A Study on the Correlation between Street Environment Characteristics and Tourists’ Emotional Perception (https://doi.org/10.63386/621466)

Yitao Wang^a , Zhian Liu^b*

The School of Arts and Media, Kunming University of Science and Technology, Kunming, Yunnan, 650500, China

^aEmail: 17736368198@163.com

^bEmail: utsky@163.com

Abstract:

In order to explore the effect of street environment characteristics on tourists’ emotional perception, this paper took Dali Old Town in Dali City, Yunnan Province, China as the research object, used Baidu Street View image data and the SegNet image semantic segmentation model of deep convolutional neural network to quantify and spatially visualize the four street environment characteristic indicators, namely, GVI, Sky Oppenness, Enclosedness, and Street Order, of the eleven main streets in Dali Old City. Then, we constructed the correlation matrix and used Random Forest to analyze the relationship between the street environment characteristics and tourists’ emotional perception. The results show that 1. In general, the higher the GVI and the lower the Sky Oppenness, the more positive the tourists’ emotional perception; the higher the Street Order, the more positive the tourists’ emotional perception; and the weaker the correlation between Enclosedness and tourists’ emotional perception compared with the other three environmental characteristics. 2. GVI is the most important factor for anxious perception, Sky Oppenness is the most important factor for pleasant perception, and Street Order is the most important factor for comfortable, satisfied, depressing, and boring perception. 3. There is a strong correlation between the environmental factors, and the emotional perception factors also affect each other. 4. In the old city, the streets with main transportation function and higher commercial value have better environmental quality and are better perceived by tourists, while the streets that are relatively narrow and have average commercial value have poor environmental quality and are poorly perceived by tourists.

Keywords: environmental characteristics; emotional perception; streetscape images; correlation; street space

Introduction:

Streets are an important part of the city, and their environmental characteristics directly affect the emotional experience of visitors [1-5]. However, in actual planning operations, the measurement of street space characteristics mostly relies on manually taken street view images, which cannot meet the large-scale and timely requirements of planning practice; some researchers also use image processing software to extract the proportion of pixels of each element in street view images sheet by sheet, thus realizing finer measurements. Although these manual methods can provide accurate research results in a small area, they are complicated and time-consuming in the process of data collection and processing due to technical constraints, which makes it difficult to be widely popularized and applied in practice. Meanwhile, machine learning technology is developing rapidly, and the machine learning algorithm represented by SegNet has become an important means for processing street view image information. With the help of street view images, many scholars have applied machine learning methods to analyze urban street environments and achieved good results. [6-10]

However, the quality of street space cannot be judged simply by identifying and counting various elements on the street, but also requires a human-centered urban planning approach to support the evolving urban landscape [11-13]. Therefore, it is necessary to quantify the feelings of tourists in the street environment from the psychological perspective of the crowd and to explore how the street environment affects the emotional perception of tourists. Many scholars have extensively studied the relationship between street environment characteristics and tourists’ emotional perception. Some studies have shown that greening is crucial to human mental health [14-19]. Many studies have listed the psychological benefits of vegetation on people from the perspective of environmental psychology [20-24]. Navarrete-Hernandez and Laffan (2019) found that greener local environments lead to higher levels of well-being and lower levels of stress [25]. Xu et al. (2023) used Random Forest Regression to reveal the nonlinear effects of street characteristics on human perception [26]. Harvey, C. et al. explored the effects of street view design on human perceptions of safety. [27] Asgarzadeh, M. et al. investigated the effects of buildings, trees, sky, and ground on human perception. [28] However, there are less systematic studies on the relationship between the street environment characteristics of the old city and the different emotional perceptions of tourists. In this paper, we took Dali Old City in Dali City, Yunnan Province, as the research object, and quantitatively analyzed the street environment characteristics by using Baidu Street View image data together with the SegNet image semantic segmentation model of deep convolutional neural network, based on which, we explored the correlation between the street environment characteristics and tourists’ emotional perception. This study aims to enrich the research system of the relationship between street environment characteristics and tourists’ emotional perception and provide theoretical support for optimizing the street environment in tourist cities.

1. Study Area and Methodology

1.1 Study Area

Dali Old City, located in the central part of Dali Bai Autonomous Prefecture in Yunnan Province, is a famous national historical and cultural city with profound historical accumulation and cultural connotation. The Old City was built in the fifteenth year of the Hongwu reign of the Ming Dynasty (1382), covering an area of 3 square kilometers. The city is about 1,000 meters wide from east to west and 2,000 meters long from north to south, with five main streets in the north-south direction and six main streets in the east-west direction. The location map of the study area is shown in Figure 1.

Figure 1 Location of the Study Area

1.2 Research Methodology

The technical roadmap of this paper is shown in Figure 2: first, using Pythonto to crawl Baidu street view images of Dali Old City Town, and then the SegNet model of deep convolutional neural network to semantically segment the street view images and extract the four environmental characteristic indicators, namely, GVI, SkyOpenness, Enclosedness and Street Order; then, collecting tourists’ emotional perception data by questionnaire survey; finally, using SPSS software to construct the correlation matrix and analyze the correlation among the factors, and Random Forest to analyze the importance of environmental characteristic indicators on tourists’ emotional perception, and discussing the research results.

Figure 2 Technology Roadmap

1.3 Measurement of Road Environment

1.3.1 Method of Selecting and Calculating Street Environmental Characteristic Factors

The selection of street characteristic factors is crucial. In this study, the following four environmental characteristic factors were selected as research objects based on the comprehensive consideration of existing literature, field work, and expert opinions: GVI, Sky Oppenness, Enclosedness, and Street Order.

Greening can not only beautify the environment and clean the air, but also improve people’s physical and mental health, which is an important way to improve the quality of life and promote public health. Traditional greening evaluation methods, such as measuring the green cover by remote sensing images, can’t simulate the street greening level from the human perspective well, so this study evaluated the street greening level by measuring the streets’ GVI, which is the proportion of green plants from the human perspective [29], and the expression of GVI is shown in Equation 1.

(Equation 1)

In Equation 1, pixcel tree is the number of pixels occupied by the tree,pixcel terrain is the number of pixels occupied by the lawn, and pixcel total is the total number of pixels in the street view image.

Sky Oppenness indicates the degree of visibility of the sky above the street, which affects the spatial perception and lighting conditions of the street. [30] The density and height of street trees and surrounding buildings all affect sky oppenness. In this study, sky oppenness was measured as the ratio of sky pixels to the total pixels in the street view image. The expression of sky oppenness is shown in Equation 2.

(Equation 2)

In Equation 2, pixcel sky is the number of pixels occupied by the sky, andpixcel total is the total number of pixels in the street view image.

Enclosedness is the degree of spatial enclosure of the street by buildings and trees on both sides of the street, which determines the spatial closure of the street. Different enclosedness will create different spatial feelings for tourists, which is measured by the ratio of the number of pixels of buildings and trees to the total number of pixels in the image. The expression of enclosedness is shown in Equation 3.

（Equation 3）

In Equation 3, pixel building is the number of pixels occupied by the building, pixcel tree is the number of pixels occupied by the trees, and pixcel total is the total number of pixels in the image.

Street order is concerned with the orderliness of the street environment, which includes many aspects such as architectural style, local cultural characteristics, pavement type, and outdoor public facilities, and is crucial for evaluating the quality of the street environment. However, it is difficult to evaluate the level of “street order” by machine learning methods, so this study constructed an evaluation system in which “street order” is used as the first-level indicator, which is divided into three second-level indicators: outdoor environment, architectural style, and the condition of public facilities. Under each second-level indicator, specific third-level indicators have been set, such as cleanliness, traffic order, and surface road condition, etc., to facilitate detailed field research and evaluation. By inviting experts to evaluate, the weights of the indicators at all levels were determined to ensure the scientific and rational nature of the evaluation system. Then, the research team composed of four landscape design students and teachers conducted field research on 11 streets in Dali Old City. The team scored the actual condition of each street according to the three-level indicators, with the scoring range from 0 to 1. The closer the value is to 1, the better the condition is, and the closer the value is to 0, the worse the condition is. Then the scoring data of the four researchers were averaged to obtain the Street Order value of each street. The evaluation indices and weight assignment of each level are shown in Table 1.

Table 1 Space Order Rating Scale

1.3.2 Data Collection

First, the SHP boundary of the Dali Old City area was sketched in ARCGIS, and then 564 sampling points were evenly distributed at an interval of 30 meters equidistant from each other, and the latitude and longitude coordinates of the sampling points were also recorded. Using the Baidu Street View API, a fixed viewing angle and a tilt angle were set for each image. In order to better simulate the human perspective, the vertical viewing angle was uniformly set to 0°, i.e., a flat view, at each sampling point. In terms of horizontal viewing angle, four street view images, four street view images parallel to the road, i.e., front view (0°), right view (90°), rear view (180°), and left view (270°) were crawled at each sampling point, and a total of 2,256 street view images were captured. The crawled images were accurately matched to the geographic coordinates of the points on the remotely sensed images, and each image was accompanied by detailed latitude and longitude information, providing a reliable database for subsequent visualization and analysis of street environment characteristics.

1.3.3 Image Segmentation

The image dataset collected in this study was semantically labeled in the labelme program for different objects on the street, such as walls, sky, trees, pedestrians, cars, roads, etc., to form a Json file. To achieve the goal of fast batch processing of images, a Json-To-Png image format batch processing program was written to preprocess the data for the next step of training in the data Segnet.

Then, a Segnet-based image semantic segmentation model was constructed, in which the encoder part used the first 13 convolutional layers of the pre-trained VGG16 network in order to extract image features, while the decoder part restored the image dimensions and performed pixel-level classification through upsampling and convolutional operations. In the model training phase, the pre-processed street view images were divided into training, validation, and test sets, and the cross-entropy loss function and Adam optimizer were used to update the weights. During the training process, the performance of the model on the validation set was monitored and the learning rate was adjusted as needed until the model performance was optimal or the preset number of iterations was reached. After the training, the model was applied to the test set of street view images to extract the street information and calculate the area fraction or the number of different elements in each image.

1.4 Selection of Tourists’ Emotional Perception Factors and Questionnaire Design

Based on the model of walkers’ psychological needs and emotions, six factors were screened: comfortable, satisfied, pleasant, anxious, depressing, boring, and five sets of word pairs were designed, which were: comfortable-uncomfortable, satisfied-dissatisfied, pleasant-unpleasant, anxious-relaxed, anxious-relaxed, depressing-happy, and boring-rich. The LikertScale was chosen to allow the subjects to rate the positive and negative strengths of each factor, to determine a 7-point subjective rating scale, and the terms of “very”, “somewhat”, and “slightly” were also used to differentiate between positive and negative feelings, and the values of -3, -2, -1, 0, 1, 2, and 3, respectively, were assigned to facilitate numerical quantification for subsequent quantitative analyses. A total of 120 questionnaires were distributed in the study, which included six rating factors, i.e., comfortable, satisfied, pleasant, anxious, depressing, and boring of the streets they were on. Volunteers rated each factor according to their subjective feelings about the street environment and summarized the data to assess the emotional perception of visitors in different streets. Thirty questionnaires were distributed on each street with the sample site, and a total of 330 questionnaires were distributed and 317 questionnaires were collected, with a validity rate of 96%.

2. Results and Analysis

2.1 Analysis of Street Environmental Characteristics

In order to compare the environmental characteristics of different streets, the image information of the sampling points of each street was statistically analyzed and the average value was taken to obtain the values of the four indicators, i.e. GVI, Sky Oppenness, Enclosedness, and Street Order for the 11 streets of Dali Old City, as shown in Table 2. And the intensity of the environmental characteristics of each street was visualized spatially, and the four environmental characteristics were represented by four different colors, as shown in Figures 3-6.

Table 2   Measurement Results of the Environmental Characteristics of Each Street

Figure 3 Spatial Distribution of GVI

Figure 4 Spatial Distribution of Sky Openness

Figure 5 Spatial Distribution of Enclosedness

Figure 6 Spatial Distribution of Street Order

Comparing the spatial distribution maps of different indicators, it can be found that for the six streets running from north to south, except for the outermost Zhonghe Road, the three indicators, i.e., GVI, Enclosedness, and Street Order, basically show the regularity that the central street (Yu’er Road) has higher value intensity, while the streets on both sides have lower value intensity.

From the field study, it can be seen that Yu’er Road, which connects the entrance of the west side of the city gate and the exit of the east side of the Old City, is wide, has many shops, high commercial value, high level of greenery, sidewalks on both sides, unified architectural style and perfect supporting facilities, which shows a high degree of street order. While Zhonghe Road, a street on the northern edge of the Old City, although with limited development and relatively low commercial value, is close on one side to the undeveloped land outside the Old City, where there is a dense distribution of tall and lush native trees, which in turn gives Zhonghe Road a higher GVI, lower sky oppenness, and higher enclosedness. People’s Road connects the west side of “Xijing Line” Highway at its western intersection, and Erhai Gate and Hongwu Road at its eastern intersection, so it definitely has motor vehicle traffic, but it is narrower and can’t provide the conditions for planting street trees on both sides of it, resulting in lower GVI, higher Sky oppenness and higher enclosedness; meanwhile, it doesn’t have continuous and complete sidewalks, resulting in mixed pedestrian and vehicular traffic, and its architectural style is also more chaotic, resulting in a lower degree of street order.

For the five streets running from east to west, the three indicators, i.e., GVI, Enclosedness, and Street Order are higher in the two streets closest to the outside (Bo’ai Road, Ye Yu Road) and lower in the central streets. According to the field survey research, Guangwu Road, the central street, is relatively narrow, with some parts are only 2.2m wide, and is limited in the area with the condition for planting street trees, resulting in a lower GIV; at the same time, Guangwu Road has mixed traffic, and lacks unity and continuity on its building facades; due to its narrowness and lack of necessary tourist service facilities, it has a lower degree of street order. Fuxing Road has a higher commercial value, reasonable width, the GVI, sky oppenness and enclosedness of this street show higher values, but the street has a more chaotic environment, resulting in a lower degree of street order.

2.2 Evaluation Results of Tourists’ Emotional Perception

In order to compare the differences in tourists’ emotional perception of different streets, the questionnaire data of 11 streets in Dali Old City were statistically analyzed, and the average value was taken to obtain the values of the six indicators of each street as shown in Table 3: Anxious, Comfortable, Depressing, Boring, Satisfied, and Pleasant. A box plot was drawn to analyze the distribution characteristics and the degree of dispersion of each emotion perception factor, as shown in Figure 7. The horizontal coordinates of this box plot are the 6 emotion perception factors and are represented by different colors. Each box plot shows the distribution of the data corresponding to the emotion perception factors, including the median, quartiles, and extreme values.

Table 3   Tourists’ Emotion Perception Results by Street

Figure 7 Box Plot of Tourists’ Emotion Perception

As can be seen from Figure 7, the mean values of the three positive emotion perception factors, i.e., comfortable, satisfied, pleasant, are higher than the three negative emotion perception factors, i.e., anxious, depressing, boring, and the median values of the positive emotion factors are above 0, while the median values of the two negative emotion factors, anxious and depressing, are below 0, indicating that the tourists’ evaluation of Dali Old City is more positive. The median value of the tourists’ perception of comfortable is located around 0.4, and the data are roughly distributed between -0.2 and 1.3, indicating that the tourists’ perception of comfortable varies widely among different streets. The median value of tourists’ perception of satisfied is located around 0.3, and the data are roughly distributed between -0.2 and 0.9, which is slightly more concentrated than the perception of comfortable. The values of tourists’ perception of pleasant are mostly above 0.2 and relatively concentrated, indicating that most of the streets are pleasant. The values of tourists’ perception of anxious and depressing are mostly concentrated above 0.2, indicating that a significant proportion of streets make tourists anxious or depressing. The median value of tourists’ perception of boring is around 0.1, and the data are roughly distributed between -0.5 and 0.3, indicating that tourists find many streets in Dali Old City boring.

2.3 Correlation Analysis of Spatial Characteristics Factors and Emotion Perception Factors

In order to explore the correlation between the street spatial characteristics factor and the emotion perception factors, a corresponding matrix was constructed to understand the correlation between the factors. The graph of the corresponding matrix is shown in Figure 8. The red dots represent positive correlation and the blue dots represent negative correlation. The larger the dots, the stronger the correlations. The cross indicates that there is no significant correlation between the variables. (p> 0.05).

Figure 8 Heat Map of Correlation between Street Environment Characteristics and Tourists’ Emotional Perception

As shown in Figure 8, most of the spatial characteristic factors show strong correlations with the emotional perception factors. Among them, GVI shows more significant positive correlation with Comfortable, Satisfied and Pleasant, and negative correlation with Anxious, Depressing and Boring. Dali Old City is located in a high-altitude area with strong ultraviolet rays, which makes people uncomfortable when exposed to sunlight for a long time. But a higher level of greening will greatly improve the comfort and satisfaction of tourists. Related studies have shown that greening can reduce anxiety and increase pleasure, which is consistent with the expectation. Sky Oppenness has a negative correlation with GVI, when the proportion of green area is higher, the proportion of sky area is lower, so in the correlation with tourists’ affective perception factors, Sky Oppenness shows the opposite result to GVI. Enclosedness shows weaker correlations with all emotion perception factors and doesn’t show significant correlations with the factors of Anxious, Depressing and Boring. Considering the fact that the relationship between Enclosedness and visitors’ perceptions is not a simple linear relationship, if Enclosedness is too low, the space lacks a sense of security, while if Enclosedness is too high, the space appears depressing and boring. Therefore, the appropriate level of Enclosedness needs to be further studied. There is a negative correlation between Street Order and Anxious, Depressing, and Boring. The streets of Dali Old City are narrow, most of them do not have dedicated sidewalks, and there are potential dangers in too many walking interferences such as motorized vehicles, battery operated vehicles, bicycles, etc., which increase tourists’ perception of anxious, and the chaotic street interface also make them feel depressing and boring. Street Order is positively correlated with Comfortable, Satisfied, and Pleasant, and it is found that some spacious streets with higher commercial value often have better street order, higher greening level and more comfortable street scale, which enhances tourists’ perception of comfort, satisfied and pleasant.

At the same time, the environmental characteristic factors of the four streets, i.e., GVI, Sky Oppenness, Enclosedness, Street Order, show strong correlation among them, in which GVI is negatively correlated with Sky Oppenness, and positively correlated with Enclosedness and Street Order. The six emotion perception characteristic factors show strong correlation with each other, with strong positive correlation with the positive emotion perception characteristic factors and negative correlation with the negative emotion perception characteristic factors respectively.

The characteristic importance evaluation based on the Random Forest regression model reflects the relative importance of 4 street environment characteristic factors, i.e., GVI, Sky Oppenness, Enclosedness and Street Order to 6 types of tourists’ emotion perception factors. The sum of 4 street environment characteristic factors is 100%. The importance scores are plotted as shown in Figure 9.

Figure 9 Importance Score Plot of the Effect of Street Environment Characteristic Factors on Emotion Perception

As shown in Figure 9, GVI is the most important influencing factor for Anxious, accounting for 40%, followed by Sky Oppenness, accounting for 31%, indicating that good green visibility can effectively alleviate the tourists’ anxiety. For Comfortable, Street Order accounts for the highest proportion of 33%, while GVI and Sky Oppenness account for 22% and 26% respectively, indicating that in addition to Street Order, GVI and Sky Oppenness are also more important for Comfortable. Depressing is more similar to Comfortable, and Street Order still has the highest importance at 38%, followed by Sky Oppenness at 23%. For Boring, Street Order has the highest percentage at 32%, followed by GVI at 31%, indicating that too little street order and too little GVI will make the whole street look boring and uninteresting. For Satisfied, Street Order has the highest percentage at 40%, indicating that Street Order significantly affects tourist satisfaction. For Pleasant, Sky Oppenness has the highest percentage at 40%, indicating that too much sky oppenness can lead to direct sunlight, which in turn affects the tourist pleasantness.

3. Discussion

This study explored the relationship between the physical dimension of streets and the psychological dimension of tourists, and it can be affirmed that the GVI has a positive effect on the emotional perception of tourists and has relatively significant importance. In order to improve the tourists’ experience, we suggest to increase the investment in street greening construction in Dali Old City, and increase the GVI by adding more trees, flowers and hedges on both sides of the streets, so as to create an ecological and pleasant walking environment. It is also worth noting that many streets in the Old City are narrow and do not have the conditions for continuous planting of street trees, resulting in a low GVI. We suggest that surfaces such as building facades, fences, railings, etc. be planted with climbing plants or installed plant walls, which can not only effectively increase the GVI, but also add a unique green landscape to the Old City. In addition, flower boxes, flower beds, or small landscaped areas can also be placed at appropriate locations on the streets to utilize the limited space for spot greening, which could not only beautify the environment but also improve the GVI.

Although some studies have shown that higher sky openness is more likely to have a positive effect, this study shows that too much sky openness has a negative effect on tourists’ emotional perceptions, which may be due to the fact that the study area is located at a high altitude with strong UV rays and dry climate, which causes high sky openness to induce hot and uncomfortable emotions in tourists, thus affecting their emotional perceptions of the street environment. Therefore, on the basis of maintaining the appearance of the Old City, the height of buildings and the width of streets should be rationally planned, while the amount of greenery or shading facilities should be increased as much as possible to prevent tourists from being directly hit by the sun.

Street order has a high importance in all the emotional perception factors, low street order is especially easy for tourists to produce the emotions of boring and depressing, thus seriously affecting the tourists’ comfort and satisfaction. We suggest to optimize the layout of sidewalks to ensure that sidewalks are spacious, continuous and barrier-free, and to do a good job of controlling electric vehicles and bicycles in some overly narrow and crowded streets to optimize the walking experience of tourists. Meanwhile, in the renewal design of the Old City, the protection of traditional buildings should be done well and a unified and clean architectural style should be maintained. Urban furniture and necessary service facilities should be added on both sides of the street to ensure the walking experience of tourists.

The results of the study show that the correlation between enclosedness and tourists’ emotional perception is weak, which may be due to the fact that the differences in street width and building density along the streets in Dali Old City are not very significant, and the building heights are also relatively uniform, resulting in a small difference in the enclosedness of different streets. However, the importance of enclosedness should not be ignored. The volunteers who participated in this study generally agreed that too high a street enclosedness tends to make people feel depressed, while too low a street enclosedness tends to make people feel uncomfortable and boring.

By optimizing the street space in Dali Old City, it can not only enhance the emotional experience of tourists, but also strengthen the tourist attraction of the Old City, and create a more livable and touristic environment for tourists and residents. This study used a machine learning method to determine the characteristics of the street environment, which proved the advantages of this method in terms of efficiency and accuracy. However, there are limitations to the study, such as the acquisition of streetscape images was easily affected by the season, such as the western section of Yu’er Road, which was in winter when the streetscape images were taken, and the bare and leafless tree branches affected the GVI measurement. Future research should be devoted to optimizing the algorithm and improving the accuracy and applicability of the model in order to better serve the sustainable development of greening in the Old City.

4. Conclusion

As an important part of urban planning and design, street environment characteristics have always been the focus of attention in both academic and practical fields. At the same time, the interactions and connections between these characteristics and tourists’ emotional perceptions have gradually become the key to understanding the quality of urban space and improving the tourism experience, the importance of which cannot be ignored. The conclusions of this study are as follows:

1. There is a certain correlation between street environment characteristics and tourists’ emotional perception. Specifically, the higher the GVI, the more positive the emotional perception of tourists; the lower the sky oppenness, the more positive the emotional perception of tourists; the higher the street order, the more positive the emotional perception of tourists. The correlation between enclosedness and tourists’ emotional perception is relatively weak.
2. Street order has the most significant effect on tourists’ emotional perception, especially in the emotional perception of comfortable, depressing, boring and satisfied, which indicates that improving the degree of street order is a key factor in improving the emotional experience of tourists in street environment design. In addition, the effects of GVI and sky openness on tourists’ emotional perception are also more obvious, especially in terms of anxious and pleasant. Therefore, in street environment design, greenery should be planted as much as possible to avoid direct sunlight to enhance the positive emotions of tourists.
3. There is a strong correlation between the factors of environmental characteristics and the factors of emotional perception, therefore, in the process of optimizing the street environment, it is necessary to comprehensively consider the interaction between various environmental characteristics in order to achieve the best optimization effect, while in improving the tourist experience, it is necessary to comprehensively consider the emotional perception of tourists.

4.There is a significant difference between the environmental quality and the emotional perception of tourists in different streets of Dali Old City, some of the streets that bear the main transportation function and have high commercial value, such as Yu’er Road and Ye’yu Road, have a better environmental quality, while Guangwu Road and Ping’e Road have a poorer environmental quality, which should be emphasized in the management and  renovation design of the Old City in order to improve the environmental quality of Dali Old City.

Bibliography:

[1]Wolf, K.L. Business District Streetscapes, Trees, and Consumer Response. J. For. 2005, 103, 396–400.

[2]Liu, Y.; Wang, R.; Grekousis, G.; Liu, Y.; Yuan, Y.; Li, Z. Neighbourhood greenness and mental wellbeing in Guangzhou, China: What are the pathways? Landsc. Urban Plan. 2019, 190, 103602.

[3]Yang, Y.; He, D.; Gou, Z.; Wang, R.; Liu, Y.; Lu, Y. Association between street greenery and walking behavior in older adults in Hong Kong. Sustain. Cities Soc. 2019, 51, 101747.

[4]Wang, R.; Helbich, M.; Yao, Y.; Zhang, J.; Liu, P.; Yuan, Y.; Liu, Y. Urban greenery and mental wellbeing in adults: Cross-sectional mediation analyses on multiple pathways across different greenery measures. Environ. Res. 2019, 176, 108535.

[5]Yue, Y.; Yang, D.; Van Dyck, D. Urban greenspace and mental health in Chinese older adults: Associations across different greenspace measures and mediating effects of environmental perceptions. Health Place 2022, 76, 102856.

[6]Ki, D.; Lee, S. Analyzing the effects of Green View Index of neighborhood streets on walking time using Google Street View and deep learning. Landsc. Urban Plan. 2021, 205, 103920.

[7]Li, X.; Zhang, C.; Li, W.; Ricard, R.; Meng, Q.; Zhang, W. Assessing street-level urban greenery using Google Street View and a modified green view index. Urban For. Urban Green. 2015, 14, 675–685.

[8]Yin, L.; Wang, Z. Measuring visual enclosure for street walkability: Using machine learning algorithms and Google Street View imagery. Appl. Geogr. 2016, 76, 147–153.

[9]Zhang, F.; Zhou, B.; Liu, L.; Liu, Y.; Fung, H.H.; Lin, H.; Ratti, C. Measuring human perceptions of a large-scale urban region using machine learning. Landsc. Urban Plan. 2018, 180, 148–160.

[10]Chen, X.; Meng, Q.; Hu, D.; Zhang, L.; Yang, J. Evaluating Greenery around Streets Using Baidu Panoramic Street View Images and the Panoramic Green View Index. Forests 2019, 10, 1109.

[11]Lin, T.-P.; Tsai, K.-T.; Hwang, R.-L.; Matzarakis, A. Quantification of the effect of thermal indices and sky view factor on park attendance. Landsc. Urban Plan. 2012, 107, 137–146.

[12]Li, X.; Santi, P.; Courtney, T.K.; Verma, S.K.; Ratti, C. Investigating the association between streetscapes and human walking activities using Google Street View and human trajectory data. Trans. GIS 2018, 22, 1029–1044.

[13]Baran, P.K.; Tabrizian, P.; Zhai, Y.; Smith, J.W.; Floyd, M.F. An exploratory study of perceived safety in a neighborhood park using immersive virtual environments. Urban For. Urban Green. 2018, 35, 72–81.

[14]Alcock, I.; White, M.P.; Wheeler, B.W.; Fleming, L.E.; Depledge, M.H. Longitudinal Effects on Mental Health of Moving to Greener and Less Green Urban Areas. Environ. Sci. Technol. 2014, 48, 1247–1255.

[15] Kondo, M.C.; Fluehr, J.; McKeon, T.; Branas, C. Urban Green Space and Its Impact on Human Health. Int. J. Environ. Res. Public Health 2018, 15, 445.

[16]Nguyen, P.-Y.; Astell-Burt, T.; Rahimi-Ardabili, H.; Feng, X. Green Space Quality and Health: A Systematic Review. Int. J. Environ. Res. Public Health 2021, 18, 11028.

[17]Barton, J.; Griffin, M.; Pretty, J. Exercise-, Nature- and Socially Interactive-Based Initiatives Improve Mood and Self-Esteem in the Clinical Population. Perspect. Public Health 2012, 132, 89–96.

[18]Engemann, K.; Pedersen, C.B.; Arge, L.; Tsirogiannis, C.; Mortensen, P.B.; Svenning, J.-C. Residential Green Space in Childhood Is Associated with Lower Risk of Psychiatric Disorders from Adolescence into Adulthood. Proc. Natl. Acad. Sci. USA 2019, 116, 5188–5193.

[19]Brooks, A.M.; Ottley, K.M.; Arbuthnott, K.D.; Sevigny, P. Nature-Related Mood Effects: Season and Type of Nature Contact. J. Environ. Psychol. 2017, 54, 91–102.

[20]Hunter, R.F.; Cleland, C.; Cleary, A.; Droomers, M.; Wheeler, B.W.; Sinnett, D.; Nieuwenhuijsen, M.J.; Braubach, M. Environmental, Health, Wellbeing, Social and Equity Effects of Urban Green Space Interventions: A Meta-Narrative Evidence Synthesis. Environ. Int. 2019, 130, 104923.

[21]Song, C.; Joung, D.; Ikei, H.; Igarashi, M.; Aga, M.; Park, B.-J.; Miwa, M.; Takagaki, M.; Miyazaki, Y. Physiological and Psychological Effects of Walking on Young Males in Urban Parks in Winter. J. Physiol. Anthr. 2013, 32, 18.

[22]Ambrey, C.L.; Fleming, C.M.; Manning, M. Greenspace and Life Satisfaction: The Moderating Role of Fear of Crime in the Neighbourhood. In Proceedings of the Australia New Zealand Society for Ecological Economics, Canberra, Australia, 11–14 November 2013; pp. 89–108.

[23]MacKerron, G.; Mourato, S. Happiness Is Greater in Natural Environments. Glob. Environ. Change 2013, 23, 992–1000.

[24]Ward Thompson, C.; Roe, J.; Aspinall, P. Woodland Improvements in Deprived Urban Communities: What Impact Do They Have on People’s Activities and Quality of Life? Landsc. Urban Plan. 2013, 118, 79–89.

[25]Navarrete-Hernandez and Laffan, 2019A greener urban environment: Designing green infrastructure interventions to promote citizens’ subjective wellbeing

Landscape and Urban Planning, 191 (2019), Article 103618, 10.1016

[26]Xu, J., Xiong, Q., Jing, Y., Xing, L., An, R., Tong, Z., Liu, Y., & Liu, Y. (2023). Understanding the nonlinear effects of the street canyon characteristics on human perceptions with street view images, Ecological Indicators, 154, 2023, 110756,

[27]Harvey, C.; Aultman-Hall, L.; Hurley, S.E.; Troy, A. Effects of skeletal streetscape design on perceived safety. Landsc. Urban Plan. 2015, 142, 18–28.

[28]Asgarzadeh, M.; Koga, T.; Hirate, K.; Farvid, M.; Lusk, A. Investigating oppressiveness and spaciousness in relation to building, trees, sky and ground surface: A study in Tokyo. Landsc. Urban Plan. 2014, 131, 36–41.

[29]Z. Cui, M. He, M. Lu An analysis of green view index in cold ｒegion city: a case study of Harbin J. Chin. Urban For., 16 (05) (2018), pp. 34-38

[30]M. Helbich, Y. Yao, Y. Liu, J. Zhang, P. Liu, R. Wang ,Using deep learning to examine street view green and blue spaces and their associations with geriatric depression in Beijing, China Environ. Int., 126 (2019), pp. 107-117