Objective scoring of streetscape walkability related to leisure walking: Statistical modeling approach with semantic segmentation of Google Street View images

Street-level built environment on SARS-CoV-2 transmission: A study of Hong Kong

Understanding the association between SARS-CoV-2 Spatial Transmission Risk (SSTR) and Built Environments (BE) is crucial for implementing effective pandemic prevention measures. Massive efforts have been made to examine the macro-built environment at the regional level, which has neglected the living service areas at the residential scale. Therefore, this study aims to explore the association between Street-level Built Environments (SLBE) and SSTR in Hong Kong from the 1st to the early 5th waves of the pandemic to address this gap. A total of 3693 visited/resided buildings were collected and clustered by spatial autocorrelation, and then Google Street View (GSV) was employed to obtain SLBE features around the buildings. Eventually, the interpretable machine learning framework based on the random forest algorithm (RFA)-based SHapley Additive exPlanations (SHAP) model was proposed to reveal the hidden non-linear association between SSTR and SLBE. The results indicated that in the high-risk cluster area, street sidewalks, street sanitation facilities, and artificial structures were the primary risk factors positively associated with SSTR, in low-risk cluster areas with a significant positive association with traffic control facilities. Our study elucidates the role of SLBE in COVID-19 transmission, facilitates strategic resource allocation, and guides the optimization of outdoor behavior during pandemics for urban policymakers.

Assessing the nonlinear impact of green space exposure on psychological stress perception using machine learning and street view images

Introduction

Urban green space (GS) exposure is recognized as a nature-based strategy for addressing urban challenges. However, the stress relieving effects and mechanisms of GS exposure are yet to be fully explored. The development of machine learning and street view images offers a method for large-scale measurement and precise empirical analysis.

Methods

This study focuses on the central area of Shanghai, examining the complex effects of GS exposure on psychological stress perception. By constructing a multidimensional psychological stress perception scale and integrating machine learning algorithms with extensive street view images data, we successfully developed a framework for measuring urban stress perception. Using the scores from the psychological stress perception scale provided by volunteers as labeled data, we predicted the psychological stress perception in Shanghai's central urban area through the Support Vector Machine (SVM) algorithm. Additionally, this study employed the interpretable machine learning model eXtreme Gradient Boosting (XGBoost) algorithm to reveal the nonlinear relationship between GS exposure and residents' psychological stress.

Results

Results indicate that the GS exposure in central Shanghai is generally low, with significant spatial heterogeneity. GS exposure has a positive impact on reducing residents' psychological stress. However, this effect has a threshold; when GS exposure exceeds 0.35, its impact on stress perception gradually diminishes.

Discussion

We recommend combining the threshold of stress perception with GS exposure to identify urban spaces, thereby guiding precise strategies for enhancing GS. This research not only demonstrates the complex mitigating effect of GS exposure on psychological stress perception but also emphasizes the importance of considering the "dose-effect" of it in urban planning and construction. Based on open-source data, the framework and methods developed in this study have the potential to be applied in different urban environments, thus providing more comprehensive support for future urban planning.

Portable Arduino-Based Multi-Sensor Device (SBEDAD): Measuring the Built Environment in Street Cycling Spaces

The built environment's impact on human activities has been a hot issue in urban research. Compared to motorized spaces, the built environment of pedestrian and cycling street spaces dramatically influences people's travel experience and travel mode choice. The streets' built environment data play a vital role in urban design and management. However, the multi-source, heterogeneous, and massive data acquisition methods and tools for the built environment have become obstacles for urban design and management. To better realize the data acquisition and for deeper understanding of the urban built environment, this study develops a new portable, low-cost Arduino-based multi-sensor array integrated into a single portable unit for built environment measurements of street cycling spaces. The system consists of five sensors and an Arduino Mega board, aimed at measuring the characteristics of the street cycling space. It takes air quality, human sensation, road quality, and greenery as the detection objects. An integrated particulate matter laser sensor, a light intensity sensor, a temperature and humidity sensor, noise sensors, and an 8K panoramic camera are used for multi-source data acquisition in the street. The device has a mobile power supply display and a secure digital card to improve its portability. The study took Beijing as a sample case. A total of 127.97 G of video data and 4794 Kb of txt records were acquired in 36 working hours using the street built environment data acquisition device. The efficiency rose to 8474.21% compared to last year. As an alternative to conventional hardware used for this similar purpose, the device avoids the need to carry multiple types and models of sensing devices, making it possible to target multi-sensor data-based street built environment research. Second, the device's power and storage capabilities make it portable, independent, and scalable, accelerating self-motivated development. Third, it dramatically reduces the cost. The device provides a methodological and technological basis for conceptualizing new research scenarios and potential applications.

Artificial intelligence-based assessment of built environment from Google Street View and coronary artery disease prevalence

BACKGROUND AND AIMS

Built environment plays an important role in the development of cardiovascular disease. Tools to evaluate the built environment using machine vision and informatic approaches have been limited. This study aimed to investigate the association between machine vision-based built environment and prevalence of cardiometabolic disease in US cities.

METHODS

This cross-sectional study used features extracted from Google Street View (GSV) images to measure the built environment and link them with prevalence of coronary heart disease (CHD). Convolutional neural networks, linear mixed-effects models, and activation maps were utilized to predict health outcomes and identify feature associations with CHD at the census tract level. The study obtained 0.53 million GSV images covering 789 census tracts in seven US cities (Cleveland, OH; Fremont, CA; Kansas City, MO; Detroit, MI; Bellevue, WA; Brownsville, TX; and Denver, CO).

RESULTS

Built environment features extracted from GSV using deep learning predicted 63% of the census tract variation in CHD prevalence. The addition of GSV features improved a model that only included census tract-level age, sex, race, income, and education or composite indices of social determinant of health. Activation maps from the features revealed a set of neighbourhood features represented by buildings and roads associated with CHD prevalence.

CONCLUSIONS

In this cross-sectional study, the prevalence of CHD was associated with built environment factors derived from GSV through deep learning analysis, independent of census tract demographics. Machine vision-enabled assessment of the built environment could potentially offer a more precise approach to identify at-risk neighbourhoods, thereby providing an efficient avenue to address and reduce cardiovascular health disparities in urban environments.

Assessing the Impact of Urban Environments on Mental Health and Perception Using Deep Learning: A Review and Text Mining Analysis

Understanding how outdoor environments affect mental health outcomes is vital in today's fast-paced and urbanized society. Recently, advancements in data-gathering technologies and deep learning have facilitated the study of the relationship between the outdoor environment and human perception. In a systematic review, we investigate how deep learning techniques can shed light on a better understanding of the influence of outdoor environments on human perceptions and emotions, with an emphasis on mental health outcomes. We have systematically reviewed 40 articles published in SCOPUS and the Web of Science databases which were the published papers between 2016 and 2023. The study presents and utilizes a novel topic modeling method to identify coherent keywords. By extracting the top words of each research topic, and identifying the current topics, we indicate that current studies are classified into three areas. The first topic was "Urban Perception and Environmental Factors" where the studies aimed to evaluate perceptions and mental health outcomes. Within this topic, the studies were divided based on human emotions, mood, stress, and urban features impacts. The second topic was titled "Data Analysis and Urban Imagery in Modeling" which focused on refining deep learning techniques, data collection methods, and participants' variability to understand human perceptions more accurately. The last topic was named "Greenery and visual exposure in urban spaces" which focused on the impact of the amount and the exposure of green features on mental health and perceptions. Upon reviewing the papers, this study provides a guide for subsequent research to enhance the view of using deep learning techniques to understand how urban environments influence mental health. It also provides various suggestions that should be taken into account when planning outdoor spaces.

Spatio-temporal effects of built environment on running activity based on a random forest approach in nanjing, China

Running activity is closely related to the urban built environment in terms of mental and physical health, and this relationship can change as a result of spatio-temporal changes. Most studies, however, do not account for this and assume a linear relationship exists between the built environment and running activity. This study, therefore, collected running data spanning 2019-2022, studied spatial distribution of four-year running activity, established built environment indicators, used a random forest approach to investigate the non-linear relationship between them, and evaluated spatio-temporal changes in the relationships over time. The findings suggested that running activities are spatially clustered and the degree of clustering varies over time, and nonlinear relationships and threshold effects between the built environment and running activity can be found through the random forest algorithm and partial dependence plots. Urban park green space, greenway, and the normalized difference vegetation index had the most significant effects on running activity. The effects of population, buildings, streets, road intersections, and points of interest on running activity changed during the Coronavirus disease 2019 pandemic. In 2022, however, these effects were consistent with those during the pre-pandemic period. Our findings fill research gaps by using spatio-temporal analysis and a non-linear approach; they can also provide a reference for urban planners in building running-suitable and healthy cities.

Effects of greenery at different heights in neighbourhood streetscapes on leisure walking: a cross-sectional study using machine learning of streetscape images in Sendai City, Japan

BACKGROUND

It has been pointed out that eye-level greenery streetscape promotes leisure walking which is known to be a health -positive physical activity. Most previous studies have focused on the total amount of greenery in the eye-level streetscape to investigate its association with walking behaviour. While it is acknowledged that taller trees contribute to greener environments, providing enhanced physical and psychological comfort compared to lawns and shrubs, the examination of streetscape metrics specifically focused on greenery height remains largely unexplored. Therefore, this study examined the relationship between objective indicators of street greenery categorized by height from a pedestrian viewpoint and leisure walking time.

METHODS

We created streetscape indices of street greenery using Google Street View Images at 50-m intervals in an urban area in Sendai City, Japan. The indices were classified into four ranges according to the latitude of the virtual hemisphere centred on the viewer. We then investigated their relationship to self-reported leisure walking.

RESULTS

Positive associations were identified between the street greenery in higher positions and leisure walking time, while there was no significant association between the greenery in lower positions.

CONCLUSION

The findings indicated that streets with rich greenery in high positions may promote residents' leisure walking, indicating that greenery in higher positions contributes to thermally comfortable and aesthetic streetscapes, thus promoting leisure walking. Increasing the amount of greenery in higher positions may encourage residents to increase the time spent leisure walking.

Association between urban green space and postpartum depression, and the role of physical activity: a retrospective cohort study in Southern California

Background

Little research exists regarding the relationships between green space and postpartum depression (PPD). We aimed to investigate the relationships between PPD and green space exposure, and the mediating role of physical activity (PA).

Methods

Clinical data were obtained from Kaiser Permanente Southern California electronic health records in 2008-2018. PPD ascertainment was based on both diagnostic codes and prescription medications. Maternal residential green space exposures were assessed using street view-based measures and vegetation types (i.e., street tree, low-lying vegetation, and grass), satellite-based measures [i.e., Normalized Difference Vegetation Index (NDVI), land-cover green space, and tree canopy cover], and proximity to the nearest park. Multilevel logistic regression was applied to estimate the association between green space and PPD. A causal mediation analysis was performed to estimate the proportion mediated by PA during pregnancy in the total effects of green space on PPD.

Findings

In total, we included 415,020 participants (30.2 ± 5.8 years) with 43,399 (10.5%) PPD cases. Hispanic mothers accounted for about half of the total population. A reduced risk for PPD was associated with total green space exposure based on street-view measure [500 m buffer, adjusted odds ratio (OR) per interquartile range: 0.98, 95% CI: 0.97-0.99], but not NDVI, land-cover greenness, or proximity to a park. Compared to other types of green space, tree coverage showed stronger protective effects (500 m buffer, OR = 0.98, 95% CI: 0.97-0.99). The proportions of mediation effects attributable to PA during pregnancy ranged from 2.7% to 7.2% across green space indicators.

Interpretation

Street view-based green space and tree coverage were associated with a decreased risk of PPD. The observed association was primarily due to increased tree coverage, rather than low-lying vegetation or grass. Increased PA was a plausible pathway linking green space to lower risk for PPD.

Funding

National Institute of Environmental Health Sciences (NIEHS; R01ES030353).

Virtual reality assessment of walking in a modifiable urban environment: a feasibility and acceptability study

Physical activity is known to be one of the most health-beneficial behaviors, and salutogenic design modifications to the built environment can facilitate increased physical activity. Unfortunately, it is not often clear in advance which environmental and urban design implementations will generate increases in activities such as walking, and which will have little impact or even reduce walking. The present study tested the feasibility and acceptability of a virtual reality (VR) model for pre-testing urban designs for their impact on walking. Using a wearable VR head-mounted display/computer, young adults (n = 40) walked freely through a large indoor gymnasium, simultaneously walking through a virtual model of an urban streetscape that was designed to be modifiable and allow for testing impacts on walking of various changes to the urban environment. The majority of participants found the experience to be acceptable: pleasant and nonaversive, and they walked freely through the VR model for approximately 20 min, on average. Using modifiable VR models to pre-test built-environment changes for their impacts on walking behavior appears to be a feasible and acceptable approach and worthy of continued research investigation.

Built Environment Features Obtained from Google Street View Are Associated with Coronary Artery Disease Prevalence: A Deep-Learning Framework

Background

Built environment plays an important role in development of cardiovascular disease. Tools to evaluate the built environment using machine vision and informatic approaches has been limited. We sought to investigate the association between machine vision-based built environment and prevalence of cardiometabolic disease in urban cities.

Methods

This cross-sectional study used features extracted from Google Street view (GSV) images to measure the built environment and link them with prevalence of cardiometabolic disease. Convolutional neural networks, light gradient boosting machines and activation maps were utilized to predict health outcomes and identify feature associations with coronary heart disease (CHD). The study obtained 0.53 million GSV images covering 789 census tracts in 7 cities (Cleveland, OH; Fremont, CA; Kansas City, MO; Detroit, MI; Bellevue, WA; Brownsville, TX; and Denver, CO). Analyses were conducted from February 2022 to December 2022. We used census tract-level data from the Centers for Disease Control and Prevention's PLACES dataset. Main outcomes included census tract-level estimated prevalence of CHD based on GSV built environment features.

Results

Built environment features extracted from GSV using deep learning predicted 63% of the census tract variation in CHD prevalence. The ExtraTrees Regressor achieved the best result among all models with the lowest average mean absolute error of 1.11% and Root mean square of error of 1.58. The addition of GSV features outperformed and improved a model that only included census-tract level age, sex, race, income and education. Activation maps from the features revealed a set of neighborhood features represented by buildings and roads associated with CHD prevalence.

Conclusions

In this cross-sectional study, a significant portion of CHD prevalence were explained by GSV-based built environment factors analyzed using deep learning, independent of census tract demographics. Machine vision enabled assessment of the built environment could help play a significant role in designing and improving heart-heathy cities.

An interpretable machine learning framework for measuring urban perceptions from panoramic street view images

The proliferation of street view images (SVIs) and the constant advancements in deep learning techniques have enabled urban analysts to extract and evaluate urban perceptions from large-scale urban streetscapes. However, many existing analytical frameworks have been found to lack interpretability due to their end-to-end structure and "black-box" nature, thereby limiting their value as a planning support tool. In this context, we propose a five-step machine learning framework for extracting neighborhood-level urban perceptions from panoramic SVIs, specifically emphasizing feature and result interpretability. By utilizing the MIT Place Pulse data, the developed framework can systematically extract six dimensions of urban perceptions from the given panoramas, including perceptions of wealth, boredom, depression, beauty, safety, and liveliness. The practical utility of this framework is demonstrated through its deployment in Inner London, where it was used to visualize urban perceptions at the Output Area (OA) level and to verify against real-world crime rate.

Machine Learning Approach for Automated Detection of Irregular Walking Surfaces for Walkability Assessment with Wearable Sensor

The walkability of a neighborhood impacts public health and leads to economic and environmental benefits. The condition of sidewalks is a significant indicator of a walkable neighborhood as it supports and encourages pedestrian travel and physical activity. However, common sidewalk assessment practices are subjective, inefficient, and ineffective. Current alternate methods for objective and automated assessment of sidewalk surfaces do not consider pedestrians' physiological responses. We developed a novel classification framework for the detection of irregular walking surfaces that uses a machine learning approach to analyze gait parameters extracted from a single wearable accelerometer. We also identified the most suitable location for sensor placement. Experiments were conducted on 12 subjects walking on good and irregular walking surfaces with sensors attached at three different locations: right ankle, lower back, and back of the head. The most suitable location for sensor placement was at the ankle. Among the five classifiers trained with gait features from the ankle sensor, Support Vector Machine (SVM) was found to be the most effective model since it was the most robust to subject differences. The model's performance was improved with post-processing. This demonstrates that the SVM model trained with accelerometer-based gait features can be used as an objective tool for the assessment of sidewalk walking surface conditions.

The influence of neighborhood quality on tourism in China: Using Baidu Street View pictures and deep learning techniques

Previous studies have investigated the determinants of urban tourism development from the various attributes of neighborhood quality. However, traditional methods to assess neighborhood quality are often subjective, costly, and only on a small scale. To fill this research gap, this study applies the recent development in big data of street view images, deep learning algorithms, and image processing technology to assess quantitatively four attributes of neighborhood quality, namely street facilities, architectural landscape, green or ecological environment, and scene visibility. The paper collects more than 7.8 million Baidu SVPs of 232 prefecture-level cities in China and applies deep learning techniques to recognize these images. This paper then tries to examine the influence of neighborhood quality on regional tourism development. Empirical results show that both levels of street facilities and greenery environment promote tourism. However, the construction intensity of the landscape has an inhibitory influence on the development of tourism. The threshold test shows that the intensity of the influence varies with the city's overall economic level. These conclusions are of great significance for the development of China's urban construction and tourism economy, and also provide a useful reference for policymakers. The methodological procedure is reduplicative and can be applied to other challenging cases.

Street images classification according to COVID-19 risk in Lima, Peru: a convolutional neural networks feasibility analysis

OBJECTIVES

During the COVID-19 pandemic, convolutional neural networks (CNNs) have been used in clinical medicine (eg, X-rays classification). Whether CNNs could inform the epidemiology of COVID-19 classifying street images according to COVID-19 risk is unknown, yet it could pinpoint high-risk places and relevant features of the built environment. In a feasibility study, we trained CNNs to classify the area surrounding bus stops (Lima, Peru) into moderate or extreme COVID-19 risk.

DESIGN

CNN analysis based on images from bus stops and the surrounding area. We used transfer learning and updated the output layer of five CNNs: NASNetLarge, InceptionResNetV2, Xception, ResNet152V2 and ResNet101V2. We chose the best performing CNN, which was further tuned. We used GradCam to understand the classification process.

SETTING

Bus stops from Lima, Peru. We used five images per bus stop.

PRIMARY AND SECONDARY OUTCOME MEASURES

Bus stop images were classified according to COVID-19 risk into two labels: moderate or extreme.

RESULTS

NASNetLarge outperformed the other CNNs except in the recall metric for the moderate label and in the precision metric for the extreme label; the ResNet152V2 performed better in these two metrics (85% vs 76% and 63% vs 60%, respectively). The NASNetLarge was further tuned. The best recall (75%) and F1 score (65%) for the extreme label were reached with data augmentation techniques. Areas close to buildings or with people were often classified as extreme risk.

CONCLUSIONS

This feasibility study showed that CNNs have the potential to classify street images according to levels of COVID-19 risk. In addition to applications in clinical medicine, CNNs and street images could advance the epidemiology of COVID-19 at the population level.

Preference for Street Environment Based on Route Choice Behavior While Walking

This study aimed to better understand the relationship between the street environment and walking behavior by deciphering the pedestrians' street environment preference based on their route choice behavior while walking. The route data of 219 residents were collected using an unobtrusive tracking method and subjected to binary logistic regression models to analyze the pedestrian route choice behavior. The results revealed that except for the walking distance, the trip purpose and travel status are the potential factors influencing the route choice of pedestrians. Furthermore, it was revealed that on-street parking, garbage bins, and streetlights could influence the pedestrians to select longer distance routes. In addition, pedestrians were more likely to select the shortest distance route when they were engaged in leisure activities with an accompanist. The findings of this study would offer insights, from different perspectives, into the micro-scale street environment and the walking behavior of pedestrians.

Impact of the COVID-19 Pandemic on Walkability in the Main Urban Area of Xi’an

The COVID-19 pandemic has greatly affected the mobility of individuals everywhere. This has been especially true in China, where many restrictions, including lockdowns, have been widely applied. This paper discusses the impact of the pandemic on walkability, an important factor in promoting urban neighborhoods, in the main urban area of Xi’an, China, one of China’s four great ancient capitals. Based on the street view data obtained before and after the pandemic, the paper quantitatively compares changes in specific components of selected streetscapes through a deep learning (DL) street view analysis. The aim is to identify the impact of the pandemic on walkability and determine the elements that influence increased walkability in Xi’an’s historical area, using a walkability evaluation model based on a regression analysis involving three factors (streetscape components, walkability check scores, and street connectivity of space syntax for every image). Although Xi’an’s urban structure did not change significantly, the pandemic has clearly impacted street vitality, especially in terms of reducing pedestrian flow and commercial value. Based on study results, the street environment has great room for improvement, especially in the city’s historical blocks, by reconsidering safety measures to pedestrians and the important role of atmospheric aspects on the streets.

A Framework of Community Pedestrian Network Design Based on Urban Network Analysis

Community is the foundation of modern cities, where urban residents spend most of their lifetime. Effective and healthy community design plays a vital role in improving residents’ living quality. Pedestrian network is an indispensable element in the community. Successful pedestrian network design can help the residents be healthy both physically and mentally, build the awareness of “Go Green” for the society, and finally contribute to low-carbon and green cities. This paper proposes a community pedestrian network design method based on Urban Network Analysis with the help of the Rhino software. A case study of a typical community in Guangzhou, China was implemented, specifying the steps of the proposed method. The findings presented include the features of the citizens and the accessibilities of the neighbors that are obtained from the community pedestrian network simulation. The limitation and scalability of this method was discussed. The proposed method can be essential to designing healthy and sustainable communities.

Measuring Perceived Psychological Stress in Urban Built Environments Using Google Street View and Deep Learning

An urban built environment is an important part of the daily lives of urban residents. Correspondingly, a poor design can lead to psychological stress, which can be harmful to their psychological and physical well-being. The relationship between the urban built environment and the perceived psychological stress of residents is a significant in many disciplines. Further research is needed to determine the stress level experienced by residents in the built environment on a large scale and identify the relationship between the visual components of the built environment and perceived psychological stress. Recent developments in big data and deep learning technology mean that the technical support required to measure the perceived psychological stress of residents has now become available. In this context, this study explored a method for a rapid and large-scale determination of the perceived psychological stress among urban residents through a deep learning approach. An empirical study was conducted in Gangnam District, Seoul, South Korea, and the SegNet deep learning algorithm was used to segment and classify the visual elements of street views. In addition, a human-machine adversarial model using random forest as a framework was employed to score the perception of the perceived psychological stress in the built environment. Consequently, we found a strong spatial autocorrelation in the perceived psychological stress in space, with more low-low clusters in the urban traffic arteries and riverine areas in Gangnam district and more high-high clusters in the commercial and residential areas. We also analyzed the street view images for three types of stress perception (i.e., low, medium and high) and obtained the percentage of each street view element combination under different stresses. Using multiple linear regression, we found that walls and buildings cause psychological stress, whereas sky, trees and roads relieve it. Our analytical study integrates street view big data with deep learning and proposes an innovative method for measuring the perceived psychological stress of residents in the built environment. The research methodology and results can be a reference for urban planning and design from a human centered perspective.

Assessing Inequity in Green Space Exposure toward a "15-Minute City" in Zhengzhou, China: Using Deep Learning and Urban Big Data

Green space exposure is considered an important aspect of a livable environment and human well-being. It is often regarded as an indicator of social justice. However, due to the difficulties in obtaining green space exposure data from a ground-based view, an effective evaluation of the green space exposure inequity at the community level remains challenging. In this study, we presented a green space exposure inequity assessment framework, integrating the Green View Index (GVI), deep learning, spatial statistical analysis methods, and urban rental price big data to analyze green space exposure inequity at the community level toward a “15-minute city” in Zhengzhou, China. The results showed that green space exposure inequality is evident among residential communities. The areas in the old city were with relatively high GVI and the new city districts were with relatively low GVI. Moreover, a spatially uneven association was observed between the degree of green space exposure and housing prices. Especially, the wealthier communities in the new city districts benefit from low green space, compared to disadvantaged communities in the old city. The findings provide valuable insights for policy and planning to effectively implement greening strategies and eliminate environmental inequality in urban areas.

Effect of Urban Green Space in the Hilly Environment on Physical Activity and Health Outcomes: Mediation Analysis on Multiple Greenery Measures

Background: Green spaces reduce the risk of multiple adverse health outcomes by encouraging physical activity. This study examined correlations between urban green space and residents’ health outcomes in hilly neighborhoods: if they are mediated by social cohesion, visual aesthetics, and safety. Methods: We used multiple green space indicators, including normalized difference vegetation index (NDVI) extracted from satellite imagery, green view index (GVI) obtained from street view data using deep learning methods, park availability, and perceived level of greenery. Hilly terrain was assessed by the standard deviation of the elevation to represent variations in slope. Resident health outcomes were quantified by their psychological and physiological health as well as physical activity. Communities were grouped by quartiles of slopes. Then a mediation model was applied, controlling for socio-demographic factors. Results: Residents who perceived higher quality greenery experienced stronger social cohesion, spent more time on physical activity and had better mental health outcomes. The objective greenery indicators were not always associated with physical activity and might have a negative influence with certain terrain. Conclusions: Perceived green space offers an alternative explanation of the effects on physical activity and mental health in hilly neighborhoods. In some circumstances, geographical environment features should be accounted for to determine the association of green space and resident health outcomes.

Measuring the Correlation between Human Activity Density and Streetscape Perceptions: An Analysis Based on Baidu Street View Images in Zhengzhou, China

Although investigators are using data sources to describe the visual characteristics of streets, few researchers have linked human perceptions of the street environment with human activity density. This study proposes a conceptualized analytical framework that explains the relationship between human activity density and the visual characteristics of the streetscape. The image-segmentation model DeepLabv3+ automatically extracts each pixel’s semantic information and classifies visual elements from 120,012 collected panoramic street view images of Zhengzhou, China, using the entropy weighting method and weighted superposition to calculate the street perception summary score. This deep learning approach can successfully describe the semantics of streets and the connection between population density and street perception. The study provides a new quantitative method for urban planning and the development of high-density cities.

Using machine learning to examine street green space types at a high spatial resolution: Application in Los Angeles County on socioeconomic disparities in exposure

BACKGROUND

Compared to commonly-used green space indicators from downward-facing satellite imagery, street view-based green space may capture different types of green space and represent how environments are perceived and experienced by people on the ground, which is important to elucidate the underlying mechanisms linking green space and health.

OBJECTIVES

This study aimed to evaluate machine learning models that can classify the type of vegetation (i.e., tree, low-lying vegetation, grass) from street view images; and to investigate the associations between street green space and socioeconomic (SES) factors, in Los Angeles County, California.

METHODS

SES variables were obtained from the CalEnviroScreen3.0 dataset. Microsoft Bing Maps images in conjunction with deep learning were used to measure total and types of street view green space, which were compared to normalized difference vegetation index (NDVI) as commonly-used satellite-based green space measure. Generalized linear mixed model was used to examine associations between green space and census tract SES, adjusting for population density and rural/urban status.

RESULTS

The accuracy of the deep learning model was high with 92.5% mean intersection over union. NDVI were moderately correlated with total street view-based green space and tree, and weakly correlated with low-lying vegetation and grass. Total and three types of green space showed significant negative associations with neighborhood SES. The percentage of total green space decreased by 2.62 [95% confidence interval (CI): -3.02, -2.21, p < 0.001] with each interquartile range increase in CalEnviroScreen3.0 score. Disadvantaged communities contained approximately 5% less average street green space than other communities.

CONCLUSION

Street view imagery coupled with deep learning approach can accurately and efficiently measure eye-level street green space and distinguish vegetation types. In Los Angeles County, disadvantaged communities had substantively less street green space. Governments and urban planners need to consider the type and visibility of street green space from pedestrian's perspective.

Using machine learning to examine street green space types at a high spatial resolution: Application in Los Angeles County on socioeconomic disparities in exposure

BACKGROUND

Compared to commonly-used green space indicators from downward-facing satellite imagery, street view-based green space may capture different types of green space and represent how environments are perceived and experienced by people on the ground, which is important to elucidate the underlying mechanisms linking green space and health.

OBJECTIVES

This study aimed to evaluate machine learning models that can classify the type of vegetation (i.e., tree, low-lying vegetation, grass) from street view images; and to investigate the associations between street green space and socioeconomic (SES) factors, in Los Angeles County, California.

METHODS

SES variables were obtained from the CalEnviroScreen3.0 dataset. Microsoft Bing Maps images in conjunction with deep learning were used to measure total and types of street view green space, which were compared to normalized difference vegetation index (NDVI) as commonly-used satellite-based green space measure. Generalized linear mixed model was used to examine associations between green space and census tract SES, adjusting for population density and rural/urban status.

RESULTS

The accuracy of the deep learning model was high with 92.5% mean intersection over union. NDVI were moderately correlated with total street view-based green space and tree, and weakly correlated with low-lying vegetation and grass. Total and three types of green space showed significant negative associations with neighborhood SES. The percentage of total green space decreased by 2.62 [95% confidence interval (CI): -3.02, -2.21, p < 0.001] with each interquartile range increase in CalEnviroScreen3.0 score. Disadvantaged communities contained approximately 5% less average street green space than other communities.

CONCLUSION

Street view imagery coupled with deep learning approach can accurately and efficiently measure eye-level street green space and distinguish vegetation types. In Los Angeles County, disadvantaged communities had substantively less street green space. Governments and urban planners need to consider the type and visibility of street green space from pedestrian's perspective.

Improving spatial data in health geographics: a practical approach for testing data to measure children's physical activity and food environments using Google Street View

Background

Geographic information systems (GIS) are often used to examine the association between both physical activity and nutrition environments, and children’s health. It is often assumed that geospatial datasets are accurate and complete. Furthermore, GIS datasets regularly lack metadata on the temporal specificity. Data is usually provided ‘as is’, and therefore may be unsuitable for retrospective or longitudinal studies of health outcomes. In this paper we outline a practical approach to both fill gaps in geospatial datasets, and to test their temporal validity. This approach is applied to both district council and open-source datasets in the Taranaki region of Aotearoa New Zealand.

Methods

We used the ‘streetview’ python script to download historic Google Street View (GSV) images taken between 2012 and 2016 across specific locations in the Taranaki region. Images were reviewed and relevant features were incorporated into GIS datasets.

Results

A total of 5166 coordinates with environmental features missing from council datasets were identified. The temporal validity of 402 (49%) environmental features was able to be confirmed from council dataset considered to be ‘complete’. A total of 664 (55%) food outlets were identified and temporally validated.

Conclusions

Our research indicates that geospatial datasets are not always complete or temporally valid. We have outlined an approach to test the sensitivity and specificity of GIS datasets using GSV images. A substantial number of features were identified, highlighting the limitations of many GIS datasets.

Panoramic Street-Level Imagery in Data-Driven Urban Research: A Comprehensive Global Review of Applications, Techniques, and Practical Considerations

The release of Google Street View in 2007 inspired several new panoramic street-level imagery platforms including Apple Look Around, Bing StreetSide, Baidu Total View, Tencent Street View, Naver Street View, and Yandex Panorama. The ever-increasing global capture of cities in 360° provides considerable new opportunities for data-driven urban research. This paper provides the first comprehensive, state-of-the-art review on the use of street-level imagery for urban analysis in five research areas: built environment and land use; health and wellbeing; natural environment; urban modelling and demographic surveillance; and area quality and reputation. Panoramic street-level imagery provides advantages in comparison to remotely sensed imagery and conventional urban data sources, whether manual, automated, or machine learning data extraction techniques are applied. Key advantages include low-cost, rapid, high-resolution, and wide-scale data capture, enhanced safety through remote presence, and a unique pedestrian/vehicle point of view for analyzing cities at the scale and perspective in which they are experienced. However, several limitations are evident, including limited ability to capture attribute information, unreliability for temporal analyses, limited use for depth and distance analyses, and the role of corporations as image-data gatekeepers. Findings provide detailed insight for those interested in using panoramic street-level imagery for urban research.