Measuring streetscape perceptions from driveways and sidewalks to inform pedestrian-oriented street renewal in Düsseldorf

https://doi.org/10.1016/j.cities.2023.104472

Abstract

The measurement of urban streetscape quality facilitates the identification of street regeneration. The emerging Street View image and semantic segmentation techniques have recently proven considerable assistance for investigating urban spatial quality. However, current studies using Google Street View (GSV) cannot explicitly reflect pedestrian perceptions, as GSVs are taken from a driveway perspective, which differs significantly from the pedestrian view. This study aims to elucidate the variance in quantitative measurements of streetscape perceptual qualities under different viewpoints and to identify the priority of street renewal based on a combined analysis between street perceptual quality and walking potential. We collected GSVs and self-photographed pedestrian views in the city center of Düsseldorf. 5300 images are gathered, and their semantic information at the pixel level is extracted by the segmentation technique DeepLab V3+. The results indicate that perspective differences have varying geographic effects on street perceptions, with the most significant effect on walkability. In addition, the combined analysis identifies renewal areas and provides planning techniques and operational standards for urban functional zones, streets, and street segments, respectively. The joint spatial syntax and image segmentation-driven approach offers a feasible paradigm for identifying fine-grained street renewal. In short, this study promotes pedestrian-centered urban measurement.

🏙 Background

Urban streets are an essential component of public space, serving multiple functions: they are channels for mobility as well as places for social interaction, recreation, commerce, and tourism. With an increasing focus on human-centered design, urban streetscapes in Germany have gradually improved. However, many studies, particularly those using Google Street View (GSV), often rely on a driveway perspective, which does not accurately reflect pedestrian perceptions. This study aims to address this gap by comparing the differences in streetscape perceptions from both the vehicle and pedestrian viewpoints to better inform pedestrian-oriented street renewal.

🎯 Research Objectives

The main objectives of this study are:

1. To quantify walking potential: Using spatial syntax analysis to assess the walking potential of urban streets.
2. To compare differences in streetscape perceptions: Comparing GSV (vehicle) and self-photographed pedestrian views (SPPV) to understand the variance in street perception.
3. To identify priority areas for street renewal: Combining streetscape perception and walking potential to offer detailed urban renewal recommendations.

🔬 Research Methodology

The study employed the following methods:

1. Data Collection: 5300 images were collected in Düsseldorf, Germany, including both GSV and self-photographed pedestrian views.
2. Image Semantic Segmentation: DeepLab V3+ was used to perform pixel-level semantic segmentation to extract street elements from the images.
3. Spatial Syntax Analysis: The sDNA tool was used to measure the walking potential of streets.
4. Integrated Analysis: Combining walking potential and streetscape perception data to identify priority areas for urban renewal.

📊 Main Results

Key findings of the study include:

1. Impact of Perspective Differences: There are significant differences in street perception between the vehicle and pedestrian perspectives, especially in terms of walkability. Pedestrian views show higher walkability scores.
2. Relationship Between Walking Potential and Street Quality: The walking potential of streets does not always align with their quality. Some streets with high walking potential have medium-quality scores, while some high-quality streets have low walking potential.
3. Variations Across Urban Functional Areas: Different urban zones (e.g., commercial, residential) show varying impacts on streetscape perceptions, indicating the need for tailored renewal strategies.

🛠 Conclusions & Recommendations

This study provides a practical framework for pedestrian-centered street renewal, offering several important insights:

1. Significant Perspective Differences: Differences between the vehicle and pedestrian perspectives have a substantial effect on walkability, with vehicle-based studies often underestimating it.
2. Identification of Renewal Priority Areas: Using the combined analysis of walking potential and street quality, several areas were identified as high-priority for renewal, particularly in commercial and densely trafficked zones.
3. Urban Planning Recommendations: The study suggests urban renewal strategies based on the unique characteristics of different functional zones, including increasing greenery, widening sidewalks, and enhancing street amenities.