Assessing Urban Street Vitality Through Visual and Auditory Perception:

A case study of historic urban area in Guangzhou, China

Abstract

Urban streets serve as both transportation networks and public spaces, accommodating activity and travel demand. Evaluating street vitality is essential in assessing urban vitality. Existing studies usually employed indicators like population density through mobile phone signalling or app check-ins data to evaluate vitality intensity, but few have considered vitality quality from pedestrians' perspectives. Our study puts forward an approach to evaluate the vitality quality of urban streets based on the visual and auditory perception of pedestrians, using the historic urban area of Guangzhou, China, as a case. We collected multi-dimensional data, including street view images, videos, sound recordings, and user preference questionnaires. Streets were classified into nine types based on varying levels of vitality quality and intensity. Our primary finding revealed a spatial mismatch: areas with better street quality often have lower activity vitality, and vice versa. This suggests urban renewal should focus on improving both aspects. Different street types require targeted improvements, such as enhancing commercial atmosphere in tree-lined streets, increasing greenness in arcade streets, and managing noise levels on high-level motorized roads. These insights can provide valuable insights for the local government and urban planners to enhance urban street vitality.

Introduction

Urban vitality, which arises from the interaction between people and places according to Jane (1993), has become a crucial topic in contemporary urban studies. In recent years, thanks to the development of multi-source big data, numerous studies have measured urban vitality making use of data such as points of interest (POIs), mobile phone signalling, Wi-Fi access points, and app check-ins (Kang, Fan et al., 2021; Kim, 2018; Klimek, 2019; Li, Q., Cui et al., 2021; Lu, Shi et al., 2019; Pan, Zhou et al., 2021). These studies have focused on various urban areas ranging from large-scale urban agglomerations and city regions to small-scale neighborhoods, commercial centers, and residential areas (Chen, Z., Dong et al., 2022; Griffin, Nordstrom et al., 2016; Lotfata and Helbich, 2023; Wu, Ye et al., 2023; Yue, Chen et al., 2021).

Streets constitute a vital component of open spaces in cities, serving as transportation networks while also collaborating with adjacent plazas and public facilities to meet the activity needs of residents (Donais, Abi-Zeid et al., 2019; Lefebvre-Ropars, Negron-Poblete et al., 2021; Von Schönfeld and Bertolini, 2017). In historic urban area, due to the insufficiency of open spaces, streets accommodate most of the residents’ travel and activity demands. Therefore, the quality of street environment has a significant impact on people's living experiences.

Urban streets vary widely, encompassing high-level motorized roads that primarily serve motorized traffic, as well as slower streets where pedestrian activities, such as commercial pedestrian streets and community streets, take place (Subadyo, Tutuko et al., 2018; Zubaidi, Amalia et al., 2024). Pedestrian-oriented streets, characterized by reduced motor vehicle traffic and lower speeds, predominate in urban areas. These streets often have a higher density of POIs on both sides, indicating that citizens are more likely to engage in leisure activities in such spaces, leading to higher levels of commercial and social vitality. However, due to narrow street widths and limited open spaces, these areas are under the pressure of diverse traffic patterns and pedestrian movements, as well as vehicular access and parking demand, goods loading and unloading demand, and leisure activities like shopping and queuing (Chen, Jingxu, Li et al., 2018; Espina, Mori et al., 2018; Li, Q., Cui et al., 2021; Zacharias, 2003). The more vibrant the area, the more prominent the spatial demand conflicts, which can result in a decline in street quality and a worse sensory experience for citizens and tourists (Hong and Jeon, 2020; Wen, Kenworthy et al., 2020).

Street vitality has always been an essential part of evaluating urban vitality. Similar to researches on urban vitality, many studies used indicators such as population volume, POI density, smart-phone based tracking and so forth to measure street vitality intensity (Delclòs-Alió, Gutiérrez et al., 2019; Huang, Hu et al., 2023; Li, Q., Cui et al., 2021). Additionally, some scholars have employed deep learning methods to obtain pedestrian flow data from video cameras (Li, Y., Yabuki et al., 2022). After measuring street vitality, the driving factors of street vitality were also investigated. Existing studies explored the association between street vitality and other built environment attributes, including diversity and accessibility (Huang, Hu et al., 2023), urban spatial patterns (Yang, Li et al., 2023), urban forms (Zumelzu and Barrientos-Trinanes, 2019), and land use (Sung, Go et al., 2013). These studies have provided a theoretical basis for improving street vitality.

However, from the perspective of street users, high-vitality streets characterized by high pedestrian flow do not necessarily guarantee a good perceptual experience. Congested and noisy streets, for example, can lead to uncomfortable and unbearable experiences. To truly create favourable street vitality, it is essential to evaluate the street quality based on people’s sensory experiences. Previous research has focused primarily on objective street attributes (Li, Y., Yabuki et al., 2022; Lunecke and Mora, 2018; Sung, Go et al., 2013; Ye, Li et al., 2018), with little attention paid to pedestrians’ subjective perceptions. Some studies have utilized street view images to conduct visual assessments (Li, Y., Yabuki et al., 2022; Liu, Ma et al., 2022; Wang, M., He et al., 2022; Ye, Zeng et al., 2019). However, lacking human-centered data, these results cannot fully capture the perceptions of pedestrians (Eom and Suzuki, 2019). Furthermore, street evaluations often focused on single sensory dimensions, with few studies examining the visual and auditory experiences of users concurrently.

In the historic urban area with diverse activity demands and prominent spatial conflicts, high-density pedestrian flows, commercial transactions, and noise related to logistics processes can lead to uncomfortable sensory experiences for street users. Consequently, poor visual and auditory perception is often closely related to high levels of vitality. It is necessary to investigate the vitality-quality types in different kinds of streets and analyze the impact of different visual and auditory elements on people's perception. Additionally, it is worth studying whether people's choices of street activity spaces are influenced by multiple factors.

Our study takes the historic urban area in Guangzhou as an example, which is representative for diverse street spaces and vigorous commercial activities in Chinese megacities. Guangzhou’s historic urban area provides a valuable case study due to its high population density, rich cultural heritage, and variety of street types, making it an ideal location for examining the interplay between street vitality and quality.

We collected street view images, videos and sound recordings taken from the pedestrians’ perspective. Semantic segmentation and sound level measurement were employed to extract indicators of visual and auditory perception, including commercial atmosphere, greenness, enclosure, crowdedness and noise level. Furthermore, a questionnaire based on 1-minute video recordings was conducted to obtain subjective evaluations of visual and auditory experiences from the respondents. This allowed us to determine the weight of different evaluation indicators affecting people’s sensory experiences and to calculate quality scores. Meanwhile, pedestrian flow intensity, which can represent street vitality, was also measured. Based on the spatial distribution of the binary relationship between vitality intensity and quality within the study area, we proposed strategies to improve both vitality and quality, considering relevant planning and policy backgrounds.

Study Area

Guangzhou is situated in the southern region of China, on the downstream part of the Pearl River. This location is characterized by a subtropical monsoon climate with mild temperatures and prominent maritime climate features, which significantly influence the urban landscape. The climate shapes urban spatial forms such as buildings and streets, exemplified by covered walkways in the form of arcade streets designed to shield pedestrians from sunlight and rain.

As one of the first batch of national famous historic and cultural cities in China, Guangzhou has a long history of urban development. The formation of its street types dates back to the early 20th century. Historical remote sensing images (Figure 1) reveal the dense urban street texture of the historic urban area (Figure 2(a)) that remains to the present day. The narrow distance between buildings (Figure 2(b)) creates a very crowded living environment. At present, the average floor area ratio (FAR) of historic urban area exceeds 2.2 (Source: Guangzhou Land Use Survey), which is at a very high level compared with other megacities with a subtropical monsoon climate. Additionally, the population density within the historic urban area is also remarkably high. Covering an area of 20.39km², the average population density of Guangzhou’s historic urban area exceeds 60,000 people/km², with some blocks exceeding 200,000 people/km² (Source: Seventh National Population Census).

Given these characteristics, we have chosen to focus our research on the historic urban area of Guangzhou, which is a more precise and manageable scale for studying street quality and vitality. The high FAR and extremely high population density have led to an incredibly tight land use within the historic urban area, with very limited open space and insufficient per capita activity area. Consequently, streets serve as the most important public space for commuting and accommodating activities. The vitality and quality of these streets have a great impact on residents' life experiences. Therefore, it is very necessary and urgent to optimize the quality of street space to make full use of these vital urban areas.

Guangzhou is commercially developed and has strong regional influence. Businesses within the historic urban area include clothing, toys, seafood, electronics and jade. Specifically, most of these businesses are involved in wholesale, retail and catering services. Shops are densely distributed on both sides of the streets, leading to frequent conflicts between commodity transactions, goods distribution, and entertainment and leisure activities in the relatively narrow pedestrian spaces (Figure 2(c)). This situation can make citizens and tourists feel crowded, uncomfortable, and even unsafe. Moreover, high-vitality streets often have lower quality. Therefore, improving the quality of these streets is a top priority for the local government.

Figure 1. Remote sensing images of Guangzhou's historic urban area in 1960s, 1970s and 2020s. (Source: https://earthexplorer.usgs.gov/)

Figure 2. The current situation of the historic urban area in Guangzhou.

Based on the current situation analysis above, we select four types of streets in the historic urban area of Guangzhou: tree-lined streets, arcade streets, alleys, and high-level motorized roads. The distribution of 242 selected sampling points is depicted in Figure 3(a). Examples of these streets are shown in Figure 3(b), and their characteristics are described as follows:

(1) Arcade streets: In 1918, inspired by the construction of arcade streets in Hong Kong, Guangzhou began to promote the development of arcade streets. Initially, these streets were the most prosperous commercial areas in Guangzhou. In the historic protection plan, arcade buildings are regarded as a characteristic feature of the city, with their street form being immutable. However, due to early construction, the street width of arcade streets is narrow, with sidewalks generally between 2.8 and 3 meters wide and mostly about 4 meters high. Pedestrian walking space is inside the arcade, with the road on one side and shopfronts on the other. As arcade streets are directly connected to storefronts, they serve as crucial spaces for both commercial and transportation activities, making them the street type with the most serious spatial conflict in the historic urban area.

(2) Tree-lined streets: After the construction of arcade streets was banned, Guangzhou vigorously promoted tree-lined streets, forming a landscape of arcade and tree-lined streets. These modern streets feature trees planted between the sidewalk and motorway, with no building or construction covering the sidewalk. Tree-lined streets predominate in the historic urban area. Typically, the ground floors of buildings on both sides of these streets are commercial, while the upper floors are residential.

(3) Alleys: Most alleys are located in communities, with Mashi alleys being the most representative. The buildings along the alleys are mainly residential, with a few small restaurants interspersed.

(4) High-level motorized roads: These roads, which include urban arterial roads and flyovers, have either been newly constructed or expanded over the past two decades.

Figure 3. (a) Distribution of sampling points of different street types. (b) Examples of street view images of different street types from the pedestrian perspective

Traditional streets and alleys are fundamental components of the historic urban area, reflecting its traditional characteristics. Alongside modern roads, they constitute the backbone of the urban public space. The protection and development of Guangzhou's historic urban area place great importance on improving street vitality and quality. Based on an evaluation of the streets in this historic area, we discuss the relationship between the vitality and quality of different types of streets. The aim is to provide targeted strategies for enhancing the vitality and quality of these streets.

Materials and Methods

Approaches

The research approaches are presented in Figure 4. Firstly, the street quality of each sampling point was evaluated from the perspective of audio-visual dual perception, considering five evaluation indicators: commercial atmosphere, greenness, enclosure, crowdedness, and noise level. Secondly, combined with perceptual preferences obtained through video scoring, these five evaluation indicators were weighted and calculated to obtain the street quality score. Thirdly, the vitality intensity of each sampling point was determined by pedestrian volume from the field survey. Finally, the spatial segregation between street quality and vitality was explored and discussed, along with the influencing factors such as POIs, population, and commercial activities.

Figure 4. Research approaches.

Data collection and processing

Data sources

Street view images, day-time equivalent sound level, videos and pedestrian volume at 242 sampling points were obtained through field research conducted between 2:00 PM and 6:00 PM. This time period was chosen to capture a representative sample of daily street activities and conditions. The distribution of these sampling points can be seen in Figure 3(a). The survey process at each sampling point included taking photos from the pedestrian's perspective, counting the number of pedestrians passing by within a 1-minute time frame, monitoring the sound level using a sound level meter, and recording a 1-minute video. Additionally, POI data was obtained from Amap.

Street view images processing

Using a pre-trained model from the Model Zoo (Chen, L.-C., Papandreou et al., 2017), semantic segmentation was employed to process street view images. The model was based on the Cityscapes Dataset and trained using DeepLabV3, with a pixel accuracy (pixAcc) of 96.4 and a mean intersection over union (mIou) of 79.4. As shown in Figure 5, the segmentation results divided the street view images into 19 categories, including road, sidewalk, building, wall, fence, pole, traffic light, traffic sign, vegetation, terrain, sky, person, rider, car, truck, bus, train, motorcycle, and bicycle. The proportion of pixels for each category was also calculated.

Figure 5. Examples of semantic segmentation results for street view images from pedestrian perspective.

Sound level measurement

According to the People's Republic of China's "Environmental Quality Standard for Noise (GB 3096-2008)" (Environmental quality standard for noise: https://www.mee.gov.cn/ywgz/fgbz/bz/bzwb/wlhj/shjzlbz/200809/t20080917_128815.shtml), if the environment is affected by fixed noise sources, the measurement time can be set as 1 minute. The sound sources in Guangzhou’s historic urban area mainly include point sources such as living sounds, mechanical sounds, and natural sounds, as well as linear sources such as traffic sounds, which exhibit varying patterns. The day-time equivalent sound level was obtained by using a multifunctional sound level meter AWA6228 for 1 minute at each sampling point. This method provides precise and standardized sound level measurements, essential for objective acoustic environment assessment. The sound level meter accurately captures sound intensity, reflecting the diverse noise conditions across the historic urban landscape.

Natural Breaks classification method

The Natural Breaks classification method is a statistical technique used for grading and classifying numerical data (Arabameri, Saha et al., 2020; Chen, Jian, Yang et al., 2013). Its goal is to minimize within-group variance and maximize between-group variance. The method identifies class breaks that best group similar values together and highlight the differences between classes. For example, when assessing street vitality, the Natural Breaks classification method would categorize streets into low, medium, and high vitality groups based on their activity levels, ensuring that streets within each group are as similar as possible in terms of vitality. This method is particularly useful for mapping data values that are not evenly distributed. Therefore, in our study, the Natural Breaks classification method was used to classify five evaluation indicators and three evaluation dimensions.

Street environment evaluation

Street quality evaluation

Based on the semantic segmentation results, the element composition of each street view image was obtained. Referring to previous studies (Dong, Jiang et al., 2023; Wang, L., Han et al., 2022; Wang, Y., Qiu et al., 2023), greenness, enclosure and crowdedness were calculated. Commercial atmosphere was defined as the number of shopfront signs in use. Noise level was obtained through a sound level meter. The calculation formulas for the five evaluation indicators are as follows:

$$V_{\mathrm{\text{GR}}}^{\left(s\right)}=V{I}_{\mathrm{\text{tree}}}^{\left(s\right)}$$	(1)
$$V_{\mathrm{\text{EN}}}^{\left(s\right)}=(V{I}_{\mathrm{\text{building}}}^{\left(s\right)}+V{I}_{\mathrm{\text{wall}}}^{\left(s\right)}+V{I}_{\mathrm{\text{fence}}}^{\left(s\right)}+V{I}_{\mathrm{\text{vegetation}}}^{\left(s\right)})/(V{I}_{\mathrm{\text{road}}}^{\left(s\right)}+V{I}_{\mathrm{\text{sidewalk}}}^{\left(s\right)}+V{I}_{\mathrm{\text{terrain}}}^{\left(s\right)})$$	(2)
$$V_{\mathrm{\text{CR}}}^{\left(s\right)}=V{I}_{\mathrm{\text{person}}}^{\left(s\right)}+V{I}_{\mathrm{\text{rider}}}^{\left(s\right)}+V{I}_{\mathrm{\text{car}}}^{\left(s\right)}+V{I}_{\mathrm{\text{truck}}}^{\left(s\right)}+V{I}_{\mathrm{\text{bus}}}^{\left(s\right)}+V{I}_{\mathrm{\text{train}}}^{\left(s\right)}+V{I}_{\mathrm{\text{motorcycle}}}^{\left(s\right)}+V{I}_{\mathrm{\text{bicycle}}}^{\left(s\right)}$$	(3)
$$V_{\mathrm{\text{CA}}}^{\left(s\right)}=n_{sh\mathrm{\text{opsign}}}^{\left(s\right)}$$	(4)
$$V_{\mathrm{\text{NO}}}^{\left(s\right)}=L_{\mathrm{\text{eq}}}^{\left(s\right)}$$	(5)

Where $V{I}_i^{\left(s\right)}$ represents the $i$ -th view index in sampling point $s$ from the semantic segmentation. In this study, the view index included the proportion of tree, building, wall, fence, person, rider, car, etc. For example, $V{I}_{\mathrm{\text{tr}}\mathrm{\text{ee}}}^{\left(s\right)}$ represents the proportion of trees in sampling point $s$ .

$n_{sh\mathrm{\text{opsign}}}^{\left(s\right)}$ is the number of shop signs that can be identified.

$L_{\mathrm{\text{eq}}}^{\left(s\right)}$ is day-time equivalent sound level.

$V_m^{\left(s\right)}$ represents the $m$ -th evaluation indicator in sampling point $s$ . $V_{\mathrm{\text{GR}}}^{\left(s\right)}$ , $V_{\mathrm{\text{EN}}}^{\left(s\right)}$ , $V_{\mathrm{\text{CR}}}^{\left(s\right)}$ , $V_{\mathrm{\text{CA}}}^{\left(s\right)}$ , and $V_{\mathrm{\text{NO}}}^{\left(s\right)}$ represent the indices of greenness, enclosure, crowdedness, commercial atmosphere, and noise level, respectively.

The five evaluation indicators were further divided into five levels using the Natural Breaks classification method. Higher scores in greenness and commercial ambiance indicate better street quality, while higher scores in enclosure, crowdedness, and noise level indicate lower street quality.

The overall street quality of sampling point $s$ , $Q_k^{\left(s\right)}$ , is a weighted sum of all evaluation indicators and user preferences, as shown in eq. (6).

$$Q_k^{\left(s\right)}=\sum _m^{}{\beta }_{m|k}\cdot V_m^{\left(s\right)}={\beta }_{GR|k}\cdot V_{\mathrm{\text{GR}}}^{\left(s\right)}+{\beta }_{EN|k}\cdot V_{\mathrm{\text{EN}}}^{\left(s\right)}+{\beta }_{CR|k}\cdot V_{\mathrm{\text{CR}}}^{\left(s\right)}+{\beta }_{CA|k}\cdot V_{\mathrm{\text{CA}}}^{\left(s\right)}+{\beta }_{NO|k}\cdot V_{\mathrm{\text{NO}}}^{\left(s\right)}$$	(6)

Where the subscript $k$ represents the user group.

${\beta }_{m|k}$ represents the user preference of indicator $m$ by user group $k$ .

Therefore, the street vitality quality was evaluated by both visual and auditory indicators calculated from observed data, and user preferences. It implies that given the same street environment, the quality evaluation could vary among different socio-economic groups. Note that, ${\beta }_{m|k}$ is irrelevant with the sample point $s$ of vitality quality.

To determine the street users’ preference, ${\beta }_{m|k}$ , we designed a questionnaire based on respondents' scoring of street video samples. All videos were recorded at the same time as the street view images and sound level samples. For each question, respondents watched two random video samples and chose their preferred street for social activities based on simultaneous visual and auditory perception. Additionally, they selected the three most crucial indicators from the five evaluation indicators that influenced their choice. This binary choice procedure was repeated, resulting in a total of 309 binary choices made by the respondents. In this preliminary study, the respondent group consisted of several undergraduate students majoring in urban planning and landscape architecture. In the ongoing study, the respondents could extend to include more socio-economic groups to identify the possible preference disparities, such that decision makers could come up with targeted vitality improvement solutions in different neighbourhood.

Eventually, the weight, reflecting user preferences for vitality quality indicators, was estimated, as shown in Table 1. And the final street vitality quality $Q_k^{\left(s\right)}$ was calculated by eq. (6), and classified into six levels using Natural Breaks classification method, with level 1 representing the poorest street quality and level 6 representing the best street quality, as shown in Figure 6(a).

Table 1. The weight of five evaluation indicators.

Indicator name	Commercial atmosphere		Greenness	Enclosure	Crowdedness	Noise level
Weight		0.167	0.276	0.215	0.183	0.159

Street vitality evaluation

Based on the field survey of pedestrian volume, pedestrian density was calculated to represent street vitality intensity, combined with subjective perceptions of the size of open space. The vitality of each sampling point was then classified into six levels using the Natural Breaks classification method. As shown in Figure 6(b), level 1 represents the lowest street vitality and level 6 represents the highest street vitality.

Activity opportunity calculation

In ArcGIS, a 50m buffer zone was generated for each sampling point, and the number of public activity POIs (including healthcare services, sports and leisure services, daily services, scientific and cultural services, shopping services, public facilities, scenic spots, and catering services) within the buffer zone was counted. The numbers of POIs were then divided into six levels using Natural breaks classification method, with level 1 indicating the least activity opportunities and level 6 indicating the most activity opportunities on the streets, as shown in Figure 6(c).

Figure 6. Street quality, vitality and activity opportunity of each sampling point.

Evaluation of Street Quality and Vitality

Spatial segregation of street quality and vitality

Using the Inverse Distance Weighting (IDW) method in ArcGIS, raster maps of street quality and activity vitality scores (level 1-6) were generated and visualized in ArcScene, as shown in Figure 7. The darker the green, the higher the street quality, while the darker the red, the higher the street vitality. A clear spatial mismatch between street quality and vitality can be observed: regions with better street quality often have lower activity vitality, while regions with higher activity vitality tend to have lower street quality. This suggests that urban renewal efforts in Guangzhou’s historic urban area should focus on improving both street quality and activity vitality.

Figure 7. Raster maps of street quality and vitality scores.

The average scores of the five quality evaluation indicators and the three street evaluation dimensions were calculated separately for the four types of streets and are shown in Table 2 and Table 3, respectively.

Table 2. Average scores of the five quality evaluation indicators (scores range from 1 to 5).

Street type	Commercial atmosphere	Greenness	Enclosure	Crowdedness	Noise level
Tree-lined streets	2.63	3.47	1.77	2.62	3.37
Arcade streets	2.71	1.32	2.57	2.96	3.75
Alleys	2.13	1.57	2.78	2.35	2.74
High-level motorized roads	2.09	1.55	1.45	2.73	3.82
Average	2.57	2.70	2.04	2.68	3.42

Table 3. Average scores of the three street evaluation dimensions (scores range from 1 to 6).

Street type	Street quality	Street vitality	Activity opportunity
Tree-lined streets	4.28	3.32	2.50
Arcade streets	2.11	3.39	3.23
Alleys	2.52	2.13	2.30
High-level motorized roads	2.64	3.64	1.82
Average	3.53	3.24	2.62

In terms of the five quality evaluation indicators, arcade streets have the strongest average commercial atmosphere, while tree-lined streets have a slightly weaker commercial presence. Tree-lined streets exhibit the highest average greenness, whereas arcade streets have the lowest. Due to their narrow widths, alleys have the highest degree of enclosure, followed by arcade streets, which also have a relatively limited field of view. There is a positive correlation between crowdedness and noise level: arcade streets, bustling with commercial activity, are noisier and more crowded than tree-lined streets. High-level motorized roads have the highest noise levels due to heavy traffic volumes.

Table 4. The number of quality-vitality types for the four types of streets.

Street type	Low quality			Medium quality			High quality
	Low vitality	Medium vitality	High vitality	Low vitality	Medium vitality	High vitality	Low vitality	Medium vitality	High vitality
Tree-lined streets	6	3	5	16	38	11	22	39	12
Arcade streets	11	20	9	6	4	4	1	0	1
Alleys	9	3	0	7	1	2	1	0	0
High-level motorized roads	0	3	2	2	3	1	0	0	0

Table 5. The proportion of quality-vitality types for the four types of streets.

Street type	Low quality			Medium quality			High quality
	Low vitality	Medium vitality	High vitality	Low vitality	Medium vitality	High vitality	Low vitality	Medium vitality	High vitality
Tree-lined streets	3.95%	1.97%	3.29%	10.53%	25.00%	7.24%	14.47%	25.66%	7.89%
Arcade streets	19.64%	35.71%	16.07%	10.71%	7.14%	7.14%	1.79%	0.00%	1.79%
Alleys	39.13%	13.04%	0.00%	30.43%	4.35%	8.70%	4.35%	0.00%	0.00%
High-level motorized roads	0.00%	27.27%	18.18%	18.18%	27.27%	9.09%	0.00%	0.00%	0.00%

In terms of the comprehensive street quality evaluation, tree-lined streets have the best quality, while the quality of arcade streets is significantly affected by the presence of wholesale markets and other commercial activities. Arcade streets offer the most abundant activity space, while tree-lined streets, despite having less activity space, achieve a similar activity score.

According to the redefined six-level quality and vitality scores (1-2 as low, 3-4 as medium, and 5-6 as high), nine types of streets can be obtained: low quality-low vitality (Q^LV^L)，low quality-medium vitality (Q^LV^M), low quality-high vitality (Q^LV^H), medium quality-low vitality (Q^MV^L)，medium quality-medium vitality (Q^MV^M), medium quality-high vitality (Q^MV^H), high quality-low vitality (Q^HV^L)，high quality-medium vitality (Q^HV^M), high quality-high vitality (Q^HV^H). The number and proportion of quality-vitality types for each type of street were calculated.

As shown in Table 5, arcade streets have the highest proportion of low quality-medium vitality type (Figure 8), while tree-lined streets have the highest proportion of high quality-medium vitality type (Figure 9). Alleys mostly have low/medium quality and low vitality, and high-level motorized roads have the highest proportion of low/medium quality-medium vitality type.

Figure 8. The distribution of quality-vitality types for arcade streets.

Figure 9. The distribution of quality-vitality types for tree-lined streets.

Based on Figure 8, apart from Enning Road, which has received adequate revitalization efforts from the government, the vitality intensity and quality of other arcade streets are generally low. This is associated with the presence of vacant buildings and shops in most arcade streets. Conversely, Enning Road, which exhibits high vitality intensity, has poor quality in most sampling points due to the influx of tourists attracted by its tourism-related businesses, resulting in crowded and noisy street environments. Figure 9 displays the distribution of vitality-quality types in tree-lined streets, revealing significant variations among different blocks. Notably, low-quality tree-lined streets are predominantly concentrated around Enning Road. Furthermore, many tree-lined streets with better quality exhibit lower vitality intensity, which can be attributed to their relatively remote locations and insufficient attractiveness in terms of landscape features.

Sample analysis

By analyzing the road segments with different characteristics at selected sampling points for the four types of streets, we aim to identify the factors that influence street quality and vitality.

Tree-lined streets

Binjiang West Road was selected as an example (Figure 10). Binjiang West Road features higher greenness and a wide view due to its proximity to the Pearl River. Additionally, the road is wide and includes a running lane that separates pedestrian flow. A comparison of different sampling points on Binjiang West Road reveals that the discontinuation of roadside tree planting around sampling point 101 results in decreased greenness and increased noise levels, leading to a lower street quality compared to sampling point 98.

Figure 10. Sample analysis of tree-lined streets.

Note: Evaluation indicators marked with an asterisk (*) indicate that higher scores correspond to lower quality, while otherwise higher scores indicate better quality.

As shown in Table 1, greenness and enclosure are the most critical evaluation indicators for the perceived street quality. From this perspective, tree-lined streets have a natural advantage. And waterfront roads with lower surrounding development intensity perform better in these two indicators.

Binjiang West Road is relatively far from the central area, resulting in a weak commercial atmosphere. To improve the quality and vitality of tree-lined streets like Binjiang West Road, it may be appropriate to add small commercial facilities or attractive urban furniture and landscape features at key nodes. Enhancing the basic infrastructure of Binjiang Greenway to meet people's basic demands is also beneficial. These improvements can encourage more citizens to participate in outdoor activities, thus strengthening street vitality. Additionally, incorporating elements such as shaded seating areas, pedestrian-friendly lighting, and interactive public art or instruments can further enhance the appeal and usability of the street, making it a more vibrant and engaging space for the community.

Arcade streets

Enning Road, also known as "the most beautiful arcade street," is the oldest and best-preserved historical and cultural district in Guangzhou, located at the core of the Xiguan cultural area. As shown in Figure 11, Enning Road has a high level of pedestrian density and overall vitality, but its quality is relatively low, although some sections have higher quality. Sampling point 17, with lower quality, and sampling point 167, with higher quality, were selected for comparison. Sampling point 167 is located at the end of Enning Road, where unobstructed road greenery and arcade facades can be seen. Although the commercial atmosphere is weakened by the closure of shops, the increase in greenness and enclosure, as well as the reduction in crowdedness and noise levels, have significantly improved the quality.

Against the dual background of preserving the historical charm of arcade streets and promoting urban vitality and economic development, the government has renovated the buildings on both sides of the Enning Road, and infused them with a multitude of commercial, retail and entertainment functions. Through publicity efforts, large numbers of residents and tourists have been attracted to arcade streets, resulting in increased traffic flow of motor vehicles and crowded queues of people on the streets, making the narrow arcade streets appear even more congested and noisy. Improving the quality of arcade streets should focus on increasing greenery and reducing logistics and traffic noise. Planning decisions should also aim to improve the commercial format and maintain its vitality as a business center. Enhancing greenery, implementing noise-reduction measures, encouraging diverse businesses, expanding pedestrian-only zones, installing public amenities, organizing cultural activities, and improving signage can balance historical preservation with modern economic and social functions, ultimately improving the quality and vitality of arcade streets.

Figure 11. Sample analysis of arcade streets.

Note: Evaluation indicators marked with an asterisk (*) indicate that higher scores correspond to lower quality, while otherwise higher scores indicate better quality.

Alleys

Taking a typical alley in the Xiguan Mansion community as an example, the vitality of the alley is low due to its poor accessibility. However, the quality of sampling point 282 is better than that of sampling point 254. Based on the five evaluation indicators in Figure 12, sampling point 282 has higher greenness and a larger depth of field, whereas sampling point 254 is more crowded and has many disorderly obstructions that negatively affect pedestrians’ sight. Typically, alleys are internal roads of old residential buildings, and the primary task of urban renewal is to improve the environment by increasing greenery, parking spaces for non-motor vehicles, drying spaces, and other amenities.

The number of POIs for public activities was calculated to characterize the activity opportunities of the two sampling points. Within the 50-meter buffer of sampling point 254, there are 18 POIs mainly related to catering and shopping, while within the 50-meter buffer of sampling point 282, there is only one POI, namely, the Duobao Street Comprehensive Elderly Service Center. Although there are more activity opportunities around sampling point 254, both sampling points have similarly low vitality. Despite being near storefronts on Longjin West Road, the commercial business types in the area are relatively traditional, and the commercial vitality does not penetrate into the alleys. To improve the quality and vitality of alleys in historical urban areas, the alleys within residential zones can supplement the business types lacking in traditional commercial streets. Enhancing vitality and quality can be achieved by incorporating functions such as cultural displays, artistic design, health and leisure facilities, and characteristic homestays. Additionally, increasing greenery, creating more parking spaces for non-motor vehicles, and adding drying spaces can further enhance the environmental quality of these alleys.

Figure 12. Sample analysis of alleys.

Note: Evaluation indicators marked with an asterisk (*) indicate that higher scores correspond to lower quality, while otherwise higher scores indicate better quality.

High-level motorized roads

Taking Huangsha Road as a typical high-level motorized road, as shown in Figure 13, these roads generally exhibit poor quality due to the absence of commercial activities, low greenness, and high noise levels. Flyovers above the road contribute significantly to the noise. Comparing the two sampling points, 40 and 139, it is found that the quality of sampling point 40 is slightly better than that of sampling point 139, as the latter is located near a subway station with denser pedestrian and vehicular traffic. For high-level motorized roads, planting trees along the road may be the most effective way to improve their quality. Trees can create a natural barrier between people and vehicles, increase greenness, and reduce noise levels. Moreover, incorporating sound barriers, enhancing pedestrian pathways, adding aesthetic elements such as murals or green walls, and introducing commercial activities such as food trucks or pop-up markets can further improve the quality and vitality of these roads.

Figure 13. Sample analysis of high-level motorized roads.

Note: Evaluation indicators marked with an asterisk (*) indicate that higher scores correspond to lower quality, while otherwise higher scores indicate better quality.

Discussion and Conclusion

This study aims to assess the quality of street vitality from the perspective of pedestrians' visual and auditory perception. We intergrated street view images, POIs, on-site data collection, and stated preference questionnaires, utilizing semantic segmentation and expert evaluation to comprehensively evaluate street vitality and quality. Our finding indicates a spatial mismatch: areas with higher street quality often exhibit lower activity vitality, and vice versa. This suggests that urban renewal efforts should aim to enhance both aspects simultaneously. The approaches proposed in this paper, based on pedestrians’ visual and auditory perception evaluation, can also be used for quantifying street vitality in other regions, especially historic urban area.

Urban streets in Guangzhou's historic urban area have high preservation value. From the results of the vitality and quality quantitative evaluation in our study, different types of streets require improvement from different perspectives. Arcade streets and tree-lined streets are the most representative and valuable types. Arcade streets, as important historical commercial streets, have strong vitality but urgently need quality improvements. Since the arcade buildings cannot be reconstructed, improvements can be made by increasing greenery and upgrading commercial activities. Transforming existing wholesale markets into cultural displays and leisure activities can enhance the overall visual appearance and significantly reduce logistics noise, thus optimizing residents’ auditory perception. In contrast, although some tree-lined streets are of high quality, they attract low activity intensity. Street vitality can be improved by transferring some central area functions to the sides of the tree-lined streets.

Future research can expand to include more sampling points on additional streets. Furthermore, obtaining perceptions and preferences from a wider range of social groups through questionnaires will provide decision-makers with targeted evidence for addressing real-world problems.

Author Contributions

Conceptualization, Y.X. and X.M.; methodology, Y.X. and X.M.; software, Y.X.; investigation, X.M.; resources, X.M.; data curation, Y.X.; writing—original draft preparation, Y.X.; writing—review and editing, X.M.; supervision, X.M. All authors have read and agreed to the published version of the manuscript.

Ethics Declaration

The authors declare that they have no conflicts of interest regarding the publication of the paper.

Acknowledgments

We would like to express our gratitude to Xuanxuan Liu, Ziying Qi, Yang Gao and Botao Zhao for their assistance in collecting the data used in this study.

References

•  Arabameri,  A.,  Saha,  S., et al. (2020). "Landslide Susceptibility Evaluation and Management Using Different Machine Learning Methods in the Gallicash River Watershed, Iran". Remote Sensing, 12(3), 475. doi: http://doi.org/10.3390/rs12030475.
•  Chen,  J.,  Li,  Z., et al. (2018). "Evaluating Bicycle–Vehicle Conflicts and Delays on Urban Streets with Bike Lane and on-Street Parking". Transportation letters, 10(1), 1-11. doi: http://doi.org/10.1080/19427867.2016.1207365.
•  Chen,  J.,  Yang,  S., et al. (2013). "Research on Geographical Environment Unit Division Based on the Method of Natural Breaks (Jenks)". The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 40, 47-50. doi: http://doi.org/10.5194/isprsarchives-XL-4-W3-47-2013.
•  Chen,  L.-C.,  Papandreou,  G., et al. (2017). "Rethinking Atrous Convolution for Semantic Image Segmentation". arXiv preprint arXiv:.05587.
•  Chen,  Z.,  Dong,  B., et al. (2022). "The Impacts of Urban Vitality and Urban Density on Innovation: Evidence from China's Greater Bay Area". Habitat International, 119, 102490. doi: http://doi.org/10.1016/j.habitatint.2021.102490.
•  Delclòs-Alió,  X.,  Gutiérrez,  A., et al. (2019). "The Urban Vitality Conditions of Jane Jacobs in Barcelona: Residential and Smartphone-Based Tracking Measurements of the Built Environment in a Mediterranean Metropolis". Cities, 86, 220-228. doi: http://doi.org/10.1016/j.cities.2018.09.021.
•  Donais , F. M.,  Abi-Zeid,  I., et al. (2019). "Assessing and Ranking the Potential of a Street to Be Redesigned as a Complete Street: A Multi-Criteria Decision Aiding Approach". Transportation research part A: policy and practice, 124, 1-19. doi: http://doi.org/10.1016/j.tra.2019.02.006.
•  Dong,  L.,  Jiang,  H., et al. (2023). "Assessing Impacts of Objective Features and Subjective Perceptions of Street Environment on Running Amount: A Case Study of Boston". Landscape and urban planning, 235, 104756. doi: https://doi.org/10.1016/j.landurbplan.2023.104756.
•  Eom,  S. and  Suzuki,  T. (2019). "Spatial Distribution of Pedestrian Space in Central Tokyo Regarding Building, Public Transportation and Urban Renewal Projects". International Review for Spatial Planning and Sustainable Development, 7(2), 108-124. doi: http://doi.org/10.14246/irspsda.7.2_108.
• Espina, J. M., Mori, S., et al. (2018). "An Analysis of Environment Behavior Relationships Towards the Design of a Local Mixed-Used Street: Based on Behavior Settings of Belgium Street in Cebu City, Philippines". Sustainability, 10(9), 3230. doi: http://doi.org/10.3390/su10093230.
•  Griffin,  M.,  Nordstrom , B. W., et al. (2016). "A Case Study: Analyzing City Vitality with Four Pillars of Activity—Live, Work, Shop, and Play". Big Data, 4(1), 60-66. doi: http://doi.org/10.1089/big.2015.0043.
• Hong, J. Y. and Jeon, J. Y. (2020). "Comparing Associations among Sound Sources, Human Behaviors, and Soundscapes on Central Business and Commercial Streets in Seoul, Korea". Building and Environment, 186, 107327. doi: http://doi.org/10.1016/j.buildenv.2020.107327.
•  Huang,  J.,  Hu,  X., et al. (2023). "How Diversity and Accessibility Affect Street Vitality in Historic Districts?". Land, 12(1), 219. doi: http://doi.org/10.3390/land12010219.
•  Jane,  J. (1993). "The Death and Life of Great American Cities". Vintage Books.
•  Kang,  C.,  Fan,  D., et al. (2021). "Validating Activity, Time, and Space Diversity as Essential Components of Urban Vitality". Environment and Planning B: Urban Analytics and City Science, 48(5), 1180-1197. doi: http://doi.org/10.1177/2399808320919771.
•  Kim,  Y.-L. (2018). "Seoul's Wi-Fi Hotspots: Wi-Fi Access Points as an Indicator of Urban Vitality". Computers, Environment and Urban Systems, 72, 13-24. doi: http://doi.org/10.1016/j.compenvurbsys.2018.06.004.
•  Klimek,  R. (2019). "Towards Recognising Individual Behaviours from Pervasive Mobile Datasets in Urban Spaces". Sustainability, 11(6), 1563. doi: http://doi.org/10.3390/su11061563.
•  Lefebvre-Ropars,  G.,  Negron-Poblete,  P., et al. (2021). "A Needs-Gap Analysis of Street Space Allocation". Journal of transport and land use, 14(1), 151-170. doi: http://doi.org/10.5198/jtlu.2021.1808.
•  Li,  Q.,  Cui,  C., et al. (2021). "Multidimensional Urban Vitality on Streets: Spatial Patterns and Influence Factor Identification Using Multisource Urban Data". ISPRS International Journal of Geo-Information, 11(1), 2. doi: http://doi.org/10.3390/ijgi11010002.
•  Li,  Y.,  Yabuki,  N., et al. (2022). "Exploring the Association between Street Built Environment and Street Vitality Using Deep Learning Methods". Sustainable Cities and Society, 79, 103656. doi: http://doi.org/10.1016/j.scs.2021.103656.
•  Liu,  Z.,  Ma,  X., et al. (2022). "Information in Streetscapes—Research on Visual Perception Information Quantity of Street Space Based on Information Entropy and Machine Learning". ISPRS International Journal of Geo-Information, 11(12), 628. doi: http://doi.org/10.3390/ijgi11120628.
•  Lotfata,  A. and  Helbich,  M. (2023). "Spatial Analysis of Neighborhood Vitality Determinants on Physical Activity: A Case Study of Chicago". GeoJournal, 88(2), 2187-2197. doi: http://doi.org/10.1007/s10708-022-10748-8.
•  Lu,  S.,  Shi,  C., et al. (2019). "Impacts of Built Environment on Urban Vitality: Regression Analyses of Beijing and Chengdu, China". International journal of environmental research and public health, 16(23), 4592. doi: http://doi.org/10.3390/ijerph16234592.
•  Lunecke , M. G. H. and  Mora,  R. (2018). "The Layered City: Pedestrian Networks in Downtown Santiago and Their Impact on Urban Vitality". Journal of Urban Design, 23(3), 336-353. doi: http://doi.org/10.1080/13574809.2017.1369869.
•  Pan,  C.,  Zhou,  J., et al. (2021). "Impact of Check‐in Data on Urban Vitality in the Macao Peninsula". Scientific Programming, 2021(1), 7179965. doi: http://doi.org/10.1155/2021/7179965.
•  Subadyo , A. T.,  Tutuko,  P., et al. (2018). "Assessment of Inclusive Historical Public Spaces in Achieving Preservation of Such Areas in Malang, Indonesia Case Study: Public Spaces Developed During the Dutch Colonial Period". International Review for Spatial Planning and Sustainable Development, 6(4), 76-92. doi: http://doi.org/10.14246/irspsd.6.4_76.
•  Sung,  H.-G.,  Go,  D.-H., et al. (2013). "Evidence of Jacobs’s Street Life in the Great Seoul City: Identifying the Association of Physical Environment with Walking Activity on Streets". Cities, 35, 164-173. doi: http://doi.org/10.1016/j.cities.2013.07.010.
•  Von  Schönfeld ,  K.  C. and Bertolini L. (2017). "Urban Streets: Epitomes of Planning Challenges and Opportunities at the Interface of Public Space and Mobility". Cities, 68, 48-55. doi: http://doi.org/10.1016/j.cities.2017.04.012.
•  Wang,  L.,  Han,  X., et al. (2022). "Measuring Residents’ Perceptions of City Streets to Inform Better Street Planning through Deep Learning and Space Syntax". ISPRS Journal of Photogrammetry and Remote Sensing, 190, 215-230. doi: http://doi.org/10.1016/j.isprsjprs.2022.06.011.
•  Wang,  M.,  He,  Y., et al. (2022). "Assessing Street Space Quality Using Street View Imagery and Function-Driven Method: The Case of Xiamen, China". ISPRS International Journal of Geo-Information, 11(5), 282. doi: http://doi.org/10.3390/ijgi11050282.
•  Wang,  Y.,  Qiu,  W., et al. (2023). "Drivers or Pedestrians, Whose Dynamic Perceptions Are More Effective to Explain Street Vitality? A Case Study in Guangzhou". Remote Sensing, 15(3), 568. doi: http://doi.org/10.3390/rs15030568.
•  Wen,  L.,  Kenworthy,  J., et al. (2020). "Higher Density Environments and the Critical Role of City Streets as Public Open Spaces". Sustainability, 12(21), 8896. doi: http://doi.org/10.3390/su12218896.
•  Wu,  C.,  Ye,  Y., et al. (2023). "Using Street View Images to Examine the Association between Human Perceptions of Locale and Urban Vitality in Shenzhen, China". Sustainable Cities and Society, 88, 104291. doi: http://doi.org/10.1016/j.scs.2022.104291.
•  Yang,  J.,  Li,  X., et al. (2023). "Exploring the Relationship between Urban Street Spatial Patterns and Street Vitality: A Case Study of Guiyang, China". International journal of environmental research and public health, 20(2), 1646. doi: http://doi.org/10.3390/ijerph20021646.
•  Ye, Y Y.,  Li,  D., et al. (2018). "How Block Density and Typology Affect Urban Vitality: An Exploratory Analysis in Shenzhen, China". Urban Geography, 39(4), 631-652. doi: http://doi.org/10.1080/02723638.2017.1381536.
•  Ye,  Y.,  Zeng,  W., et al. (2019). "The Visual Quality of Streets: A Human-Centred Continuous Measurement Based on Machine Learning Algorithms and Street View Images". Environment and Planning B: Urban Analytics and City Science, 46(8), 1439-1457. doi: http://doi.org/10.1177/2399808319828734.
•  Yue,  W.,  Chen,  Y., et al. (2021). "Identifying Urban Vitality in Metropolitan Areas of Developing Countries from a Comparative Perspective: Ho Chi Minh City Versus Shanghai". Sustainable Cities and Society, 65, 102609. doi: http://doi.org/10.1016/j.scs.2020.102609.
•  Zacharias,  J. (2003). "The Search for Sustainable Transport in a Developing City: The Case of Tianjin". International Development Planning Review, 25(3), 283-299. doi: http://doi.org/DOI10.3828/idpr.25.3.4.
•  Zubaidi,  F.,  Amalia,  N., et al. (2024). "Developing the Public Spaces Alternative in Palu City". International Review for Spatial Planning and Sustainable Development, 12(2), 201-220. doi: http://doi.org/10.14246/irspsd.12.2_201.
•  Zumelzu,  A. and  Barrientos-Trinanes,  M. (2019). "Analysis of the Effects of Urban Form on Neighborhood Vitality: Five Cases in Valdivia, Southern Chile". Journal of Housing and the Built Environment, 34(3), 897-925. doi: http://doi.org/10.1007/s10901-019-09694-8.