indicators such as plant height, crown width, leaf area, flowering status, and wilting status. Environmental monitoring indicators included light intensity, air temperature, soil moisture content, and wind speed. The appearance or disappearance of plants, changes in crown coverage, and changes in species composition were compared between adjacent observation times to identify community succession events, including the addition of new vegetation, the withdrawal of existing vegetation, and the transfer of dominant vegetation. The plant growth records, environmental monitoring indicators, and community succession events at each observation time were integrated in chronological order and by coordinate numbering to construct a flower border data table with time as the index and coordinate number as the field.

[0028] Specifically, spatial reference points are fixed, immovable markers (such as cement piles buried underground) set at the boundary or core location of the sample plot. These serve as reference points for positioning, avoiding deviations in location recording. The planar coordinate grid divides the sample plot into several uniformly sized rectangular grids (similar to a chessboard distribution) with the spatial reference points as the origin. Each grid corresponds to a unique coordinate range, and the coordinate number is a unique identifier assigned to each plant within its assigned grid. For example, the plant in the first grid to the right and the second grid in front of the reference point is numbered X1Y2, ensuring accurate tracking of plant locations. Growth morphology indicators reflect the external characteristics of the plant's growth status. Among these, plant height is the vertical distance from the ground to the top of the plant; crown width is the maximum horizontal diameter of the plant's crown; leaf area is the total area of ​​leaves on a single plant or branch; and flowering status includes non-flowering, newly opened, and flowering stages. During the flowering and withering stages, the withering state refers to the extent of withering of leaves and branches (e.g., partial withering, complete withering). Environmental monitoring indicators are external factors affecting plant growth. Light intensity is the intensity of sunlight received at this coordinate location. Air temperature is the real-time air temperature around the plant. Soil moisture content is the proportion of soil moisture content around the roots. Wind speed is the real-time wind force within the sample plot. Plant distance represents the straight-line distance between two adjacent plants. Plant growth record is a comprehensive record of the location, growth form, and surrounding environment of a single plant. Community succession events are the dynamic changes in species within the flower border community, including the addition of new vegetation (e.g., weed germination, new seedling growth), the withdrawal of existing vegetation (plant withering, death, and disappearance), and the transfer of dominant vegetation (the originally dominant vegetation is replaced by other vegetation, such as shrubs declining due to disease and being overtaken by herbaceous plants).

[0029] Understandably, a spatial reference point is set up within the sample plot, and several planar coordinate grids are divided based on this. The position of each plant is marked with its corresponding grid, and a unique coordinate number is assigned to each plant to ensure accurate location tracking. The observation time is within the complete growth cycle of the mixed flower border. Specific time points are pre-selected for concentrated observation and recording to accurately collect plant growth data, environmental monitoring indicators, and identify community succession events. For example, for a flower border dominated by annual herbaceous flowers, the observation time is from morning to noon and from noon to evening every day after the spring temperature has stabilized and risen. At each observation time, for the marked plants, their plant height, crown width, leaf area, flowering status, and other growth morphological indicators are recorded one by one. Simultaneously, the coordinates of the plant are collected. Environmental monitoring indicators such as light intensity, air temperature, and soil moisture content at the designated locations are recorded, along with the distance between adjacent plants. This data, combined with plant distance and growth morphology indicators, forms a plant growth record for each plant. Between two adjacent observation times, the presence or disappearance of plants, changes in the horizontal area of ​​canopy coverage, and alterations in the vegetation species composition within the area are compared. This allows for the identification of community succession events such as the addition of new vegetation, the withdrawal of existing vegetation, and the shift of dominant vegetation. The plant growth records, environmental monitoring indicators, and community succession events at each observation time are integrated according to time sequence and coordinate numbering to construct a flower border data table indexed by time and using coordinate number as a field, ensuring accurate correspondence of relevant data at each time point and coordinate location. The precise location and tracking of plant positions were achieved through spatial reference points and planar coordinate grids, solving the problems of ambiguous and easily confused plant positions during observation. Combined with multi-dimensional growth morphology indicators and environmental monitoring indicators, the plant growth status and environmental influencing factors were comprehensively captured. At the same time, community succession events were accurately identified by comparing changes at adjacent observation times, ensuring the integrity and relevance of data collection. The flower border data table achieved the orderly integration of various types of data, providing a precise data foundation for subsequent flower border analysis.

[0030] In some embodiments of this application, when performing parameter normalization on the flower border data table, extracting characteristic data of plant morphology, growth rhythm, and tolerance performance, and determining plant feature vectors, the process includes: merging plant growth records appearing at consecutive observation times within the same plane coordinate grid based on the time and coordinate number in the flower border data table, determining growth sequences at the unit of a single plant or plant combination; in each growth sequence, determining structural morphological parameters based on plant height, crown width, leaf area, and clump width, including average height, maximum height, crown width expansion, and leaf area index; determining growth rhythm parameters based on the plant height change, crown width expansion rate, and the time of flowering start, full bloom, and fruiting at adjacent observation times, including growth rate, seasonal change cycle length, and seasonal transition time interval; statistically analyzing tolerance performance parameters during the damage period, including the number of surviving plants, the degree of leaf damage, and the decrease in growth rate; and normalizing the structural morphological parameters, growth rhythm parameters, and tolerance performance parameters to determine the plant feature vector of a single plant or plant combination.

[0031] Specifically, a growth sequence is a complete time series formed by integrating the growth records of a single plant or a group of plants with similar habits within the same planar coordinate grid, based on continuous observations. For example, three *Liriope muscari* plants within the same grid are considered a group, and their growth status is continuously observed to form the growth sequence of this group. Structural morphological parameters reflect indicators of the plant's static growth structure. Among these, average height is the average plant height at different observation times within the growth sequence; maximum height is the height of the tallest plant within the growth sequence; crown spread is the difference between the crown spread at the end of the growth sequence and the initial crown spread; and leaf area index is the ratio of the total leaf area of ​​the plant per unit land area to the corresponding land area (not the area of ​​a single leaf). Rather, it is an indicator of the total leaf cover. Growth rhythm parameters reflect the dynamic growth patterns of plants. Growth rate is the ratio of the change in plant height between adjacent observation times to the observation interval. Seasonal change cycle length is the complete duration from the beginning of flowering to the end of fruiting. Seasonal transition interval is the time difference between adjacent seasons (e.g., the interval from the beginning of flowering to full bloom, or from full bloom to fruiting). Damage period is the period when the flower border encounters unfavorable environmental conditions such as high temperature, low temperature, or low soil moisture content (e.g., periods of sustained high temperatures above 35℃ in summer, periods of low temperatures below 0℃ in winter, or periods of prolonged dry soil moisture content in spring and autumn). Tolerance performance parameters are the plant's resistance to adverse conditions during the damage period. The number of surviving plants is the number of plants that survive after the damage period ends. The number of surviving individuals in a single plant or plant combination; the degree of leaf damage is defined as the range of yellowing, wilting, and scorching of the leaves (e.g., slight damage with only leaf tips scorched, moderate damage with half of the leaves yellowing, severe damage with most leaves wilting); and the rate of decline in growth rate is the ratio of the plant's growth rate during the damage period to its growth rate during the normal environment period. Based on the time and coordinate number in the flower border data table, a growth sequence is determined for each individual plant (e.g., a single rose) or plant combination (e.g., several Liriope muscari plants in the same grid). This ensures that each sequence fully reflects the continuous growth process of the corresponding plant or combination. Within each growth sequence, combined with data such as plant height, crown width, and leaf area, structural parameters such as average height, maximum height, and crown width expansion are determined. Morphological parameters are used to accurately depict the plant's population structure and static characteristics. By analyzing changes in plant height, crown expansion rate, and the timing of flowering initiation, peak flowering, and fruiting at adjacent observation times, growth rhythm parameters such as growth rate and the length of seasonal phase change cycles are determined. This allows for a clear capture of the temporal patterns of plant growth speed and seasonal phase transitions. For example, the 7-day interval between the start of flowering and peak flowering of a rose is the seasonal phase transition time interval. Subsequently, for periods of damage caused by high temperature, low temperature, or low water content, the tolerance parameters of corresponding individual plants or plant combinations are statistically analyzed. For example, after a period of high summer temperatures, the number of surviving plants in a Liriope muscari combination, the degree of leaf scorch, and the decrease in growth rate compared to the normal temperature period are statistically analyzed.Structural morphology parameters, growth rhythm parameters, and tolerance performance parameters are normalized to eliminate unit differences between different parameters, determining the plant characteristic vectors for individual plants or plant combinations, thereby achieving precise quantification of the plant's comprehensive characteristics. Growth sequences are formed using time and coordinate numbering, ensuring data continuity and specificity, avoiding feature extraction bias caused by scattered data. Tolerance indicators are statistically analyzed for high temperature, low temperature, and low moisture content damage periods, comprehensively covering the static, dynamic, and stress resistance characteristics of plant growth, avoiding the limitations of analysis focusing only on a single growth indicator. Normalization allows for the synergistic analysis of different types of parameters, enabling the plant characteristic vectors to accurately and comprehensively quantify the plant's comprehensive characteristics, avoiding analytical errors caused by parameter confusion or incomplete characteristic characterization.

[0032] In some embodiments of this application, when classifying all plant feature vectors based on a decision tree model and determining the plant combination structure based on the splitting node features, the process includes: constructing a plant feature vector set from all plant feature vectors; using structural morphology parameters, growth rhythm parameters, and tolerance performance parameters as candidate splitting attributes in the plant feature vector set; using the candidate splitting attributes as the splitting basis; dividing the plant feature vector set into several attribute subsets based on the differences in different attributes; determining splitting nodes and node splitting paths based on the distribution changes of structural morphology differences, growth rhythm differences, and tolerance characteristic differences within each attribute subset; classifying each plant feature vector into the corresponding splitting node level by level according to the hierarchical order of the node splitting path; determining a plant category tree; determining the set of plant feature vectors at the same end node in the plant category tree; constructing a plant combination group with the end node as the category boundary; comparing the distribution of the plant combination group with the distribution in the entire category tree to determine the similarity; merging combination groups with similarity greater than or equal to the similarity threshold to determine the plant combination structure.

[0033] Specifically, candidate splitting attributes are parameters selected from plant feature vectors for classifying categories. These include structural morphology parameters, growth rhythm parameters, and tolerance performance parameters. For example, drought tolerance (tolerance performance parameter), growth rate (growth rhythm parameter), and crown spread (structural morphology parameter) can all be candidate splitting attributes. Attribute subsets are several subsets formed by splitting the plant feature vector set according to attribute differences based on candidate splitting attributes. For example, using drought tolerance as the splitting attribute, the plant feature vector set can be split into two attribute subsets: drought-tolerant and drought-intolerant. Splitting nodes are nodes used to classify categories in the decision tree model. Each node corresponds to a candidate splitting attribute. The node splitting path is the path formed by the various splitting nodes and attribute selection logic that the plant feature vector passes through from the root node to the terminal node of the decision tree. For example, drought tolerance → slow growth → small crown spread is a node splitting path. The plant category tree is formed by hierarchical splitting. The resulting tree-like plant classification system has a root node representing the complete set of plant feature vectors, branches at each level representing attribute subsets, and terminal nodes representing the final classification. Terminal nodes are the lowest-level nodes in the plant category tree, each corresponding to a set of plant feature vectors with highly similar characteristics. A plant combination group is formed by aggregating all plant feature vectors from the same terminal node. Plants within a combination group have similar overall characteristics; for example, the feature vectors of *Liriope muscari* and *Hosta* from the same terminal node can form a combination group. The hierarchical span is the difference in hierarchical level between two plant combination groups in the plant category tree; the smaller the hierarchical span, the closer the classification levels of the two groups. Similarity is the degree of matching between two plant combination groups in the plant category tree, determined based on hierarchical span and splitting path. The similarity threshold is set to 0.8. The plant combination structure is the plant combination system formed by merging plant combination groups that meet the similarity threshold.

[0034] Understandably, all plant feature vectors are integrated to construct a plant feature vector set. From this set, structural morphology parameters, growth rhythm parameters, and tolerance performance parameters are selected as candidate splitting attributes. Based on these candidate splitting attributes, and according to the differences in attributes among different vectors, the plant feature vector set is divided into several attribute subsets. The distribution and changes in structural morphology, growth rhythm, and tolerance performance within each attribute subset are analyzed. If the differences within a certain attribute subset are still significant, the corresponding candidate splitting attribute is used as a new splitting node, thereby determining the splitting nodes at each level and the node splitting path. For example, drought tolerance is first used as the root node to split into drought-tolerant and drought-intolerant attribute subsets. Then, the drought-tolerant attribute subset is split into slow-growing and fast-growing attribute subsets using growth rate as the splitting node, and so on, forming a multi-level node distribution. The splitting path method assigns each plant feature vector to its corresponding splitting node in a hierarchical order, ultimately forming a complete plant category tree. Within this tree, plant feature vectors at the same terminal node are grouped together to construct plant combination groups with the terminal node as the category boundary. Plants within each combination group exhibit highly similar characteristics. Based on the hierarchical span and splitting path, the distribution of each plant combination group is compared with that in the entire category tree to determine similarity. First, the hierarchical span of the terminal nodes of the two groups is compared; a smaller hierarchical span indicates a higher basic similarity. Next, the node splitting paths of the two groups are compared; the more overlapping splitting nodes in the paths and the more consistent the attribute selection logic, the higher the similarity. Combining the hierarchical span and path overlap, the final similarity between the two groups is determined. Similarity can be determined using methods such as the cosine theorem and Euclidean distance. By selecting multi-dimensional candidate splitting attributes to construct a decision tree model, reinforcement learning is used to achieve refined classification of plant feature vectors, replacing the traditional experience-based plant combination method. This avoids the blindness of combination configuration. The similarity is determined based on the hierarchical span and splitting path, ensuring that the plant characteristics in the merged plant combination group are highly compatible, and guaranteeing the compatibility of plants in each combination in terms of structural morphology, growth rhythm, and tolerance performance.

[0035] In some embodiments of this application, when determining the distance parameters between different plant combinations based on the K-nearest neighbor algorithm, judging the stability trend and substitution relationship of each plant combination based on the distance parameters, and determining the community discrimination result based on the judgment result and the plant combination structure, the process includes: merging the plant feature vectors of nodes in the same category in the plant combination structure to determine the combination feature vector; calculating the vector difference between every two combination feature vectors in the feature space; determining the distance parameter based on the vector difference; arranging the distance parameters in the feature space in ascending order to determine several neighboring combinations; comparing the neighboring combinations of the same plant combination to determine the persistence of the neighboring relationship of the plant combination; counting the number of differences between neighboring combinations at different stages to determine the volatility of the neighboring relationship of the plant combination; determining the stability trend of each plant combination based on persistence and volatility; determining the substitution relationship of each plant combination in the structural hierarchy based on the distance change range between adjacent plant combinations within the neighboring combination; determining the substitution order based on the continuity interval; and determining the community discrimination result based on the stability trend, substitution relationship, and substitution order.

[0036] Specifically, the feature space is an abstract space that carries all combined feature vectors, and each combined feature vector corresponds to a unique position in the feature space. The distance parameter is an index corresponding to the difference between any two combined feature vectors in the feature space. The smaller the distance parameter, the more similar the comprehensive characteristics of the two plant combinations are. The persistence of the proximity relationship of plant combinations is the maintenance of the neighboring combinations of the same plant combination at different growth stages. If the neighboring combinations are basically consistent at different stages, it indicates stronger persistence, and vice versa. The volatility of the proximity relationship of plant combinations is the number of differences between the neighboring combinations of the same plant combination at different growth stages. The more differences, the stronger the volatility, and vice versa. The stability trend is the ability of a plant combination to maintain its own structure and characteristics and coexist harmoniously with the surrounding combinations during long-term growth. It is determined by both persistence and volatility. The substitution relationship is the potential association that other plant combinations with similar characteristics can fill the ecological niche when a plant combination is difficult to maintain due to environmental changes or its own decline. The duration interval is the time range from the beginning of the substitution relationship to its complete formation. The feature vectors of all plants belonging to the same category in the plant combination structure are merged to form a combination feature vector that represents the overall characteristics of each plant combination. Within the feature space containing all combination feature vectors, the vector difference is calculated for every two combination feature vectors; this distance parameter reflects the similarity of their characteristics. All distance parameters are arranged in ascending order, and the value of K can be dynamically adjusted according to the number of distance parameters. Based on the K value, the plant combinations corresponding to the first K distance parameters after sorting are selected. These selected plant combinations are the nearest neighbors with the most similar plant combination characteristics. For example, if K is pre-set to 3, the plant combinations corresponding to the first 3 distance parameters after sorting are selected as the nearest neighbors. If the plant combination is *Liriope muscari*, its distance parameter ranking with *Hosta*, *Liriope spicata*, *Iris*, and *Rosa* is *Liriope muscari* - *Hosta* < *Liriope muscari* - *Liriope spicata* < *Liriope muscari* - *Iris* < *Liriope spicata* - *Rosa*. If K is set to 2, the selected nearest neighbors are *Hosta* and *Liriope spicata*.For the same plant combination, its neighboring combinations at different growth stages are compared to determine the persistence of the neighboring relationship. For example, if a certain *Liriope spicata* combination is consistently paired with *Hosta* in spring, summer, and autumn, its neighboring combination is strong. Simultaneously, the number of differences between neighboring combinations at different stages is counted to determine the volatility of the neighboring relationship. For example, if the neighboring combination of *Liriope spicata* changes to *Ophiopogon japonicus* in winter (a difference of 1), its volatility is weak. Based on persistence and volatility, the stability trend of each plant combination is determined. Plant combinations with strong persistence and weak volatility in neighboring relationships have better stability, while those with weaker persistence and volatility have poorer stability and are prone to decline. To mitigate risk, the substitution relationship of each plant combination in the structural hierarchy is determined based on the range of distance variation between adjacent plant combinations. The smaller the range of distance variation, the stronger the substitution suitability of the adjacent combination. At the same time, the substitution order is determined based on the duration of the substitution relationship from its manifestation to its formation. The shorter the duration of the adjacent combination, the higher the priority of replacement. By combining the stability trend, substitution relationship, and substitution order of each plant combination, the community discrimination result for the overall state and development trend of the flower border community is determined. For example, if the plant combination in the central area has a good stability trend and no substitution demand, while a certain combination in the peripheral area has a poor stability trend, its optimal substitution combination is a certain adjacent combination. By merging plant feature vectors into combined feature vectors, a leap from single-plant or small-combination analysis to community-level analysis is achieved, which aligns with the actual growth characteristics of flower border communities. Based on the K-nearest neighbor algorithm, distance parameters are calculated and neighboring combinations are screened, providing a quantitative basis for the stability and substitution relationship analysis of plant combinations. By analyzing the persistence and volatility of neighbor relationships, stability trends are determined, and substitution relationships and order are determined by combining the magnitude of distance changes and the duration interval. This allows for the early prediction of community succession risks such as the overexpansion of dominant species and the extinction of weak species. The final community discrimination results can accurately reflect the overall state of the flower border community.

[0037] In some embodiments of this application, when analyzing the vertical hierarchy, coverage relationship, and color coordination relationship of plants based on the plant combination structure to determine the hierarchical flower border structure layout, the process includes: extracting the height parameters from the feature vectors of each plant and arranging the height parameters in ascending order to determine several height intervals; mapping the several height intervals to the planar coordinate grid of the sample plot; determining the spatial hierarchy diagram of each plant combination; determining the coverage overlap ratio between each plant combination based on the coverage range and crown extension of the plant combinations at adjacent levels in the spatial hierarchy diagram; dividing the dominant coverage combination and subordinate coverage combination based on the coverage overlap ratio; and determining the coverage relationship structure diagram.

[0038] In some embodiments of this application, when analyzing the vertical hierarchy, coverage relationship, and color coordination relationship of plants based on the plant combination structure to determine the hierarchical flower border structure layout, the method further includes: extracting the color attributes corresponding to each plant combination, determining the color distribution points based on hue, brightness, and saturation, determining the color transition sequence based on the distance between the color distribution points, determining the color arrangement order between each plant combination based on the color transition sequence, superimposing the color arrangement order onto the coverage relationship structure diagram, determining the coverage-color correspondence structure based on the color distance between each dominant coverage combination and adjacent subordinate coverage combinations, and superimposing the coverage-color correspondence structure onto the spatial hierarchy diagram to determine the hierarchical position of different height intervals in the superimposed space.

[0039] Specifically, a height interval refers to several continuous height ranges divided by sorting height parameters from low to high, namely, the ground cover layer height interval, the middle layer height interval, and the upper layer height interval. A spatial hierarchy map is a visual representation of the vertical distribution of each plant combination, formed by mapping height intervals onto a plot's planar coordinate grid. Coverage is the projected area of ​​the plant combination's crown on the plane, crown extension is the edge boundary of the plant combination's crown projection, and cover overlap ratio is the proportion of the overlapping area of ​​adjacent plant combinations to the coverage area of ​​a particular plant combination. The dominant cover combination is the plant combination with a larger coverage area and a higher cover overlap ratio, occupying a dominant position in space. Cover subordinate groups... A combination refers to a plant combination with a small coverage area and a low overlap ratio, growing dependent on the dominant mulch. The mulch relationship structure diagram reflects the dominant and subordinate relationships among the plant combinations. Color attributes are the color characteristics of the flowers or leaves of the plant combinations; hue represents the basic appearance of a color, such as red, green, and yellow; brightness is the lightness or darkness of a color; saturation is the vividness or saturation of a color; color distribution points are spatial points formed by mapping color attributes according to the three dimensions of hue, brightness, and saturation onto the color space; the color transition sequence is a gradual sequence of colors from light to dark or from warm to cool, formed by sorting the color distribution points according to the distance between them. The height parameter is extracted from the feature vector of each plant, and the height parameter is... The plants are arranged in ascending order and divided into several height intervals. Using a GIS (Geographic Information System), these height intervals are mapped onto a planar coordinate grid of the sample plots. ArcGIS is then used to generate a spatial hierarchy map, visually presenting the vertical distribution of each plant combination within the grid. For example, the height intervals of ground cover (0-30cm), middle layer (30cm-2m), and upper layer (above 2m) are mapped onto the coordinate grid of the park's flower border sample plots. ArcGIS automatically renders the colors of different height intervals, clearly displaying the vertical distribution of *Liriope muscari* (ground cover), *Lagerstroemia indica* (middle layer), and *Prunus cerasifera* (upper layer). Based on ArcGIS, the coverage of adjacent plant combinations in the spatial hierarchy map is extracted. Based on the canopy extension, the overlap ratio between each combination is calculated. Combinations are categorized as dominant or subordinate based on this overlap ratio. For example, shrub combinations with a large coverage area and a high overlap ratio with ground cover combinations are considered dominant, while ground cover combinations are subordinate. Coordinate grids and plant combinations are imported using SketchUp and Enscape to help determine vertical distribution and coverage, ensuring the reliability of the cover relationship structure map generated by ArcGIS. Color attributes corresponding to each plant combination are extracted. Based on hue, lightness, and saturation, the color attributes are transformed into color distribution points in an abstract color space using the HSV color space algorithm. The distances between these color distribution points are calculated using tools such as Adobe Photoshop and GIMP to determine the color transition sequence, which is ordered from smallest to largest distance between color distribution points.The colors are arranged in a gradient from similar to gradually changing. Closer distances indicate a more natural transition, while greater distances indicate greater color differences and a more abrupt transitions. The color attributes in the color arrangement are precisely assigned to each dominant and adjacent subordinate combination in the coverage relationship diagram according to the correspondence between plant combinations. For example, the dominant combination is the rose combination, corresponding to warm red in the color arrangement, and its adjacent subordinate combination is the lilyturf combination, corresponding to emerald green in the color arrangement. Next, the color distance between each dominant combination and its adjacent subordinate combination is determined to verify the harmony of the color scheme. Too small a distance will result in low color recognition and inability to distinguish coverage levels, while too large a distance will cause color conflict and damage the visual effect. For example, the emerald green of the lilyturf combination is slightly adjusted to a light yellowish-green to make the distance between it and the warm red of the rose moderate. Finally, the coverage relationship is further refined. In the composition, overlapping areas are determined by color transition rules based on the colors of the dominant and subordinate overlay combinations. For example, a gradient from the dominant to the subordinate color is used to avoid color clutter in the overlapping areas. The overlay relationships and color matching rules, after color matching, coordination verification, and overlapping area processing, are integrated to determine the color scheme, color distance parameters, and overlapping area color rules for each dominant-subordinate combination pair. Simultaneously, using software such as ArcGIS and QGIS, the overlay relationship fields and color fields are linked by attributes, or corresponding materials are assigned in SketchUp for visualization verification. This ultimately forms an overlay-color correspondence structure that combines overlay relationships and color coordination. Layer overlay algorithms are then used to overlay the overlay-color correspondence structure with the spatial hierarchy map to clarify the hierarchical positions of different height ranges in the overlaid space, ultimately forming a hierarchically related flower border structure layout.

[0040] Understandably, GIS software enabled precise spatial mapping of height ranges, ensuring the reliability of the hierarchical distribution of various plant combinations. The overlap ratio of coverage was determined based on the coverage range and crown extension, clarifying the dominant and subordinate relationships of coverage, improving the ecological stability of the flower border community, and avoiding the overexpansion of dominant species. The HSV color space algorithm and layer overlay algorithm were used to achieve quantitative analysis and structural overlay of colors, thus taking into account the landscape aesthetics of the flower border, making the color matching more harmonious and natural, and achieving full-dimensional quantitative analysis from vertical hierarchy, coverage relationship to color coordination.

[0041] In some embodiments of this application, when determining the association results between community discrimination results and flower border structure layout based on association rule algorithms, the process includes: generating several candidate itemsets from hierarchical positions and community discrimination results using the Eclat algorithm; determining frequent itemsets based on the support of the candidate itemsets; determining association rules based on the frequent itemsets; and determining the association results between flower border structure layout and community discrimination results based on the association rules.

[0042] Specifically, the Eclat algorithm is an association rule mining algorithm. Hierarchical position refers to the specific location of different height intervals in the superimposed space, that is, the precise location of each plant combination in the flower border structure layout within the vertical layers (such as ground cover, middle layer, and upper layer) and the planar coordinate grid. The community discrimination result comprehensively assesses the stability trend, substitution relationship, and substitution order of plant combinations, making a comprehensive judgment on the overall state, development trend, and potential risks of the flower border community. Using the hierarchical position and community discrimination results in the flower border dataset as the analysis object, based on the vertical data processing logic of the Eclat algorithm, various feature items of the hierarchical position (such as ground cover, middle layer, and upper layer position) and various feature items of the community discrimination result (such as good stability trend, poor stability trend, existence of substitution demand, and priority of substitution order) are combined in pairs and multiple groups to generate several candidate option sets containing different feature combinations. For example, {middle layer position, poor stability trend} and {upper layer position, no substitution demand} are generated as candidate option sets. The frequency of each candidate option set in all sample data is counted. Frequent itemsets are selected based on support levels. From these itemsets, association rules reflecting the causal or correlational relationship between hierarchical location and community classification results are extracted. For example, from the frequent itemset {upper-level location, good stability trend}, association rules can be extracted indicating that if a plant combination is in an upper-level location, its stability trend is good; from {ground cover layer location, existence of substitution demand}, association rules can be extracted indicating that if a plant combination is in a ground cover layer location, its substitution demand exists. Based on these statistically validated association rules, the intrinsic connection between the flower border structure layout (hierarchical location) and community classification results is identified, thus determining specific association patterns. The efficient intersection operation logic of the Eclat algorithm improves the efficiency of candidate itemset generation and frequent itemset selection, avoiding erroneous associations caused by chance. This allows the association results to clearly reveal the intrinsic patterns between the hierarchical location of the flower border structure layout and the community state (community classification results), providing a data-driven basis for targeted optimization of flower border structure layout and early avoidance of community succession risks.

[0043] In some embodiments of this application, when determining management recommendations based on association results and planning recommendation models, the process includes: acquiring a flower border dataset and dividing the flower border dataset into a training set and a test set; finding model building parameters and building a decision model based on grid search; training the decision model based on the training set; and testing the trained decision model based on the test set, ultimately determining a planning recommendation model with association results as input and management recommendations as output.

[0044] Specifically, the flower border dataset includes data on geographical location, topography, soil type, soil fertility, light duration distribution, annual average temperature and humidity, precipitation, family and genus classification of all plants within the flower border, growth habits, plant type characteristics, historical management measures for different flower border problems (such as plant combination replacement, water and fertilizer regulation, pruning frequency adjustment, and pest and disease control programs), implementation time and cycle of management measures, and feedback on the effects after implementation. The flower border dataset can be obtained through public geographic information service platforms, meteorological data networks, soil survey databases, the official website of urban greening associations, and technical regulations documents from landscaping bureaus. The flower border dataset is divided into training and testing sets in a 4:1 ratio to ensure the model's generalization ability. The training set allows the model to learn patterns and features in the data, while the testing set tests the model's generalization ability, preventing overfitting and enabling it to better cope with different data variations. Grid search iterates through parameters, verifying the model's effectiveness for each parameter group to select the optimal parameter combinations. A decision model is built based on this grid search. The training set is input into the decision model, which repeatedly learns the logic and patterns between data during training. The test set is then input into the trained decision model to test its accuracy, mean squared error, and other metrics until convergence, resulting in a planning suggestion model. Substituting the association results into this model yields management recommendations. Association rule algorithms uncover the potential correlations between flower border structure layout (hierarchical location) and community status (community discrimination results). Flower border management recommendations must simultaneously adapt to ecological stability and landscape longevity. Relying solely on experience or a single data dimension to derive management recommendations will lack data support and fail to accurately match the dynamic changes in the flower border. For example, knowing only that ground cover plants have poor stability (community discrimination results) without understanding their connection to the imbalance in ground cover hierarchy (hierarchical location characteristics) leads to management recommendations that blindly replace plants, failing to address the root cause of the problem. By clarifying the correlation between certain hierarchical location features and certain community discrimination results through association rule algorithms, management suggestions can accurately address the root causes of flower border problems, rather than merely maintaining the surface. This avoids the limitations of traditional management suggestions that rely on experience and lack logical support. It achieves full-chain data-driven management from association rules, model reinforcement learning, and precise suggestions. The management suggestions can adapt to the structural layout characteristics of flower borders and address community succession risks in a targeted manner, thus balancing the long-term effectiveness and low maintenance requirements of flower borders and ensuring the level of intelligence in flower border management.

[0045] In summary, the beneficial effects of this invention are as follows: By collecting spatial distribution, plant growth morphology, and environmental characteristic data and constructing a flower border data table, it avoids the traditional experience-based approach and provides multi-dimensional data support for plant combination configuration. This allows for accurate prediction of the long-term coexistence stability of plant combinations, avoiding risks such as overexpansion of dominant species, extinction of weak species, and disordered landscape hierarchy, thereby ensuring the integrity of the flower border community structure and landscape stability. Based on the classification of plant feature vectors using a decision tree model and the quantification of plant combination distance parameters using the K-nearest neighbor algorithm, it achieves a leap from qualitative description to quantitative analysis in flower border community assessment, accurately determining the trend of combination stability and substitution relationships, thereby identifying community degradation risks in advance and providing a basis for the long-term maintenance of flower borders. By analyzing the vertical layers, coverage, and color relationships of plants to determine the hierarchical layout, and combining the association rule algorithm to establish the relationship between community discrimination results and structural layout, a full-chain transformation from data to plant combination structure and community discrimination results to management suggestions is realized. This avoids single-dimensional experience-based management, and the output control suggestions can reduce the blindness of maintenance, thereby reducing the input of human and material resources and improving the long-term effectiveness of flower border landscape and the level of intelligent management.

[0046] In another preferred embodiment based on the above embodiments, see [reference] Figure 2 As shown, this embodiment provides a management suggestion generation system for long-lasting, low-maintenance mixed flower borders, used to apply a management suggestion generation method for long-lasting, low-maintenance mixed flower borders, including: The data acquisition unit is configured to acquire plant growth data and community succession data based on the spatial distribution characteristics of the sample plot, plant growth morphology characteristics, and environmental characteristics during the growth cycle of the flower border, and to construct the plant growth data and community succession data into a flower border data table.

[0047] The landscape structure unit is configured to perform parameter normalization on the flower border data table, extract characteristic data of plant morphology, growth rhythm and tolerance performance, determine plant feature vectors, classify all plant feature vectors based on the decision tree model, and determine the plant combination structure based on the characteristics of the split nodes.

[0048] The community discrimination unit is configured to determine the distance parameters between different plant combinations based on the K-nearest neighbor algorithm, judge the stability trend and substitution relationship of each plant combination based on the distance parameters, and determine the community discrimination result based on the judgment result and the plant combination structure.

[0049] The management suggestion unit is configured to analyze the vertical hierarchy, coverage relationship and color coordination relationship of plants based on the plant combination structure, determine the hierarchical flower border structure layout, determine the association result between the community discrimination result and the flower border structure layout based on the association rule algorithm, and determine management suggestions based on the association result and the planning suggestion model.

[0050] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program goods according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0051] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for generating management recommendations for long-term, low-maintenance mixed flower borders, characterized in that, include: During the growth cycle of the flower border, plant growth data and community succession data were obtained based on the spatial distribution characteristics, plant growth morphology characteristics and environmental characteristics of the sample plots, and the plant growth data and community succession data were constructed into a flower border data table. The parameters of the flower border data table are adjusted, the characteristic data of plant morphology, growth rhythm and tolerance are extracted, the plant feature vectors are determined, all plant feature vectors are classified based on the decision tree model, and the plant combination structure is determined according to the split node characteristics. The distance parameters between different plant combinations are determined based on the K-nearest neighbor algorithm. The stability trend and substitution relationship of each plant combination are judged based on the distance parameters. The community discrimination result is determined based on the judgment result and the plant combination structure. Based on the plant combination structure, the vertical hierarchy, coverage relationship and color coordination relationship of the plants are analyzed to determine the hierarchical flower border structure layout. Based on the association rule algorithm, the association result between the community discrimination result and the flower border structure layout is determined. Based on the association result and the planning suggestion model, management suggestions are determined.

2. The method for generating management recommendations for long-lasting, low-maintenance mixed flower borders according to claim 1, characterized in that, When acquiring plant growth data and community succession data based on the spatial distribution characteristics, plant growth morphology characteristics, and environmental characteristics of the sample plots during the flower border growth cycle, and constructing the plant growth data and community succession data into a flower border data table, the following is included: Spatial reference points were set up within the sample plot and several planar coordinate grids were established. The position of each individual plant was marked with the planar coordinate grids, and the coordinate number of each plant was determined. At each observation time, the growth morphology indicators and environmental monitoring indicators of the marked plants are collected, and the plant distance between adjacent plants is recorded. Based on the plant distance and growth morphology indicators, plant growth records are determined. The growth morphology indicators include plant height, crown width, leaf area, flowering status and wilting status. The environmental monitoring indicators include light intensity, air temperature, soil moisture content and wind speed. By comparing the appearance or disappearance of plants, changes in canopy coverage, and changes in species composition between adjacent observation times, community succession events are determined. These community succession events include the addition of new vegetation, the withdrawal of existing vegetation, and the transfer of dominant vegetation. Plant growth records, environmental monitoring indicators, and community succession events at each observation time were integrated in chronological order and by coordinate number to construct a flower border data table with time as the index and coordinate number as the field.

3. The method for generating management recommendations for long-lasting, low-maintenance mixed flower borders according to claim 2, characterized in that, When performing parameter normalization on the flower border data table, extracting characteristic data on plant morphology, growth rhythm, and tolerance performance, and determining plant feature vectors, the following steps are included: Based on the time and coordinate number in the flower border data table, plant growth records appearing at consecutive observation times within the same plane coordinate grid are merged to determine the growth sequence of a single plant or plant combination. In each growth sequence, structural morphological parameters are determined based on plant height, crown width, leaf area, and clump width. The structural morphological parameters include average height, maximum height, crown width expansion, and leaf area index. Based on the changes in plant height, the rate of crown expansion, and the timing of flowering start, full bloom, and fruiting at adjacent observation times, growth rhythm parameters are determined. These growth rhythm parameters include growth rate, length of seasonal change cycle, and seasonal transition time interval. The tolerance performance parameters during the injury period were statistically analyzed, including the number of surviving plants, the degree of leaf damage, and the rate of decline in growth rate. The structural morphology parameters, growth rhythm parameters, and tolerance performance parameters are normalized to determine the plant feature vectors of individual plants or plant combinations.

4. The method for generating management recommendations for long-lasting, low-maintenance mixed flower borders according to claim 3, characterized in that, When classifying all plant feature vectors based on a decision tree model and determining the plant combination structure based on the features of split nodes, the following steps are included: All plant feature vectors are constructed into a plant feature vector set. In the plant feature vector set, structural morphology parameters, growth rhythm parameters, and tolerance performance parameters are used as candidate splitting attributes. The candidate splitting attributes are used as the splitting basis. The plant feature vector set is divided into several attribute subsets according to the differences of different attributes. Based on the distribution changes of structural morphology differences, growth rhythm differences, and tolerance characteristic differences within each attribute subset, the splitting nodes and node splitting paths are determined. Based on the hierarchical order of the node splitting path, each plant feature vector is assigned to the corresponding splitting node level by level to determine the plant category tree. In the plant category tree, the set of plant feature vectors located at the same end node is determined, and a plant combination group with the end node as the category boundary is constructed. The distribution of the plant combination group is compared with the distribution in the entire category tree to determine the similarity. Combination groups with similarity greater than or equal to the similarity threshold are merged to determine the plant combination structure.

5. The method for generating management recommendations for long-lasting, low-maintenance mixed flower borders according to claim 4, characterized in that, When determining the distance parameters between different plant combinations based on the K-nearest neighbor algorithm, judging the stability trend and substitution relationship of each plant combination based on the distance parameters, and determining the community discrimination result based on the judgment result and the plant combination structure, the process includes: The plant feature vectors of the same category nodes in the plant combination structure are merged to determine the combined feature vector. The vector difference is calculated for every two combined feature vectors in the feature space, and the distance parameter is determined based on the vector difference. The distance parameters in the feature space are arranged in ascending order to determine several neighboring combinations. The neighboring combinations of the same plant combination are compared to determine the persistence of the neighboring relationship of the plant combination. The number of differences in neighboring combinations at different stages is counted to determine the volatility of the neighboring relationship of the plant combination. Based on the aforementioned persistence and volatility, the stability trend of each plant combination is determined. According to the distance variation between adjacent plant combinations within a neighboring combination, the substitution relationship of each plant combination in the structural hierarchy is determined, and the substitution order is determined according to the continuity interval. The community discrimination result is determined based on the stability trend, substitution relationship, and substitution order.

6. The method for generating management recommendations for long-lasting, low-maintenance mixed flower borders according to claim 5, characterized in that, When analyzing the vertical hierarchy, coverage, and color coordination of plants based on the plant combination structure to determine the hierarchical flower border layout, the following steps are included: Extract the height parameters from the feature vectors of each plant, and arrange the height parameters in ascending order to determine several height intervals; Several height intervals are mapped to the planar coordinate grid of the sample plot to determine the spatial hierarchy of each plant combination. Based on the coverage range and crown extension of the plant combinations at adjacent levels in the spatial hierarchy, the coverage overlap ratio between each plant combination is determined. Based on the coverage overlap ratio, the dominant and subordinate combinations are divided, and the coverage relationship structure diagram is determined.

7. The method for generating management recommendations for long-lasting, low-maintenance mixed flower borders according to claim 6, characterized in that, When analyzing the vertical hierarchy, coverage, and color coordination of plants based on the plant combination structure to determine the hierarchical flower border layout, the method also includes: Extract the color attributes corresponding to each plant combination, determine the color distribution points based on hue, brightness and saturation, determine the color transition sequence based on the distance between the color distribution points, and determine the color arrangement order between each plant combination based on the color transition sequence; The color arrangement order is superimposed on the coverage relationship structure diagram. Based on the color distance between each dominant coverage combination and the subordinate coverage combination, the coverage-color correspondence structure is determined. The coverage-color correspondence structure is then superimposed on the spatial hierarchy diagram to determine the hierarchical position of different height intervals in the superimposed space.

8. The method for generating management recommendations for long-lasting, low-maintenance mixed flower borders according to claim 7, characterized in that, When determining the association results between the community discrimination results and the flower border structure layout based on the association rule algorithm, the following are included: The Eclat algorithm is used to generate several candidate item sets based on the hierarchical location and community discrimination results. Frequent itemsets are determined based on the support of the candidate item sets. Association rules are determined based on the frequent itemsets. Association results are determined based on the association rules.

9. The method for generating management recommendations for long-lasting, low-maintenance mixed flower borders according to claim 8, characterized in that, When determining management recommendations based on the aforementioned correlation results and planning recommendation model, the following are included: Obtain the flower border dataset and divide the flower border dataset into a training set and a test set; The model establishes parameters and builds a decision model based on grid search. The decision model is trained on the training set and tested on the test set. Finally, the input is determined to be the association result, and the output is a planning suggestion model for management suggestions.

10. A system for generating management recommendations for long-lasting, low-maintenance mixed flower borders, used to apply the method for generating management recommendations for long-lasting, low-maintenance mixed flower borders as described in any one of claims 1-9, characterized in that, include: The data acquisition unit is configured to acquire plant growth data and community succession data based on the spatial distribution characteristics, plant growth morphology characteristics and environmental characteristics of the sample plot during the flower border growth cycle, and to construct the plant growth data and community succession data into a flower border data table. The landscape structure unit is configured to perform parameter adjustment on the flower border data table, extract feature data of plant morphology, growth rhythm and tolerance performance, determine plant feature vectors, classify all plant feature vectors based on the decision tree model, and determine the plant combination structure according to the split node features. The community discrimination unit is configured to determine the distance parameters between different plant combinations based on the K-nearest neighbor algorithm, determine the stability trend and substitution relationship of each plant combination based on the distance parameters, and determine the community discrimination result based on the judgment result and the plant combination structure. The management suggestion unit is configured to analyze the vertical hierarchy, coverage relationship and color coordination relationship of the plants based on the plant combination structure, determine the hierarchical flower border structure layout, determine the association result between the community discrimination result and the flower border structure layout based on the association rule algorithm, and determine management suggestions based on the association result and the planning suggestion model.

Images