Urban park quality evaluation method based on landscape, behavior and emotion collaborative differentiation

Abstract

The invention specifically discloses an urban park quality evaluation method based on landscape, behavior and emotion collaborative differentiation, and relates to the technical field of urban planning. According to the method, firstly, landscape data, behavior data and emotion perception data are collected and preprocessed; secondly, constructing an urban park quality evaluation index system based on three dimensions of park landscape quality, park access behaviors and park emotion perception; then calculating a comprehensive experience index, and quantifying a synergistic effect of park landscape quality, park access behaviors and park emotion perception; performing differentiation analysis on park landscape quality, park access behaviors and park emotion perception; and finally, generating a park quality space-time evaluation result and a differentiated quality improvement strategy. According to the method, comprehensive and accurate evaluation of the quality of the urban park is realized, a scientific basis is provided for differentiated quality improvement, and efficient utilization of park resources and maximization of service benefits are facilitated.

Description

Technical Field

[0001] This invention relates to the field of urban planning technology, and in particular to a method for evaluating the quality of urban parks based on the synergistic differentiation of landscape, behavior and emotion. Background Technology

[0002] As core public spaces in cities, urban parks play a crucial role in improving residents' quality of life, maintaining ecological balance, and promoting social interaction. With the acceleration of urbanization, the public's demand for high-quality parks is growing, making the scientific assessment of park quality a core prerequisite for optimizing urban green space planning and management.

[0003] Existing park quality assessment methods have significant limitations: On the one hand, traditional studies often focus on the physical attributes of parks, such as green coverage and the number of facilities, neglecting the public's actual user experience and emotional feedback. This leads to a disconnect between assessment results and actual service effectiveness, making it difficult to match public needs. On the other hand, although some studies have incorporated social media big data to mine public behavior or emotions, they mostly analyze visit volume or emotional tendencies in isolation, failing to establish a systematic relationship between landscape, behavior, and emotion. This makes it impossible to reveal the inherent synergy and differentiation between the park's objective environment and subjective perception. Furthermore, existing assessments lack indicators to quantify the combined effect of these three factors, making it difficult to accurately identify shortcomings in park quality and providing targeted guidance for differentiated quality improvement of parks of different types and in different regions.

[0004] Therefore, there is an urgent need to construct a multi-dimensional and collaborative evaluation framework that integrates objective landscape data, public behavior data, and emotional perception data. By quantifying collaborative relationships and differentiation characteristics, it is possible to achieve accurate evaluation of the quality of urban parks and provide a scientific basis for optimizing urban park planning and improving management. Summary of the Invention

[0005] The purpose of this invention is to propose a method for evaluating the quality of urban parks based on the synergistic differentiation of landscape, behavior, and emotion. This method aims to address the problems of the separation of subjective and objective indicators, lack of synergistic analysis, and insufficient evaluation accuracy in existing urban park quality evaluation methods. It provides a method for evaluating the quality of urban parks based on the synergy and differentiation of landscape, behavior, and emotion, enabling multi-dimensional quantitative evaluation of park quality, accurately identifying quality shortcomings, and providing targeted strategies for improving park quality.

[0006] To achieve the above objectives, this invention proposes a method for evaluating the quality of urban parks based on the synergistic differentiation of landscape, behavior, and emotion. The specific steps are as follows: Step S1: Collect landscape data, behavioral data, and emotional perception data, and preprocess the collected data respectively; Step S2: Construct an urban park quality evaluation index system based on three dimensions: park landscape quality, park visit behavior, and park emotional perception; determine the core indicators of the index system and standardize them. Step S3: Calculate the comprehensive experience index to quantify the synergistic effect of park landscape quality, park visit behavior, and park emotional perception. Step S4: Conduct differential analysis on park landscape quality, park visit behavior, and park emotional perception, including moderating effect analysis, bivariate relationship mapping, and typological differential analysis. Step S5: Generate the spatiotemporal assessment results of park quality and differentiated quality improvement strategies.

[0007] Preferably, in step S1, the landscape data includes park boundary data and remote sensing image data. After preprocessing the landscape data, the maximum likelihood method is used to classify the park land use type and calculate the landscape pattern index. The behavioral data is the amount of social media comments about the park, which is represented by the number of park visits. The sentiment data is the text content of the park comments. After deduplication, stop word removal and word segmentation preprocessing, the text content is used for sentiment analysis.

[0008] Preferably, the park land use types include vegetation, water bodies, roads and impermeable surfaces, building and facility land and other land types.

[0009] Preferably, in step S2, a comprehensive evaluation of the park's landscape quality is conducted based on four dimensions: green coverage rate, landscape shape complexity, landscape heterogeneity, and landscape diversity, as detailed below: Green coverage rate includes vegetation coverage rate and water body coverage rate, as shown in the following formula: vegetation coverage ; Water coverage rate = ; in, For vegetation area, For water body area, The area of ​​the park; Landscape shape complexity refers to the complexity of the shapes of patches within the park, and is calculated using the Landscape Shape Index (LSI). The formula is as follows: ; in, E The total length of all patches in the park; Landscape heterogeneity refers to the degree to which different land use types are separated by boundaries, and is calculated using edge density (ED), as shown in the following formula: ; Landscape diversity, representing the richness of different land use types, is calculated using the Shannon Diversity Index (SHDI), as shown in the following formula: ; in, m This represents the total number of plaque types. For the first i The probability of plaque-like formation; All indicators were normalized using the Min-Max standardization method, and the four dimensions of the processed indicators were combined into a final score for the park's landscape quality using the full permutation polygon comprehensive index method, as shown in the following formula: ; in, S For the quality of the park landscape, For the first i A single indicator value, For the first j A single indicator value, n The number of individual indicators.

[0010] Preferably, in step S2, the emotional perception of the park is mined, and the steps are as follows: Step S21: Word frequency statistics. Sentiment words are extracted based on the text feature weighting algorithm TF-IDF, as shown in the following formula: ; in, For word frequency, for , For words i In the text j The number of times it appears in The total number of texts in the corpus. For words contained in the corpus i The number of texts; Step S22: Select comment data as the basis for evaluating park satisfaction. Based on the sentiment dictionary, score the extracted sentiment words. Positive sentiment words are scored as positive numbers, and negative sentiment words are scored as negative numbers. Sum the scores of all sentiment words to obtain the sentiment value of the comment text. Calculate the mean sentiment value to obtain the park satisfaction value, as shown in the following formula: ; in, For park satisfaction, For the first i The first park j The score of each comment, For the first i Total number of comments for each park.

[0011] Preferably, in step S3, the normalized park landscape quality, park visit volume, and park satisfaction are calculated using vector weighting to quantify the overall performance of the park across different time and spatial dimensions. The formula is as follows: ; ; ; ; in, For the overall experience index, To normalize the quality of the park landscape, This represents the normalized park visit count. For normalized park satisfaction, and These represent the maximum and minimum values ​​for park landscape quality, respectively. and These represent the maximum and minimum park visitor numbers, respectively. and These represent the maximum and minimum values ​​of park satisfaction, respectively.

[0012] Preferably, in step S4, the moderating effect analysis uses park visit volume as the dependent variable, park satisfaction as the independent variable, and park landscape quality as the moderating variable. Through regression analysis, the relationship between park satisfaction and park landscape quality on park visit volume is explored, as shown in the following formula: ; in, As the dependent variable, The intercept is... For regression coefficients, To adjust the regression coefficient of the variable, The regression coefficients of the interaction term, As the independent variable, To adjust variables, For interactive items, This is the error term.

[0013] Preferably, in step S4, the bivariate relation mapping is used to analyze the interrelationships between park visit volume, park satisfaction, and park landscape quality in the park quality assessment indicators. Specifically, three bivariate color-coded matrices are constructed, each focusing on the following pairwise relationships: park visit volume and park satisfaction, park visit volume and park landscape quality, and park satisfaction and park landscape quality. In the bivariate color-coded matrices, the natural discontinuity grading method is used to divide each variable into high and low categories, and each pair of variables is combined to form four different combination methods, corresponding to the following four relationships: high-high, high-low, low-high, and low-low, in order to explore the relationships between the variables in the park quality assessment indicators and their distribution characteristics.

[0014] Therefore, this invention proposes a method for evaluating the quality of urban parks based on the synergistic differentiation of landscape, behavior, and emotion, with the following beneficial effects: (1) This invention integrates three-dimensional indicators of landscape, behavior and emotion for the first time, which solves the problem of the separation of subjective and objective evaluation in traditional evaluation. By quantifying the synergistic effect of the three through the comprehensive experience index, a comprehensive evaluation of park quality is achieved.

[0015] (2) This invention uses moderating effect analysis and bivariate mapping to accurately identify different types of parks, such as those with high visit volume and low satisfaction, and those with high landscape quality and low visit volume, providing a basis for differentiated quality improvement.

[0016] (3) This invention integrates remote sensing images, social media big data and natural language processing technology, with low data acquisition cost, high timeliness, and strong scalability and wide applicability of the evaluation method.

[0017] (4) The spatial differentiation and type differentiation quality improvement strategies generated by this invention can directly provide operational guidelines for urban planners and park managers, helping to maximize the efficient use of park resources and service benefits.

[0018] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0019] Figure 1 This is a flowchart of an urban park quality assessment method based on the synergistic differentiation of landscape, behavior, and emotion, according to the present invention. Figure 2 This is a schematic diagram of the spatiotemporal statistics of park landscape quality in an embodiment of the present invention; wherein, Figure 2 (a) in the figure is a schematic diagram of the spatiotemporal statistics of park landscape quality in 2017. Figure 2 (b) in the figure is a schematic diagram of the spatiotemporal statistics of park landscape quality in 2020. Figure 2 (c) in the figure is a schematic diagram of the spatiotemporal statistics of park landscape quality in 2023. Figure 2(d) in the figure is a comparative diagram of park landscape quality in 2017, 2020, and 2023; Figure 3 This is a schematic diagram of the spatial distribution of park review numbers in an embodiment of the present invention; wherein, Figure 3 (a) in the figure is a spatial distribution diagram of the number of park reviews in 2017. Figure 3 (b) in the diagram shows the spatial distribution of the number of park reviews in 2020. Figure 3 (c) in the figure is a spatial distribution diagram of the number of park reviews in 2023; Figure 4 This is a schematic diagram illustrating the monthly changes in the number of park reviews in an embodiment of the present invention; wherein, Figure 4 (a) in the figure shows the monthly changes in the number of park reviews in 2017. Figure 4 (b) in the figure shows the monthly changes in the number of park reviews in 2020. Figure 4 (c) in the figure is a schematic diagram showing the monthly changes in the number of park reviews in 2023; Figure 5 This is a park emotion heatmap in an embodiment of the present invention; wherein, Figure 5 (a) in the image is a heat map of park emotions in 2017. Figure 5 (b) in the image is the 2020 park emotional heat map. Figure 5 (c) in the figure is the 2023 park emotional heat map; Figure 6 This is a park emotional grid diagram in an embodiment of the present invention; wherein, Figure 6 (a) in the image is the 2017 park emotional grid map. Figure 6 (b) in the image is the 2020 park emotional grid map. Figure 6 (c) in the image represents the 2023 park emotional grid map; Figure 7 This is a schematic diagram of the spatiotemporal distribution of the park's comprehensive experience index in an embodiment of the present invention; wherein, Figure 7 (a) in the figure is a schematic diagram of the spatiotemporal distribution of the park comprehensive experience index in 2017. Figure 7 (b) in the diagram is a schematic diagram of the spatiotemporal distribution of the park comprehensive experience index in 2020. Figure 7 (c) in the figure is a schematic diagram of the spatiotemporal distribution of the park comprehensive experience index in 2023. Figure 7 (d) in the figure is a comparative diagram of the comprehensive experience index of different types of parks in 2017, 2020 and 2023; Figure 8 This is a schematic diagram illustrating the comparative analysis of the park's comprehensive experience index in different years in an embodiment of the present invention; wherein, Figure 8 (a) in the figure is a comparative analysis diagram of the park's comprehensive experience index in 2017 and 2020. Figure 8(b) in the figure is a schematic diagram comparing the park's comprehensive experience index in 2020 and 2023; Figure 9 This is a schematic diagram of a bivariate relationship mapping in an embodiment of the present invention; wherein, Figure 9 In the diagram, (a) represents the bivariate mapping between park satisfaction and visitor volume. Figure 9 (b) in the figure represents the bivariate mapping of the relationship between park landscape quality and visitor volume. Figure 9 (c) in the figure represents the bivariate relationship mapping between park landscape quality and satisfaction; Figure 10 This is a geometrical diagram of the Comprehensive Experience Index (CEI) in this invention. Detailed Implementation

[0020] To make the technical solutions, advantages, and objectives of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below. The described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention without creative effort are within the protection scope of the present invention.

[0021] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0022] Example like Figure 1 As shown, this invention provides a method for assessing the quality of urban parks based on the synergistic differentiation of landscape, behavior, and emotion. The specific steps are as follows: Step S1: Collect landscape data, behavioral data, and emotional perception data, and preprocess the collected data respectively; This study focuses on City D, which has a resident population of 21.893 million and a total green area of ​​94,135.52 hectares, representing a green coverage rate of 47.28%. Of this, 37,238.20 hectares are parkland. Given City D's unique high population density and high green space usage intensity, this study selects the area within the Sixth Ring Road as its research scope and, based on data from the G-Review platform, selects 82 representative parks with over 100 reviews as specific research subjects.

[0023] Vector data of the city's boundaries and ring roads, as well as basic data of the city's parks, were obtained from the OpenStreetMap platform. Additionally, 10-meter resolution remote sensing imagery data released by the F Aviation Administration in 2023 was acquired. After image preprocessing, the maximum likelihood method in supervised classification was used in ENVI 5.3 software to classify land use types within the parks into five categories: vegetation, water bodies, roads and impervious surfaces, building and facility land, and other land types. The supervised classification results showed an accuracy rate of 93.3138% and a Kappa coefficient of 0.9035 in 2017; 92.0081% and a Kappa coefficient of 0.8835 in 2020; and 93.6164% and a Kappa coefficient of 0.9049 in 2023. Finally, the landscape pattern index of the parks was calculated using Fragstats 4.2 software.

[0024] Using G-Reviews as the data source, as of April 2023, the platform had 18.9 million daily active users. User reviews of 82 parks within the Sixth Ring Road of City D were obtained. To ensure the timeliness and representativeness of the data, reviews from three complete years—2017, 2020, and 2023—were selected, yielding 169,845 valid data entries. Because the original review texts contained a large amount of information irrelevant to the topic, the obtained data underwent preprocessing, including deduplication, stop word removal, and word segmentation.

[0025] Step S2: Construct an urban park quality evaluation index system based on three dimensions: park landscape quality, park visit behavior, and park emotional perception; determine the core indicators of the index system and standardize them. The park's landscape quality is comprehensively evaluated based on four dimensions: green coverage rate, landscape shape complexity, landscape heterogeneity, and landscape diversity, as detailed below: Green coverage rate includes vegetation coverage rate and water body coverage rate, as shown in the following formula: vegetation coverage ; Water coverage rate = ; in, For vegetation area, For water body area, The area of ​​the park; Landscape shape complexity refers to the complexity of the shapes of patches within the park, and is calculated using the Landscape Shape Index (LSI). The formula is as follows: ; in, E The total length of all patches in the park; Landscape heterogeneity refers to the degree to which different land use types are separated by boundaries, and is calculated using edge density (ED), as shown in the following formula: ; Landscape diversity, representing the richness of different land use types, is calculated using the Shannon Diversity Index (SHDI), as shown in the following formula: ; in, m This represents the total number of plaque types. For the first i The probability of plaque-like formation; All indicators were normalized using the Min-Max standardization method, and the four dimensions of the processed indicators were combined into a final score for the park's landscape quality using the full permutation polygon comprehensive index method, as shown in the following formula: ; in, S For the quality of the park landscape, For the first i A single indicator value, For the first j A single indicator value, n The number of individual indicators.

[0026] The steps to explore emotional perceptions of parks are as follows: Step S21: Word frequency statistics. Sentiment words are extracted based on the text feature weighting algorithm TF-IDF, as shown in the following formula: ; in, For word frequency, for , For words i In the text j The number of times it appears in The total number of texts in the corpus. For words contained in the corpus i The number of texts; Step S22: Select comment data as the basis for evaluating park satisfaction. Based on the sentiment dictionary, score the extracted sentiment words. Positive sentiment words are scored as positive numbers, and negative sentiment words are scored as negative numbers. Sum the scores of all sentiment words to obtain the sentiment value of the comment text. Calculate the mean sentiment value to obtain the park satisfaction value, as shown in the following formula: ; in, For park satisfaction, For the first i The first park jThe score of each comment, For the first i Total number of comments for each park.

[0027] like Figure 2 As shown, based on the calculation results of the full permutation polygon comprehensive index method, the park's landscape quality is between 0.3 and 0.8. Figure 2 As shown in (a), (b), and (c), the spatial distribution of park landscape quality exhibits a significant unevenness: parks with lower landscape quality are mainly concentrated within the Second Ring Road and between the Fifth and Sixth Ring Roads, while parks with higher landscape quality are more evenly distributed among the ring roads. In terms of temporal variation, the overall landscape quality of the parks shows relatively little change. Although the landscape quality of parks within the Second Ring Road has gradually improved, the landscape quality of some parks within the Second Ring Road and between the Fifth and Sixth Ring Roads remains relatively low. Park H maintained a high level of landscape quality in 2017, 2020, and 2023, while the landscape quality of Forest Park I has consistently needed improvement. Figure 2 As shown in (d), the dispersion of park landscape quality was relatively large in 2017, while the dispersion of park landscape quality data decreased in 2020 and 2023. This indicates that although the overall level of park landscape quality within the Sixth Ring Road of City D did not change much, the differences between parks gradually narrowed over time.

[0028] like Figure 3 As shown, in 2017, although the overall visitor volume of parks was relatively low, parks within the Second Ring Road still attracted a relatively large number of visitors. In 2020 and 2023, park visitor volume increased significantly, accompanied by increased spatial heterogeneity: parks within the Third Ring Road were popular destinations with high visitor density, while visitor volume gradually decreased as the geographical location extended outwards; in the area from the Fourth to the Sixth Ring Road, although there were a certain number of parks with medium and high visitor volumes, their distribution was relatively sparse, indicating that park use is easily affected by geographical location.

[0029] like Figure 4 As shown, in 2017, park visits saw a slight peak in April, with some parks also experiencing peak visits in October. While park visits were influenced by seasonality that year, the impact was relatively limited. In 2020, park visits exhibited a more pronounced seasonal fluctuation, with significant peaks in April and October, a slight peak in July, and a significant trough in February. In 2023, April and October remained peak periods for visits, with relatively gradual changes in visits in other months, indicating that park visits still exhibited a certain seasonal pattern. Although the fluctuation range of park visits varied across years, the seasonal trend was significant, suggesting a peak season for park usage.

[0030] The TF-IDF algorithm was used to obtain the feature words of the park review text. After filtering, the top 20 feature words for both large and small parks were extracted. The higher the ranking, the more important the word is in all park reviews. The feature word ranking is shown in Table 1.

[0031] Table 1. Top 20 TF-IDF Feature Words for Parks

[0032] Table 1 shows that the inclusion of keywords such as "tickets," "free," "children," "good," "suitable," "like," "large," "weather," "scenery," "time," and "weekend" in all three years indicates that tourists generally gave the park a high rating, both before and after the pandemic. They also paid close attention to the park's scenery and accessibility, the best time to visit, and weather conditions, all of which influenced their willingness to visit. In 2017 and 2023, tourists focused more on the park's distinctive architecture and attractions, while in 2020, keywords such as "pandemic," "reservation," "entrance," "transportation," "parking lot," and "parking" emerged, highlighting tourists' concerns about park service management, health and safety, and accessibility during the pandemic. This change reflects how the pandemic shifted tourists' needs and focus, with service facilities and mobility becoming important factors in their park evaluation. Furthermore, from 2017 to 2023, the ranking of "tickets" declined, while "good" rose to the top, indicating that public attention to parks in City D gradually shifted from park accessibility to a better overall experience.

[0033] like Figure 5 As shown, in 2017, parks within the Third Ring Road had high emotional value, forming a cluster of high-emotional-value parks; a small number of high-emotional-value parks were distributed between the Third and Fourth Ring Roads; although a few high-emotional-value parks were also distributed near the North Fifth Ring Road and East Fifth Ring Road, parks outside the Fourth Ring Road generally had lower emotional value. In 2020, the emotional value of parks within the Fourth Ring Road decreased, while the emotional value of parks outside the Fourth Ring Road slightly increased, with the area within the Third Ring Road, between the North Fourth Ring Road and the North Fifth Ring Road becoming the emotionally vibrant zone. In 2023, although the emotional value of some parks within the Second Ring Road and between the North Fourth Ring Road and the North Fifth Ring Road decreased, the emotional value of parks within the Third Ring Road remained strong; the emotional value of parks far from the city center increased, while the overall emotional vibrancy decreased. The pandemic affected public emotional perception of parks, leading to increased demands for park services and decreased satisfaction. Although the city center has always been a zone of high emotional park sentiment, over time, emotional perception of park spaces has gradually become more balanced, and the differences in public park emotional experiences between different areas have gradually narrowed.

[0034] like Figure 6As shown, the spatial distribution of park emotional values ​​remained largely consistent across 2017, 2020, and 2023. Parks within the Second Ring Road exhibited significantly higher emotional values, forming an emotional "high ground," while the emotional values ​​gradually decreased as one moved outwards from the city center. In the northeastern area of ​​the Fifth Ring Road, significant differences in emotional values ​​were observed, reflecting a diverse range of emotional experiences. The spatial gradient characteristic of park emotional values, gradually decreasing from the city center to the periphery, reflects the hierarchical and regional differences in emotional distribution.

[0035] Step S3: Calculate the comprehensive experience index to quantify the synergistic effect of park landscape quality, park visit behavior, and park emotional perception. like Figure 10 As shown, the normalized park landscape quality, park visit volume, and park satisfaction are calculated using vector weighting to quantify the overall performance of the park across different time and spatial dimensions. The formula is as follows: ; ; ; ; in, For the overall experience index, To normalize the quality of the park landscape, This represents the normalized park visit count. For normalized park satisfaction, and These represent the maximum and minimum values ​​for park landscape quality, respectively. and These represent the maximum and minimum park visitor numbers, respectively. and These represent the maximum and minimum values ​​of park satisfaction, respectively.

[0036] like Figure 7As shown, the quality of urban parks is measured by the comprehensive experience index. Parks with a higher comprehensive experience index are considered to be of better quality. The comprehensive experience index is divided into five categories using the natural discontinuity grading method, presented in map form. In 2017, the overall experience of parks within the Sixth Ring Road of City D was poor; some parks within the Third Ring Road and between the North Fourth and North Fifth Ring Roads had a better experience; and parks between the East Fourth and East Fifth Ring Roads and outside the Fifth Ring Road had a poor experience. In 2020, the overall comprehensive experience of parks improved significantly, with the experience index of parks within each ring road area showing varying degrees of growth. However, the comprehensive experience index of parks outside the Fifth Ring Road remained at a low level, representing a weakness hindering the balanced development of park service quality within the Sixth Ring Road of City D. In 2023, the overall comprehensive experience of parks tended to stabilize. The experience of parks within the Second Ring Road improved significantly; the comprehensive experience of parks between the Second and Third Ring Roads was good and remained stable; the experience of parks between the Third and Fifth Ring Roads varied greatly, with some experiencing good and some bad; and the experience of parks outside the Fifth Ring Road decreased somewhat, remaining at a low level. These changes profoundly reveal the spatial heterogeneity of park quality: parks closer to the city center generally have higher quality; while parks farther from the city center are relatively lagging in quality due to various constraints.

[0037] According to the classification standards proposed by the D City Parks and Green Spaces Bureau, the 82 parks are divided into four categories: historical parks, specialized parks, comprehensive parks, and community parks. Among them, there are 13 historical parks, 34 comprehensive parks, 7 community parks, and 28 specialized parks. Figure 7 As shown in (d), comparing the average comprehensive experience index of various park types in 2017, 2020, and 2023, the ranking of the comprehensive experience intensity of each type of park is as follows: Historical Parks > Comprehensive Parks > Community Parks > Specialized Parks. The comprehensive experience of historical parks remained at the highest level throughout the three years, with the smallest annual variation; while the comprehensive experience of community parks showed greater annual fluctuations, with the largest differences. This indicates that City D has achieved the protection and utilization of historical and cultural heritage in its park and green space services and development, and also highlights the public's demand for cultural experience. The comprehensive experience index of all four types of parks reached its highest value in 2020, indicating that the public's perception of park experiences improved during the pandemic, and that the quality of the parks was relatively high.

[0038] like Figure 8 As shown in (a) of the table, compared with 2017, the overall experience index of most parks, except for a few, has significantly improved in 2020. The quality of parks has been effectively optimized and improved during this period, and visitors have experienced richer, more diverse, and higher-quality services. Figure 8As shown in (b), the overall experience index of the park remained relatively stable between 2020 and 2023. The quality of the park did not change significantly during this period and maintained a relatively stable level.

[0039] Step S4: Conduct differential analysis on park landscape quality, park visit behavior, and park emotional perception, including moderating effect analysis, bivariate relationship mapping, and typological differential analysis. The moderating effect analysis used park visit volume as the dependent variable, park satisfaction as the independent variable, and park landscape quality as the moderating variable. Through regression analysis, it explored the relationship between park satisfaction and park landscape quality on park visit volume. The formula is as follows: Model 1: Basic Relationship between Independent and Dependent Variables ; Model 2: Incorporating moderating variables: ; Model 3: Adding interactive elements: ; in, As the dependent variable, The intercept is... For regression coefficients, To adjust the regression coefficient of the variable, The regression coefficients of the interaction term, As the independent variable, To adjust variables, For interactive items, This is the error term.

[0040] Taking 2023 as an example, the independent variable (park satisfaction) was included in the regression equation, followed by the moderating variable (park landscape quality). Finally, the interaction term between the independent and moderating variables was incorporated into the regression equation. Table 2 shows the results of the moderating effect analysis of the relationship between park landscape quality and park visit volume and park satisfaction in 2023. The results of Model 1 show that, without considering the interference of the moderating variable (park landscape quality), the independent variable (park satisfaction) is significant (t=2.809, p=0.006<0.05), meaning that park satisfaction has a positive impact on park visit volume. The results of Model 2 show that the moderating variable (park landscape quality) is significant (t=3.924, p=0.000<0.01), indicating that park landscape quality also has a positive impact on park visit volume. Based on this, the results of Model 3 show that the interaction term (park satisfaction × park landscape quality) is not significant (t=1.090, p=0.279>0.05), which means that when park satisfaction affects park visits, the influence of park landscape quality remains consistent at different levels. Therefore, park landscape quality does not have a moderating effect on the relationship between park satisfaction and park visits.

[0041] Table 2. Results of the Moderating Effect Analysis

[0042] like Figure 9 As shown, taking 2023 as an example, the spatial distribution of the relationships between different variables shows significant differences. Parks with high values ​​for all three variables are concentrated within the Second Ring Road, while parks with low values ​​for all three variables are more abundant near the Fifth Ring Road and between the Fifth and Sixth Ring Roads. Figure 9 As shown in (a), parks with high visitor volume and low satisfaction, and parks with low visitor volume and high satisfaction, are widely distributed along the ring roads. For example... Figure 9 As shown in (b), there are fewer parks with high visitor volume and low landscape quality, while there are more parks with low visitor volume and high landscape quality, concentrated in the eastern and northern regions. Figure 9 As shown in (c), parks with high satisfaction but low landscape quality are concentrated between the North Fourth Ring Road and the North Fifth Ring Road; parks with low satisfaction but high landscape quality are mainly distributed within the Second Ring Road, with fewer outside the Fifth Ring Road. However, not all highly visited parks have good landscape quality and satisfaction.

[0043] The quality of park landscape, park visits and park satisfaction also vary among different types of parks: (1) Historical parks usually have high quality of park landscape, park visits and park satisfaction. These parks attract a large number of tourists due to their profound cultural heritage, unique historical value and high quality of park landscape, and the overall satisfaction of tourists is high. (2) Parks with low quality of park landscape, park visits and park satisfaction are mostly specialized parks, especially those within and outside the Fifth Ring Road. They are mostly natural parks or ecological parks with rich vegetation resources, lacking sufficient attractiveness and diverse service facilities. In planning and design, it is necessary to further improve the public's experience under the premise of protecting the ecological effect. Attention should be paid to the internal landscape design of the park, improving service facilities, and exploring the park's characteristics to attract more tourists to visit and experience. (3) The biggest problem of comprehensive parks is that the park visits are low, which cannot match the park satisfaction and park landscape quality. Some parks may have low visits due to their remote location, and some parks may be affected by the siphon effect of the surrounding well-known parks. Therefore, in the planning and design process, the focus should be on increasing park visits. The outer ring comprehensive park should focus on optimizing accessibility, while the inner ring needs to combine with surrounding facilities to provide diverse and distinctive activity spaces to enhance overall attractiveness. (4) Most community parks have relatively good landscape quality, but there is still room for improvement in terms of park visits and park satisfaction. These parks are usually places for community residents to relax, exercise and socialize. In planning and design, attention should be paid to the construction of park infrastructure and service management to ensure that park facilities are complete, the environment is beautiful and easy to maintain, so as to create a comfortable and pleasant leisure space and enhance residents' sense of happiness.

[0044] Step S5: Generate the spatiotemporal assessment results of park quality and differentiated quality improvement strategies.

[0045] The above analysis reveals that the overall experience index of parks on the outskirts of City D remains consistently low, and the quality of parks within the Sixth Ring Road exhibits significant spatial heterogeneity, indicating a need to prioritize landscape enhancement and facility improvement in areas outside the Fifth Ring Road. For different types of parks, historical parks demonstrate the highest overall experience index, exhibiting high levels of landscape quality, visitor numbers, and satisfaction, highlighting the value of historical and cultural heritage in park design. In contrast, specialized parks show lower levels of landscape quality, visitor numbers, and satisfaction, lacking service diversity, emphasizing the need to strike a balance between ecological protection and recreational functions, and focusing on improving park attractiveness. Comprehensive parks have low visitor numbers, and efforts should be made to increase park awareness and enhance their usability. Community parks have relatively good landscape quality, but there is room for improvement in visitor numbers and satisfaction; therefore, infrastructure construction and management should be strengthened to create more comfortable urban recreational spaces.

[0046] Seasonal visitation trends show peak park usage in spring and autumn, which can provide a basis for resource allocation in park management, maintenance, and activity planning. During peak seasons, crowd control and facility maintenance should be strengthened. Park usage preference analysis shows that park features, services, and accessibility play a significant role in public perception. Meanwhile, the increase in park visits and decrease in satisfaction caused by the pandemic highlights the important role of urban parks in public health, prompting policymakers to prioritize the equitable distribution and accessibility of parks and to prepare adequately for and respond to emergencies.

[0047] It is worth noting that all contents not described in detail in this invention are existing technologies and are well known to those skilled in the art.

[0048] Therefore, this invention provides a method for evaluating the quality of urban parks based on the synergistic differentiation of landscape, behavior, and emotion. By integrating three-dimensional data of landscape, behavior, and emotion, it quantifies the synergistic effects and differentiation characteristics, achieving a comprehensive and accurate evaluation of the quality of urban parks. This provides a scientific basis for differentiated quality improvement and helps maximize the efficient use of park resources and service benefits.

[0049] 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 them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for evaluating the quality of urban parks based on the synergistic differentiation of landscape, behavior, and emotion, characterized in that, The specific steps are as follows: Step S1: Collect landscape data, behavioral data, and emotional perception data, and preprocess the collected data respectively; Step S2: Construct an urban park quality evaluation index system based on three dimensions: park landscape quality, park visit behavior, and park emotional perception; determine the core indicators of the index system and standardize them. Step S3: Calculate the comprehensive experience index to quantify the synergistic effect of park landscape quality, park visit behavior, and park emotional perception. Step S4: Conduct differential analysis on park landscape quality, park visit behavior, and park emotional perception, including moderating effect analysis, bivariate relationship mapping, and typological differential analysis. Step S5: Generate the spatiotemporal assessment results of park quality and differentiated quality improvement strategies.

2. The urban park quality assessment method based on the synergistic differentiation of landscape, behavior, and emotion as described in claim 1, characterized in that, In step S1, the landscape data includes park boundary data and remote sensing image data. After preprocessing the landscape data, the maximum likelihood method is used to classify the park land use type and calculate the landscape pattern index. The behavioral data is the amount of social media comments about the park, which is represented by the number of park visits. The sentiment data is the text content of the park comments. After deduplication, stop word removal and word segmentation preprocessing, the text content is used for sentiment analysis.

3. The urban park quality assessment method based on the synergistic differentiation of landscape, behavior, and emotion as described in claim 2, is characterized in that... Park land use types include vegetation, water bodies, roads and impermeable surfaces, building and facility land and other land types.

4. The urban park quality assessment method based on the synergistic differentiation of landscape, behavior, and emotion as described in claim 1, characterized in that... In step S2, a comprehensive evaluation of the park's landscape quality is conducted based on four dimensions: green coverage rate, landscape shape complexity, landscape heterogeneity, and landscape diversity, as detailed below: Green coverage rate includes vegetation coverage rate and water body coverage rate, as shown in the following formula: vegetation coverage ; Water coverage rate = ; in, For vegetation area, For water body area, The area of ​​the park; Landscape shape complexity refers to the complexity of the shapes of patches within the park, and is calculated using the Landscape Shape Index (LSI). The formula is as follows: ； in, E The total length of all patches in the park; Landscape heterogeneity refers to the degree to which different land use types are separated by boundaries, and is calculated using edge density (ED), as shown in the following formula: ； Landscape diversity, representing the richness of different land use types, is calculated using the Shannon Diversity Index (SHDI), as shown in the following formula: ； in, m This represents the total number of plaque types. For the first i The probability of plaque-like formation; All indicators were normalized using the Min-Max standardization method, and the four dimensions of the processed indicators were combined into a final score for the park's landscape quality using the full permutation polygon comprehensive index method, as shown in the following formula: ； in, S For the quality of the park landscape, For the first i A single indicator value, For the first j A single indicator value, n The number of individual indicators.

5. The urban park quality assessment method based on the synergistic differentiation of landscape, behavior, and emotion as described in claim 1, characterized in that, In step S2, the emotional perception of the park is explored, and the steps are as follows: Step S21: Word frequency statistics. Sentiment words are extracted based on the text feature weighting algorithm TF-IDF, as shown in the following formula: ； in, For word frequency, for , For words i In the text j The number of times it appears in The total number of texts in the corpus. For words contained in the corpus i The number of texts; Step S22: Select comment data as the basis for evaluating park satisfaction. Based on the sentiment dictionary, score the extracted sentiment words. Positive sentiment words are scored as positive numbers, and negative sentiment words are scored as negative numbers. Sum the scores of all sentiment words to obtain the sentiment value of the comment text. Calculate the mean sentiment value to obtain the park satisfaction value, as shown in the following formula: ； in, For park satisfaction, For the first i The first park j The score of each comment, For the first i Total number of comments for each park.

6. The urban park quality assessment method based on the synergistic differentiation of landscape, behavior, and emotion as described in claim 1, characterized in that, In step S3, the normalized park landscape quality, park visit volume, and park satisfaction are calculated using vector weighting to quantify the overall performance of the park across different time and spatial dimensions. The formula is as follows: ； ； ； ； in, For the overall experience index, To normalize the quality of the park landscape, This represents the normalized park visit count. For normalized park satisfaction, and These represent the maximum and minimum values ​​for park landscape quality, respectively. and These represent the maximum and minimum park visitor numbers, respectively. and These represent the maximum and minimum values ​​of park satisfaction, respectively.

7. The urban park quality assessment method based on the synergistic differentiation of landscape, behavior, and emotion as described in claim 1, characterized in that, In step S4, the moderating effect analysis uses park visit volume as the dependent variable, park satisfaction as the independent variable, and park landscape quality as the moderating variable. Through regression analysis, it explores the relationship between park satisfaction and park landscape quality on park visit volume, as shown in the following formula: ； in, As the dependent variable, The intercept is... For regression coefficients, To adjust the regression coefficient of the variable, The regression coefficients of the interaction term, As the independent variable, To adjust variables, For interactive items, This is the error term.

8. The urban park quality assessment method based on the synergistic differentiation of landscape, behavior, and emotion as described in claim 1, characterized in that... In step S4, the bivariate relation mapping is used to analyze the interrelationships between park visit volume, park satisfaction and park landscape quality in the park quality assessment indicators. Specifically, three bivariate color scale matrices are constructed, each focusing on the following pairwise relationships: park visit volume and park satisfaction, park visit volume and park landscape quality, and park satisfaction and park landscape quality. In the bivariate color scale matrix, the natural discontinuity grading method is used to divide each variable into two categories: high and low. Each pair of variables is then combined to form four different combinations, corresponding to the following four relationships: high-high, high-low, low-high, and low-low, in order to explore the relationships between the variables in the park quality assessment indicators and their distribution characteristics.

Images