A quantitative method for plant landscape matching based on green view rate

Abstract

This invention proposes a quantitative plant landscape matching method based on green view rate, relating to the field of environmental landscape design. The method includes: acquiring a three-dimensional model of the target underground space, dividing it into multiple functional subspaces, and calculating the sight distance at each viewpoint; collecting physiological data and subjective rating data, identifying abrupt change thresholds, and constructing a psychological response grading threshold model based on these thresholds; establishing a plant configuration database, and using a two-level nested approach to calculate the equivalent green view contribution coefficient of each plant at each viewpoint; constructing a multi-objective optimization model with the overall goal of maximizing the green view rate at each viewpoint; using plant species, planting locations, and planting densities from the plant configuration database as encoding objects, and employing a genetic algorithm to solve the multi-objective optimization model, outputting the plant species, planting locations, and planting densities for each functional subspace, thus obtaining the plant landscape configuration scheme. This invention achieves precise configuration based on the functional differences of underground spaces.

Description

Technical Field

[0001] This invention relates to the field of environmental landscape design technology, and in particular to a quantitative method for plant landscape matching based on green view rate. Background Technology

[0002] Due to their highly enclosed environment, lack of natural lighting, and limited ventilation, deep underground spaces can easily cause users to experience negative psychological reactions such as spatial oppression, disorientation, and mental fatigue when exposed to them for extended periods. Research indicates that plant landscaping has a significant effect on alleviating these negative psychological responses. Green view ratio, as an objective and quantifiable indicator measuring the proportion of green vegetation within a field of vision, has gradually become one of the core parameters that can be quantified and manipulated in indoor and outdoor landscape design, and is widely used in landscape evaluation for urban streets, public buildings, and other similar settings.

[0003] However, introducing green view rate into the field of underground space plant landscape configuration still faces significant technical obstacles at the design methodology level. Firstly, existing methods for quantifying green view rate typically express the relationship between green view rate and user psychological response as a continuous regression function. While this continuous mapping has mathematical completeness, it is difficult to directly translate into operational criteria in engineering design, and designers cannot derive executable quantitative boundaries for plant configuration from it. Secondly, existing estimation models for plant green view contribution usually only use the plant's own morphological parameters as input, neglecting the significant impact of observation distance on the plant's visual contribution. This results in insufficient accuracy in calculating the equivalent green view contribution when comparing and optimizing different planting locations.

[0004] Chinese invention patent application number 202111050252.1 discloses a greening construction method for pixelated plant landscapes. This method generates pixel modules based on design drawings or photographic patterns, then designs the pixel landscape by color schemes and plant selection. It corresponds the number of soilless cultivation planting cotton to the number of pixel modules, shrinking the modules to achieve a large-scale effect within a small space, and designing pixel landscape patterns that closely resemble the drawings or photographic patterns. Customers can also freely change the landscape effect, ensuring the landscape remains evergreen and fresh. However, it fails to fully consider the visual characteristics of underground spaces (such as limited visibility, fixed viewing angles, and clear movement paths), resulting in a mismatch between plant configuration and the user's actual visible range, leading to a waste of landscape resources. Summary of the Invention

[0005] In view of this, the present invention provides a quantitative plant landscape matching method based on green view rate to solve the technical problems in the prior art, such as the lack of engineering operable quantitative criteria for the mapping relationship between green view rate and user psychological response, insufficient evaluation accuracy due to the lack of observation distance correction in the calculation of plant green view contribution coefficient, and the lack of active perception and automatic response mechanism for dynamic deviation of green view rate.

[0006] The technical solution of this invention is implemented as follows: On the one hand, this invention provides a quantitative plant landscape matching method based on green view rate, including: S1. Obtain a three-dimensional model of the target underground space, divide it into multiple functional subspaces, set up viewpoints in each functional subspace, and calculate the line-of-sight distance of each viewpoint. S2. Collect physiological data and subjective rating data of users under different green view rate levels, identify the abrupt change critical point of psychological response with green view rate through segmented fitting analysis, and construct a psychological response grading threshold model based on the abrupt change critical point. S3. Establish a plant configuration database. Using the plant's leaf area index, viewing angle occlusion correction coefficient, and combined with the observation distance segmentation correction coefficient and maintenance attenuation coefficient corresponding to the viewing distance of each viewpoint, calculate the equivalent green view contribution coefficient of each plant at each viewpoint using a two-level nested approach. S4. Using low and high critical green view rates as constraints, and the sum of the equivalent green view contribution coefficients of each plant at each viewpoint as the objective function, a multi-objective optimization model is constructed with the goal of maximizing the comprehensive green view rate at each viewpoint. S5. For each functional subspace, using the plant species, planting locations and planting densities in the plant configuration database as encoding objects, a genetic algorithm is used to solve the multi-objective optimization model, outputting the plant species, planting locations and planting densities of each functional subspace, and obtaining the plant landscape configuration scheme.

[0007] Based on the above technical solutions, preferably, the functional subspaces include an entrance / exit transition area, a passageway area, a waiting area, a commercial service area, and a rest area. The sight distance within each functional subspace is the three-dimensional straight-line distance between the viewpoint and the corresponding plant planting point.

[0008] Based on the above technical solutions, preferably, step S2 specifically includes: Multiple discrete green visual rate gradient levels arranged from low to high are pre-set, and user physiological data and subjective rating data are collected simultaneously; the physiological data includes eye movement data, skin conductance response and heart rate variability, and the subjective rating data includes recovery perception scale score and spatial pressure score. The ratio of the single-step increase in the restorative perception scale score to the total increase over the entire process, and the ratio of the single-step decrease in the spatial oppression score to the total decrease over the entire process, are used as the criteria for judgment. In the green view rate gradient sequence, the first gradient node that is not lower than the preset judgment threshold for either the single-step increase ratio of the restorative perception scale score or the single-step decrease ratio of the spatial oppression score is identified. The first gradient node that meets the condition is recorded as the low critical green view rate, and the next gradient node that meets the same judgment condition is recorded as the high critical green view rate. The low critical green view rate is less than the high critical green view rate. The green view rate range is divided into three psychological benefit segments, with the low and high critical green view rates as boundaries. The segment with a green view rate below the low critical rate is marked with a value of [value missing]. The state marker value for the section between the low and high critical green view rates is [value missing]. The status marker value for sections with a green visibility rate higher than the high critical green visibility rate is [value missing]. This constitutes a psychological response grading threshold model.

[0009] Based on the above technical solutions, preferably, the plant configuration database includes plant species, morphological types, and leaf area index. Morphological correction coefficient View occlusion correction coefficient Maintenance attenuation coefficient and light adaptability threshold, among which leaf area index The sum of the total leaf area per unit projected area of ​​the plant; morphological correction factor. Based on plant morphology classification, the morphology types include hanging, clump-forming, and ground cover types, reflecting the differences in visual occlusion efficiency among these types; visual occlusion correction coefficient. The attenuation coefficient was determined by conducting visual occlusion experiments at a preset observation line-of-sight angle; The growth and degradation status of plants are observed and calibrated within a preset operating cycle; the light adaptability threshold is calibrated based on the plant's light tolerance level and is used to constrain light matching for plant selection during the optimization solution stage.

[0010] Based on the above technical solutions, the preferred calculation formula for the two-level nesting is: in, This represents the basic green vision index. This represents the equivalent green visual contribution coefficient. This represents the piecewise correction coefficient for the observation distance.

[0011] Based on the above technical solutions, preferably, step S4 specifically includes: The low critical green view rate and the high critical green view rate are used as the constraint boundaries for optimizing the green view rate of each functional subspace. The target green view rate interval of each viewpoint is set as a closed interval with the low critical green view rate as the lower limit and the high critical green view rate as the upper limit, which serves as the feasible region constraint. The cumulative value of the equivalent green visual contribution coefficient of each plant at all viewpoints in each functional subspace is used as the optimization objective function. The optimization is guided by maximizing the overall restorative perception benefit. At the same time, a multi-objective optimization model is constructed based on engineering constraints, including: the plant light adaptability threshold is matched with the illuminance of the functional subspace, the net height of the space after plant installation is not lower than the preset safe height, and the number of plant species in each functional subspace is not lower than the preset lower limit.

[0012] Based on the above technical solutions, preferably, step S5 specifically includes: Using plant species combinations, planting density arrays at each viewpoint, and spatial distribution parameters of plant planting points as encoding objects, the candidate plant configuration schemes of each functional subspace are encoded into population individuals of a genetic algorithm using real number encoding. A set of candidate schemes is generated by performing selection, crossover, and mutation operations on the population individuals. For each candidate scheme in the candidate scheme set, the cumulative green view rate of each viewpoint is calculated based on the sum of the equivalent green view contribution coefficients of each plant at each viewpoint. The condition that the cumulative green view rate of any viewpoint is lower than the low critical green view rate is used as the criterion for insufficient green view rate. Candidate schemes that meet the criterion are judged as not meeting the minimum green view benefit requirements specified by the psychological response grading threshold model and are removed from the candidate scheme set. Candidate schemes with all viewpoints having a cumulative green view rate that is not lower than the low critical green view rate are retained to obtain the screened candidate scheme set. Pareto front analysis was introduced into the candidate scheme set after screening. With the optimization guideline of maximizing the overall green view rate of each viewpoint, the plant species, planting points and planting density of each functional subspace were output to obtain the plant landscape configuration scheme.

[0013] Based on the above technical solutions, preferably, the quantitative plant landscape configuration method further includes: The real-time green view rate of each viewpoint is calculated using an image recognition algorithm according to a preset collection cycle; Continuous real-time green view rate at each viewpoint If the green visibility rate is below the low critical green visibility rate during a single acquisition cycle, a replanting trigger signal is activated for that viewpoint. Continuous real-time green view rate at each viewpoint If the number of acquisition cycles exceeds the high critical green view rate, it is a condition for determining that the green view rate is excessive, triggering a reduction recommendation signal for that viewpoint. The condition for achieving the green view rate target is that the real-time green view rate at each viewpoint remains within a closed interval with a lower critical green view rate as the lower limit and a higher critical green view rate as the upper limit, without triggering intervention; among which... This is the preset number of consecutive judgment cycles, which is a positive integer.

[0014] Based on the above technical solutions, preferably, the trigger signal for replanting specifically includes: Using the types, planting locations, and planting densities of the plants already installed in the current plant landscape configuration as fixed constraints, and the constraint of not removing any installed plants as a hard constraint, we only re-optimize the types, planting locations, and planting densities of newly added plants. The optimization objective is to trigger the cumulative green view rate of the viewpoint to return to a closed interval with a low critical green view rate as the lower limit and a high critical green view rate as the upper limit, thus obtaining the replanting scheme. The deviation between the real-time green view rate of each viewpoint and the theoretical cumulative value of the equivalent green view contribution coefficient of each plant at each viewpoint during the re-optimization process is recorded in the plant configuration database. The maintenance attenuation coefficient of the corresponding plant is iteratively updated based on the recorded deviations.

[0015] More preferably, the trigger signal for the reduction suggestion specifically includes: Among the installed plants, plants whose equivalent green visual contribution coefficient is lower than the preset contribution threshold are identified as priority plants to be reduced, and a set of priority plants to be reduced is obtained. Based on the principle of minimum removal quantity, determine the plants to be removed from the priority reduction target set and generate a reduction suggestion list; The cumulative green view rate of each viewpoint after removing the plants listed in the reduction recommendation list is predicted and verified. The condition for passing is that the predicted cumulative green view rate of all viewpoints is not lower than the low critical green view rate. After passing the verification, the reduction recommendation list is output.

[0016] The present invention has the following advantages over the prior art: (1) By spatial zoning and visual field modeling, psychological response grading threshold calibration, parametric plant configuration library construction, multi-objective optimization solution and dynamic closed-loop correction, the limitations of traditional design relying on subjective experience are overcome, the transformation of underground space plant landscape configuration from qualitative description to vectorized decision-making is realized, the quantitative and scientific nature of plant landscape design is realized, and the precise configuration of underground space functional differentiation is realized. (2) By segmenting and identifying the collected user physiological data and subjective rating data, a graded threshold model with low critical green view rate and high critical green view rate as the boundary was established. The relationship between green view rate and psychological response was transformed from a continuous regression curve into a graded criterion with clear interval division. Based on the two-level nesting, the actual contribution of the same plant to the viewpoint green view rate at different observation distances can be expressed differently, thereby improving the calculation accuracy of plant planting point optimization. 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 description of the embodiments or the prior art 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 This is a flowchart of a quantitative plant landscape matching method based on green view rate according to the present invention; Figure 2 This is a schematic diagram of the functional subspace division and viewpoint layout of a quantitative plant landscape matching method based on green visibility rate according to the present invention. Figure 3 This is a schematic diagram of a two-level nested calculation structure for a quantitative plant landscape matching method based on green view rate according to the present invention. Figure 4 This is a flowchart illustrating the dynamic monitoring and closed-loop correction of a quantitative plant landscape matching method based on green visibility rate according to the present invention. Detailed Implementation

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

[0020] like Figure 1 and Figure 2 As shown, this invention provides a quantitative plant landscape matching method based on green view rate, including: S1. Obtain a three-dimensional model of the target underground space, divide it into multiple functional subspaces, set up viewpoints in each functional subspace, and calculate the line-of-sight distance of each viewpoint. The functional subspaces include entrance / exit transition areas, passage corridor areas, waiting areas, commercial service areas, and rest areas. The line-of-sight distance in each functional subspace is the three-dimensional straight-line distance between the viewpoint position and the corresponding plant planting point.

[0021] In one embodiment of the present invention, the underground space is taken as the concourse level of an underground rail transit station. A three-dimensional model of the target underground space is constructed using building information modeling (BIM) software. The three-dimensional model includes the geometric information, material information, and spatial dimension information of each structural component. Within each functional subspace, viewpoints are evenly distributed along the main flow lines and resting positions of the users, based on the standing eye level of the user in normal use. The horizontal spacing of the viewpoints is set according to the spatial scale of each functional subspace to ensure that each viewpoint is representative of the field of vision of that subspace.

[0022] S2. Collect physiological data and subjective rating data of users under different green view rate levels, identify the abrupt change critical point of psychological response with green view rate through segmented fitting analysis, and construct a psychological response grading threshold model based on the abrupt change critical point. Specifically, step S2 includes: Multiple discrete green visual rate gradient levels arranged from low to high are pre-set, and user physiological data and subjective rating data are collected simultaneously; the physiological data includes eye movement data, skin conductance response and heart rate variability, and the subjective rating data includes recovery perception scale score and spatial pressure score. Using the ratio of the single-step increase in the restorative perception scale score to the total increase over the entire green view rate gradient, and the ratio of the single-step decrease in the spatial oppression score to the total decrease over the entire green view rate gradient, as the criteria, the first gradient node in the green view rate gradient sequence that is not lower than a preset threshold for either the single-step increase ratio of the restorative perception scale score or the single-step decrease ratio of the spatial oppression score is identified. The first gradient node that meets the condition is recorded as the low critical green view rate. The next gradient node that satisfies the same criterion is denoted as the high critical green visibility rate. Among them, low critical green vision rate Less than the high critical green view rate ; The green view rate range is divided into three psychological benefit segments, with the low and high critical green view rates as boundaries. The segment with a green view rate below the low critical rate is marked with a value of [value missing]. The state marker value for the section between the low and high critical green view rates is [value missing]. The status marker value for sections with a green visibility rate higher than the high critical green visibility rate is [value missing]. This constitutes a psychological response grading threshold model.

[0023] Understandable, Green View Rate The green pixel area of ​​plants accounts for the proportion of the total field of view within a 360° (or specified viewing angle) field of view of a human eye at a given viewpoint. The cumulative green view rate at a viewpoint is determined by the equivalent green view contribution coefficient of each plant within the visible range of that viewpoint. The approximation obtained by linear accumulation is as follows: in This represents the total number of plants within the visible range of this viewpoint. For the first The equivalent green visual contribution coefficient of a plant to this viewpoint.

[0024] The Perceived Restorativeness Scale (PRS) is a standardized scale used to assess the psychological recovery effects of the environment on users, such as attention recovery and stress relief. The PRS scale used in this invention includes several evaluation dimensions (such as sense of distance, sense of abundance, sense of coherence, and sense of fit to purpose). After viewing landscape images at various green visibility levels, subjects rate their environmental perception on a Likert scale of 1 to 7 for each dimension. The average score of each dimension is then used as the comprehensive PRS score corresponding to that green visibility level, with a range of 1 to 7.

[0025] The spatial oppression rating uses a 5-point Likert scale (1 point represents "no oppression at all", and 5 points represent "extremely strong oppression"). The higher the score, the stronger the user's perception of spatial oppression. As the green visibility level increases, the spatial oppression rating generally shows a downward trend. "Single-step decrease" refers to the amount of reduction in the spatial oppression rating within adjacent green visibility gradient intervals (i.e., the absolute value of the score at the higher green visibility level minus the score at the lower green visibility level); "total decrease over the entire range" refers to the total decrease in the spatial oppression rating from the lowest green visibility gradient to the highest green visibility gradient.

[0026] By introducing a critical point identification mechanism for abrupt changes in psychological response, the continuous mapping relationship between green visibility rate and psychological response is transformed into a three-tiered criterion with clear interval boundaries. This allows the setting of optimization constraint boundaries to directly correspond to the critical changes in users' psychological benefits, making it engineering-operable.

[0027] S3. Establish a plant configuration database. Using the plant's leaf area index, viewing angle occlusion correction coefficient, and combined with the observation distance segmentation correction coefficient and maintenance attenuation coefficient corresponding to the line-of-sight distance at each viewpoint, calculate the equivalent green view contribution coefficient of each plant at each viewpoint using a two-level nested approach. The plant configuration database includes plant species, morphological types, and leaf area index. Morphological correction coefficient View occlusion correction coefficient Maintenance attenuation coefficient and light adaptability threshold, leaf area index The sum of the total leaf area per unit projected area of ​​the plant; morphological correction factor. Based on plant morphology classification, the morphology types include hanging, clump-forming, and ground cover types, reflecting the differences in visual occlusion efficiency among these types; visual occlusion correction coefficient. Visual occlusion experiments were conducted at a preset viewing angle to characterize the proportion of green area obstructed by plants at that angle; maintenance attenuation coefficient was also measured. By observing and calibrating the growth and degradation status of plants within a preset operating cycle, the green vision contribution loss rate caused by factors such as growth degradation and leaf drop is characterized; the light adaptability threshold is calibrated based on the plant's light tolerance level and is used to constrain light matching for plant selection during the optimization solution stage.

[0028] Understandable, shape correction factor Based on the visual shading efficiency of ground cover plants ( The difference in shading efficiency between trailing and clump-forming plants and ground cover plants was corrected. Trailing plants, due to their drooping branches and spreading leaves, have a larger vertical shading area on the viewer's line of sight. Clump-forming plants have a visual obstruction efficiency close to the baseline. Slightly greater than or equal to 1; ground cover plants, due to their low stature and leaves mainly distributed below eye level, have relatively low shading efficiency in the direction of human eye level, therefore The specific calibration is obtained by calculating the mean value of shading experiments on representative plants of different morphological types under the same observation conditions, and the value can be taken in the range of [0.5, 1.5].

[0029] In one embodiment of the present invention, the view occlusion correction coefficient The determination method is as follows: The plant to be tested is placed at a standard shooting distance. Using the human eye's standing eye level as a reference, images of the plant are taken at preset viewing angles (such as frontal eye level, 15° downward view, 30° side view, etc.). The proportion of green pixel area of ​​the plant to the total image area at that viewing angle is calculated using an image green pixel extraction algorithm. This proportion is the viewing angle occlusion correction coefficient at that viewing angle. .

[0030] The closer to 1, the higher the visual occlusion efficiency of the plant at that line of sight (the larger the proportion of green canopy area in the field of vision). The smaller the value, the higher the transparency of the plant canopy or the lower the effective green view contribution from that perspective.

[0031] like Figure 3 As shown, in one embodiment of the present invention, the two-level nested calculation formula is as follows: in, This represents the basic green vision index. This represents the equivalent green visual contribution coefficient. This represents the piecewise correction coefficient for the observation distance.

[0032] Understandable, observation distance correction factor Based on the line of sight distance between the viewpoint and the planting point Segmented values ​​are taken at preset distances from nodes. , ( () serves as the boundary, and the line of sight distance does not exceed hour Take the larger value, with the line-of-sight distance between and Between Take the secondary value; the line-of-sight distance exceeds... hour The minimum value is taken, and the values ​​of the three segments satisfy a monotonically decreasing relationship. The coefficient values ​​of each segment are calibrated based on the measured data of the visual proportion of green area at different viewing distances in the visual occlusion experiment of the subjects. Equivalent green visual contribution coefficient Characterizing the actual green visual contribution of a plant at a specific observation distance and maintenance condition, the green visual contribution of the same plant at positions closer to the viewpoint. The observation distance is significantly higher than that at more distant locations, allowing the optimal selection of plant planting sites to reflect the actual impact of observation distance on the greening effect.

[0033] visual distance The calculation system explicitly incorporates the equivalent green visual contribution coefficient. Through a two-level nested structure, it distinguishes the influence of the plant's inherent contribution capacity and actual observation conditions on the green visual effect, enabling the contribution coefficient of the same plant at different planting sites to be expressed differently, thereby improving the calculation accuracy of planting site optimization.

[0034] set up This is the angular occlusion correction coefficient for plants in the observation direction. It represents the amount of occlusion efficiency correction actually occupied by a single plant in the human eye's field of view, reflecting the coupled influence of plant canopy morphology, density, and observation direction. hour, , The growth rate is constrained; when When smaller, An increase indicates that the contribution of plant leaf volume in that viewing direction is significantly reduced. Regarding Plants under artificial lighting conditions will experience leaf aging, leaf area reduction, and crown shrinkage over time, resulting in a continuous decline in their actual green view contribution compared to their initial planting state. A combination of measured deviations and database iterations is used to ensure that the coefficients are updated synchronously with the actual conditions. This invention establishes a graded threshold model with low and high critical green visibility rates as boundaries by segmenting and identifying collected user physiological data and subjective rating data. It transforms the relationship between green visibility rate and psychological response from a continuous regression curve into a graded criterion with clear interval division. Based on two-level nesting, the actual contribution of the same plant to the green visibility rate at different viewing distances can be expressed differently, thereby improving the calculation accuracy of plant planting site optimization.

[0035] In one embodiment of the present invention, after the two-level nested calculation is completed, the viewpoint-planting point position of each group is corrected: Based on the 3D model of the target underground space, a 3D line path is established between the viewpoint and the planting point. The presence of any building structures or canopy projections of existing plants along this path is detected. The transparency ratio of the sight path is calculated and recorded as the sight transparency coefficient. Its value range is When there are no obstructions in the path of sight When the line of sight is completely blocked Approaching 0 and excluding this combination from the contribution accumulation; Effective values ​​of equivalent green visual contribution coefficients at each viewpoint Revised to: in This is the equivalent green view contribution coefficient calculated from the original two-level nested formula. This is the visual transparency coefficient between the viewpoint and the corresponding plant planting location; The corrected cumulative green view rate at each viewpoint is determined by the effective contribution coefficient of each plant. The accumulated value is calculated.

[0036] in The detection is based on a 3D model of the target underground space, using ray projection calculations. For each candidate planting point and each viewpoint combination, this is performed synchronously during the fitness calculation phase of the optimization solution process. Relying on the canopy projection of the configured plants, its calculation needs to be dynamically updated during the fitness evaluation of individuals in each generation of the genetic algorithm population. That is, different configuration schemes (different combinations of plant species and planting sites) correspond to different... The distribution allows the optimization process to automatically avoid the loss of actual green view contribution caused by mutual shading of plant communities.

[0037] S4. Using low and high critical green view rates as constraints, and the sum of the equivalent green view contribution coefficients of each plant at each viewpoint as the objective function, a multi-objective optimization model is constructed with the goal of maximizing the comprehensive green view rate at each viewpoint. Specifically, step S4 includes: The low critical green view rate and the high critical green view rate are used as the constraint boundaries for optimizing the green view rate of each functional subspace. The target green view rate interval of each viewpoint is set as a closed interval with the low critical green view rate as the lower limit and the high critical green view rate as the upper limit, which serves as the feasible region constraint. The cumulative value of the equivalent green visual contribution coefficient of each plant at all viewpoints in each functional subspace is used as the optimization objective function. The optimization is guided by maximizing the overall restorative perception benefit. At the same time, a multi-objective optimization model is constructed based on engineering constraints, including: the plant light adaptability threshold is matched with the illuminance of the functional subspace, the net height of the space after plant installation is not lower than the preset safe height, and the number of plant species in each functional subspace is not lower than the preset lower limit.

[0038] Understandably, the light adaptability threshold of plants should not be higher than the measured illuminance of the functional subspace to ensure normal growth of plants under artificial lighting conditions; the net height of the space after plant installation should not be lower than the preset safety height to meet passage safety requirements; and the number of plant species in each functional subspace should not be lower than the preset lower limit to ensure landscape diversity.

[0039] Restorative perception benefit refers to the psychological restorative effect produced by users visually perceiving green plants in underground spaces. Its strength is positively correlated with the visual proportion of green plants in the field of vision (i.e., green view rate). Therefore, taking the maximization of the cumulative value of the equivalent green view contribution coefficient of each viewpoint (i.e., cumulative green view rate) as the optimization objective function, under the constraints specified by the psychological response grading threshold model, is equivalent to taking the maximization of the overall restorative perception benefit as the optimization guide.

[0040] Let's take the rest area as an example for a specific explanation: Target green visibility range The optimization model is set as follows: The objective function is to maximize the restorative perception benefit, and the constraints include: green view rate constraint. That is, the cumulative green view rate of each viewpoint must fall within or Sections, and must not contain The viewpoint is defined by the following constraints: Lighting constraint: the selected plants' light adaptability level must be no lower than the actual light conditions in the area (LED supplemental lighting in this area provides approximately 200 lux, requiring plants to be shade-tolerant); Maintenance constraint: no more than 3 plant species; Spatial clearance constraint: hanging plants must be suspended at least 2.2 meters high (the clearance in this area is 4.2 meters, which meets the requirement); Distance segmentation constraint: the distance between each candidate planting point and the viewpoint during optimization. It must be clearly classified into one of the three distance segments to ensure During calculation The uniqueness of the value.

[0041] A genetic algorithm is used to solve the problem, with a population size of 100, 200 iterations, a crossover probability of 0.8, a mutation probability of 0.05, and an elite retention rate of 10%. After introducing Pareto front analysis, a set of candidate solutions is output.

[0042] The optimized plant configuration is as follows: A three-tiered arrangement of trailing pothos (hanging above eye level, 2.5 meters from the ground), clump-forming schefflera (placed in flower boxes 0.8 meters above eye level), and ground cover peace lilies (placed in flower boxes below eye level); planting density is 3 pothos plants per linear meter, 2 schefflera plants per square meter, and 4 peace lily clumps per square meter; recommended viewing distance is 0.8 to 2.5 meters; the cumulative green view rate at each typical viewpoint is calculated to be 23.6%, falling within the target range. Within, corresponding status flags The plan passed the feasibility test.

[0043] In one embodiment of the present invention, based on feasible domain constraints and engineering constraints, a green visual clustering constraint is introduced to ensure that the green visual information perceived by a user continuously passing through various viewpoints along the same functional subspace has spatial coherence: Two adjacent viewpoints within the same functional subspace on the user's main movement path and Defined as adjacent viewpoint pairs, calculate the cumulative green view rate of each adjacent viewpoint pair. and The absolute value of the difference is recorded as the green visibility fluctuation between adjacent viewpoints. ,Require Not exceeding the preset fluctuation threshold ,Right now: in The maximum allowable fluctuation in green visibility between adjacent viewpoints is set according to the type of functional subspace: for passageways with faster movement speeds, where users stay for short periods and green visibility perception is primarily achieved through rapid scanning. The value can be appropriately relaxed, but it is recommended that it not exceed 50% of the target green visibility range width of the functional subspace; for waiting areas and rest areas with slow movement speeds or lingering behaviors, The value should be set strictly, and it is recommended that it not exceed 30% of the width of the target green visibility range, so as to ensure that users perceive a relatively uniform green environment during longer periods of time and slow eye movement.

[0044] S5. For each functional subspace, using the plant species, planting locations and planting densities in the plant configuration database as encoding objects, a genetic algorithm is used to solve the multi-objective optimization model, outputting the plant species, planting locations and planting densities of each functional subspace, and obtaining the plant landscape configuration scheme.

[0045] Specifically, step S5 includes: Using plant species combinations, planting density arrays at each viewpoint, and spatial distribution parameters of plant planting points as encoding objects, the candidate plant configuration schemes of each functional subspace are encoded into population individuals of a genetic algorithm using real number encoding. A set of candidate schemes is generated by performing selection, crossover, and mutation operations on the population individuals. For each candidate scheme in the candidate scheme set, the cumulative green view rate of each viewpoint is calculated based on the sum of the equivalent green view contribution coefficients of each plant at each viewpoint. The condition that the cumulative green view rate of any viewpoint is lower than the low critical green view rate is used as the criterion for insufficient green view rate. Candidate schemes that meet the criterion are judged as not meeting the minimum green view benefit requirements specified by the psychological response grading threshold model and are removed from the candidate scheme set. Candidate schemes with all viewpoints having a cumulative green view rate that is not lower than the low critical green view rate are retained to obtain the screened candidate scheme set. Pareto front analysis was introduced into the candidate scheme set after screening. With the optimization guideline of maximizing the overall green view rate of each viewpoint, the plant species, planting points and planting density of each functional subspace were output to obtain the plant landscape configuration scheme.

[0046] Understandably, crossover operations reorganize species combinations and spatial distribution parameters, while mutation operations introduce random perturbations in planting density and location coordinates to maintain population diversity.

[0047] A multi-objective optimization model contains two mutually constraining optimization objectives: Objective 1: Maximize the overall green view rate of all viewpoints (i.e., all viewpoints within each functional subspace). (Maximum sum of accumulated values); Objective 2: Minimize the overall maintenance cost of plant configuration (using the sum of the maintenance requirement coefficients of each selected plant as a proxy indicator).

[0048] This invention overcomes the limitations of traditional design relying on subjective experience by using spatial zoning and visual field modeling, psychological response grading threshold calibration, parametric plant configuration library construction, multi-objective optimization solution and dynamic closed-loop correction. It realizes the transformation of underground space plant landscape configuration from qualitative description to vectorized decision-making, realizes the quantification and scientification of plant landscape design, and achieves precise configuration of underground space functions.

[0049] like Figure 4 As shown, in one embodiment of the present invention, the quantitative plant landscape configuration method further includes: The real-time green view rate of each viewpoint is calculated using an image recognition algorithm according to a preset collection cycle; Continuous real-time green view rate at each viewpoint If the green visibility rate is below the low critical green visibility rate during a single acquisition cycle, a replanting trigger signal is activated for that viewpoint. Continuous real-time green view rate at each viewpoint If the number of acquisition cycles exceeds the high critical green view rate, it is a condition for determining that the green view rate is excessive, triggering a reduction recommendation signal for that viewpoint. The condition for achieving the green view rate target is that the real-time green view rate at each viewpoint remains within a closed interval with a lower critical green view rate as the lower limit and a higher critical green view rate as the upper limit, without triggering intervention; among which... This is the preset number of consecutive judgment cycles, which is a positive integer.

[0050] Furthermore, the specific signals that trigger replanting include: Using the types, planting locations, and planting densities of the plants already installed in the current plant landscape configuration as fixed constraints, and the constraint of not removing any installed plants as a hard constraint, we only re-optimize the types, planting locations, and planting densities of newly added plants. The optimization objective is to trigger the cumulative green view rate of the viewpoint to return to a closed interval with a low critical green view rate as the lower limit and a high critical green view rate as the upper limit, thus obtaining the replanting scheme. The deviation between the real-time green view rate of each viewpoint and the theoretical cumulative value of the equivalent green view contribution coefficient of each plant at each viewpoint after the implementation of the replanting plan will be recorded in the plant configuration database. The maintenance attenuation coefficient of the corresponding plant is iteratively updated based on the recorded deviations.

[0051] Furthermore, the specific signals that trigger the feature reduction recommendation include: Among the installed plants, plants whose equivalent green visual contribution coefficient is lower than the preset contribution threshold are identified as priority plants to be reduced, and a set of priority plants to be reduced is obtained. Based on the principle of minimum removal quantity, determine the plants to be removed from the priority reduction target set and generate a reduction suggestion list; The cumulative green view rate of each viewpoint after removing the plants listed in the reduction recommendation list is predicted and verified. The condition for passing is that the predicted cumulative green view rate of all viewpoints is not lower than the low critical green view rate. After passing the verification, the reduction recommendation list is output.

[0052] The principle of minimum removal quantity means that when determining the plants to be removed, priority is given to the option that minimizes the number of plants listed in the reduction recommendation list. That is, under the premise that the cumulative green view rate of all viewpoints after removal is not lower than the low critical green view rate, plants are added to the list of plants to be removed one by one from the priority reduction target set in order of contribution coefficient from low to high. Each time a plant is added, a prediction verification is performed. When the verification is successful, the addition is stopped. This ensures that the number of plants removed is minimized and avoids excessive reduction that would cause the green view rate to fall below the low critical value.

[0053] This invention periodically monitors the real-time green visibility rate at each viewpoint, automatically determines the green visibility rate status based on the low and high critical green visibility rates in the psychological response grading threshold model, and triggers a replanting and re-optimization process or outputs a reduction suggestion accordingly. At the same time, the deviation between the measured green visibility rate and the theoretical calculated value during the re-optimization process is fed back to the plant configuration database, and the maintenance attenuation coefficient of the corresponding plants is iteratively updated, realizing the continuous self-correction of the configuration parameter library as operational data accumulates.

[0054] The technical solution of the present invention will be specifically described using a specific embodiment: Obtain a 3D model of the subway station concourse (clear height 4.2m, area approximately 3200㎡), and divide the concourse into the following functional subspaces: entrance / exit transition area (4 locations, each approximately 80㎡), passageway area (main passageway 6m wide, total length approximately 180m), waiting / waiting area (waiting area above the platform, area approximately 600㎡), and rest area (rest area with seating, 3 locations, each approximately 50㎡). Set viewpoints within each functional subspace: Entrance / Exit Transition Zone: A static viewpoint is set 3m inside each entrance / exit, with the line of sight pointing towards the interior of the station hall; Passage corridor area: A dynamic viewpoint is set every 10m along the center line of the passage, with the line of sight along the axis of the passage and the walls on both sides; Waiting / waiting area: Viewpoints are evenly distributed in the waiting area at 5m intervals, with a 360° line of sight. Rest and rest area: A viewpoint is set at each seat position, with the line of sight at eye level when seated (1.2m above the ground).

[0055] The visual field analysis algorithm is used to calculate the visual field information of each viewpoint, including the visible area in each direction, the line of sight distance, and the distribution of wall and column occlusion.

[0056] The experiment established a quantitative mapping relationship between green visual rate and psychological response indicators. Sixty subjects (aged 25-55, half male and half female) were recruited and viewed landscape images at different green visual rate levels (5%, 10%, 15%, 20%, 25%, 30%, and 35%) in a simulated underground space environment in the laboratory. Each level was viewed for 5 minutes. Eye movement data (fixation duration, fixation frequency, and saccade path), physiological data (dermal conductance response (EDA), heart rate variability (HRV)), and subjective scores (restorative perception scale (PRS), spatial oppression score (5 levels), and orientation recognition test) were collected. A multivariate regression analysis was used to establish a relationship model between green view rate (GR) and restorative perception score (R), and a relationship between green view rate and spatial oppression index (P) was also established. Plant species suitable for artificial lighting conditions (LED supplemental lighting, 200 lux) in underground spaces were selected to establish a plant configuration parameter library. The parameters are shown in Table 1. Table 1 Plant Configuration Parameter Database The target green view rate ranges for each functional subspace are set as follows: rest and rest area: 25%-30%, waiting / bus waiting area: 18%-25%, passageway area: 10%-15%, entrance and exit transition area: 8%-12%; Taking the rest and relaxation area as an example, the optimization model is set as follows: Objective function: Maximize restorative perception score Constraints: Green space ratio constraints: 25%≤GR≤30% Light constraints: The light adaptability level of the selected plants shall not be lower than the actual light conditions of the area (this area is supplemented by LED lighting, with an illuminance of about 200 lux). Maintenance constraints: The number of plant species is subject to both a lower limit constraint (to ensure landscape diversity) and an upper limit constraint (to control maintenance costs), and the total maintenance demand coefficient does not exceed the threshold. Space clearance constraints: The hanging height of trailing plants shall not be less than 2.2m; A genetic algorithm is used to solve the problem, with a population size of 100 and 200 iterations, and the Pareto front candidate solutions are output.

[0057] The optimization result is as follows: Plant arrangement: trailing pothos (hanging above the line of sight, 2.5m from the ground) + clump-forming schefflera (placed in the line of sight area, in a flower box 1.2m above the ground) + ground cover peace lily (placed below the line of sight, in a ground flower box). Planting density: 3 pots of pothos per linear meter, 2 plants of schefflera per square meter, and 4 clumps of peace lily per square meter; Recommended viewing distance: 0.8m-2.5m; Overall green view rate: 27.3%; The distance segmentation node is set as rice, The distance correction factor calibration value for the three sections is meters. , , .

[0058] The optimization results are mapped onto the 3D model of the station hall to generate: a visual layout map: marking the plant species, planting points, and planting density of each area; a construction parameter table: including plant specifications, flower box dimensions, hanging device locations, irrigation points, and supplemental lighting locations; and a maintenance manual: including watering cycles, pruning requirements, and supplemental lighting operating parameters.

[0059] Six months after the landscape wall was put into use, actual green view rate data (through regular photography and image recognition) and user satisfaction questionnaires were collected (once per quarter, with no fewer than 200 valid questionnaires collected each time). Comparing the data with the mapping relationship in step S2, it was found that the actual user satisfaction rating for the rest area was about 8% lower than the predicted value.

[0060] Analysis revealed the cause to be that the selected plants were growing slower than expected, with an actual green visibility rate of only 23.5%. Based on this, the plant configuration parameter database was revised, increasing the growth rate coefficient of the pothos plant and recommending a 10% increase in supplemental lighting during subsequent maintenance to bring the actual green visibility rate back to the target range.

[0061] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A quantitative plant landscape matching method based on green view rate, characterized in that, include: S1. Obtain a three-dimensional model of the target underground space, divide it into multiple functional subspaces, set up viewpoints in each functional subspace, and calculate the line-of-sight distance of each viewpoint. S2. Collect physiological data and subjective rating data of users under different green view rate levels, identify the abrupt change critical point of psychological response with green view rate through segmented fitting analysis, and construct a psychological response grading threshold model based on the abrupt change critical point. S3. Establish a plant configuration database. Using the plant's leaf area index, viewing angle occlusion correction coefficient, and combined with the observation distance segmentation correction coefficient and maintenance attenuation coefficient corresponding to the viewing distance of each viewpoint, calculate the equivalent green view contribution coefficient of each plant at each viewpoint using a two-level nested approach. S4. Using low and high critical green view rates as constraints, and the sum of the equivalent green view contribution coefficients of each plant at each viewpoint as the objective function, a multi-objective optimization model is constructed with the goal of maximizing the comprehensive green view rate at each viewpoint. S5. For each functional subspace, using the plant species, planting locations and planting densities in the plant configuration database as encoding objects, a genetic algorithm is used to solve the multi-objective optimization model, outputting the plant species, planting locations and planting densities of each functional subspace, and obtaining the plant landscape configuration scheme.

2. The quantitative plant landscape matching method based on green view rate as described in claim 1, characterized in that: The functional subspaces include entrance / exit transition areas, passageway areas, waiting areas, commercial service areas, and rest areas. The sight distance within each functional subspace is the three-dimensional straight-line distance between the viewpoint and the corresponding plant planting point.

3. The quantitative plant landscape matching method based on green view rate as described in claim 1, characterized in that: Step S2 specifically includes: Multiple discrete green visual rate gradient levels arranged from low to high are pre-set, and user physiological data and subjective rating data are collected simultaneously; the physiological data includes eye movement data, skin conductance response and heart rate variability, and the subjective rating data includes recovery perception scale score and spatial pressure score. The ratio of the single-step increase in the restorative perception scale score to the total increase over the entire process, and the ratio of the single-step decrease in the spatial oppression score to the total decrease over the entire process, are used as the criteria for judgment. In the green view rate gradient sequence, the first gradient node that is not lower than the preset judgment threshold for either the single-step increase ratio of the restorative perception scale score or the single-step decrease ratio of the spatial oppression score is identified. The first gradient node that meets the condition is recorded as the low critical green view rate, and the next gradient node that meets the same judgment condition is recorded as the high critical green view rate. The low critical green view rate is less than the high critical green view rate. The green view rate range is divided into three psychological benefit segments, with the low and high critical green view rates as boundaries. The segment with a green view rate below the low critical rate is marked with a value of [value missing]. The state marker value for the section between the low and high critical green view rates is [value missing]. The status marker value for sections with a green visibility rate higher than the high critical green visibility rate is [value missing]. This constitutes a psychological response grading threshold model.

4. The quantitative plant landscape matching method based on green view rate as described in claim 1, characterized in that: The plant configuration database includes plant species, morphological types, and leaf area index. Morphological correction coefficient View occlusion correction coefficient Maintenance attenuation coefficient and light adaptability threshold, among which leaf area index The sum of the total leaf area per unit projected area of ​​the plant; morphological correction factor. Based on plant morphology classification, the morphology types include hanging, clump-forming, and ground cover types, reflecting the differences in visual occlusion efficiency among these types; visual occlusion correction coefficient. The attenuation coefficient was determined by conducting visual occlusion experiments at a preset observation line-of-sight angle; The growth and degradation status of plants are observed and calibrated within a preset operating cycle; the light adaptability threshold is calibrated based on the plant's light tolerance level and is used to constrain light matching for plant selection during the optimization solution stage.

5. The quantitative plant landscape matching method based on green view rate as described in claim 4, characterized in that: The calculation formula for the two-level nesting is as follows: ； ； in, This represents the basic green vision index. This represents the equivalent green visual contribution coefficient. This represents the piecewise correction coefficient for the observation distance.

6. The quantitative plant landscape matching method based on green view rate as described in claim 4, characterized in that: Step S4 specifically includes: The low critical green view rate and the high critical green view rate are used as the constraint boundaries for optimizing the green view rate of each functional subspace. The target green view rate interval of each viewpoint is set as a closed interval with the low critical green view rate as the lower limit and the high critical green view rate as the upper limit, which serves as the feasible region constraint. The cumulative value of the equivalent green visual contribution coefficient of each plant at all viewpoints in each functional subspace is used as the optimization objective function. The optimization is guided by maximizing the overall restorative perception benefit. At the same time, a multi-objective optimization model is constructed based on engineering constraints, including: the plant light adaptability threshold is matched with the illuminance of the functional subspace, the net height of the space after plant installation is not lower than the preset safe height, and the number of plant species in each functional subspace is not lower than the preset lower limit.

7. The quantitative plant landscape matching method based on green view rate as described in claim 1, characterized in that: Step S5 specifically includes: Using plant species combinations, planting density arrays at each viewpoint, and spatial distribution parameters of plant planting points as encoding objects, the candidate plant configuration schemes of each functional subspace are encoded into population individuals of a genetic algorithm using real number encoding. A set of candidate schemes is generated by performing selection, crossover, and mutation operations on the population individuals. For each candidate scheme in the candidate scheme set, the cumulative green view rate of each viewpoint is calculated based on the sum of the equivalent green view contribution coefficients of each plant at each viewpoint. The condition that the cumulative green view rate of any viewpoint is lower than the low critical green view rate is used as the criterion for insufficient green view rate. Candidate schemes that meet the criterion are judged as not meeting the minimum green view benefit requirements specified by the psychological response grading threshold model and are removed from the candidate scheme set. Candidate schemes with all viewpoints having a cumulative green view rate that is not lower than the low critical green view rate are retained to obtain the screened candidate scheme set. Pareto front analysis was introduced into the candidate scheme set after screening. With the optimization guideline of maximizing the overall green view rate of each viewpoint, the plant species, planting points and planting density of each functional subspace were output to obtain the plant landscape configuration scheme.

8. The quantitative plant landscape matching method based on green view rate as described in claim 5, characterized in that: The quantitative plant landscape configuration method also includes: The real-time green view rate of each viewpoint is calculated using an image recognition algorithm according to a preset collection cycle; Continuous real-time green view rate at each viewpoint If the green visibility rate is below the low critical green visibility rate during a single acquisition cycle, a replanting trigger signal is activated for that viewpoint. Continuous real-time green view rate at each viewpoint If the number of acquisition cycles exceeds the high critical green view rate, it is a condition for determining that the green view rate is excessive, triggering a reduction recommendation signal for that viewpoint. The condition for achieving the green view rate target is that the real-time green view rate at each viewpoint remains within a closed interval with a lower critical green view rate as the lower limit and a higher critical green view rate as the upper limit, without triggering intervention; among which... This is the preset number of consecutive judgment cycles, which is a positive integer.

9. A quantitative plant landscape matching method based on green view rate as described in claim 8, characterized in that: The trigger signal for replanting specifically includes: Using the types, planting locations, and planting densities of the plants already installed in the current plant landscape configuration as fixed constraints, and the constraint of not removing any installed plants as a hard constraint, we only re-optimize the types, planting locations, and planting densities of newly added plants. The optimization objective is to trigger the cumulative green view rate of the viewpoint to return to a closed interval with a low critical green view rate as the lower limit and a high critical green view rate as the upper limit, thus obtaining the replanting scheme. The deviation between the real-time green view rate of each viewpoint and the theoretical cumulative value of the equivalent green view contribution coefficient of each plant at each viewpoint after the implementation of the replanting plan will be recorded in the plant configuration database. The maintenance attenuation coefficient of the corresponding plant is iteratively updated based on the recorded deviations.

10. A quantitative plant landscape matching method based on green view rate as described in claim 9, characterized in that: The trigger signal for the configuration reduction suggestion specifically includes: Among the installed plants, plants whose equivalent green visual contribution coefficient is lower than the preset contribution threshold are identified as priority plants to be reduced, and a set of priority plants to be reduced is obtained. Identify plants to be removed from the priority reduction target set and generate a reduction recommendation list; The cumulative green view rate of each viewpoint after removing the plants listed in the reduction recommendation list is predicted and verified. The condition for passing is that the predicted cumulative green view rate of all viewpoints is not lower than the low critical green view rate. After passing the verification, the reduction recommendation list is output.

Images