Page 136 - JCAU-7-2

Journal of Chinese
Architecture and Urbanism Urban features of PRD in online image

In the digital era, perceptions of cities are no longer (i) How does the image of towns in the PRD distinguish
solely dependent on direct physical interaction with it from the imagery of urban or rural areas?
urban environments (Negroponte, 1995). Instead, these (ii) What are the distinguishing and memorable
perceptions are increasingly shaped by internet media characteristics of certain central towns in PRD?
and social media platforms (Zhao et al., 2014). These (iii) How can town images, analyzed through new
platforms facilitate open public spaces for content technologies and multi-source data, provide insights
creation and exchange (Yu & Chen, 2024), allowing them into the development level of small-town landscapes?
to play a significant role in “extensively participating in Based on these research questions, this study applies
the generation of urban images” (Kong, 2020, p. 241). Lynch’s city image theory to conduct an empirical analysis.
Scholars now employ Internet platform images and Methods include image type classification, text-based
related technologies to identify and analyze city images. word frequency analysis, and a comprehensive evaluation
For example, Miao (2018) used street-view image data of the landscape in central towns of the PRD. Through
to assess the spatial quality of urban streets, while Zhao this approach, the study aims to expand and adapt Lynch’s
et al. (2015) analyzed Google images of 21 cities in image theory to better understand town-level landscapes.
Guangdong province, China, to compare urban imagery
in cyberspace. 2. Literature review
The advent of artificial intelligence (AI) and Research on city image and landscape in Western countries
advancements in Internet technologies have expanded started earlier than that in China. Early studies primarily
the tools available for such analyses. AI platforms such as analyzed the imageability of urban elements and trajectories,
ChatGPT and Tencent Cloud provide vast datasets (Zhou, while later research focused on calculating the spatial
2024) and highly accurate semantic information (Li et al., perception of Internet images. Psychologist Edward Tolman
2018). These technologies enable the visualization of urban first proposed and generalized the concept of cognitive
spaces through social media feeds and AI-based recognition maps, describing them as comprehensive representations of
tools. For instance, Naik et al. (2014) utilized computer local environments and spaces (Tolman, 1948). Lynch (2001)
vision technology to analyze extensive street-view image further explored city image through psychology and behavior,
datasets, demonstrating its potential for urban research. proposing five key elements of city image. Other studies
Subsequent studies have further explored this approach, derived perceptions of various image attributes using rigorous
refining methodologies and expanding applications (Han mathematical, statistical, and qualitative analyses (Cranshaw
et al., 2022; Naik et al., 2016; Wei et al., 2022; Yao et al., et al., 2012; Hawelka et al., 2014; Hillier, 1999; Hunter, 2008).
2019). Notably, Naik’s work also focused on using computer As information technology has evolved, network media has
vision techniques to predict the formation of “perceptual emerged as a critical medium for comprehending the world,
maps” (Naik et al., 2014). Similarly, as highlighted by Xu with representations of urban images increasingly shifting
(2013), LeCun employed neural networks to train models toward indirect dissemination and data-driven portrayals
on large datasets, enabling machines to recognize and facilitated by digital platforms (McLuhan, 2000; Zhao &
categorize images effectively. Liu, 2012; Zhao, 2012). For example, Salesses et al. (2013)
Since the beginning of the 21 century, the rapid explored urban space perception in terms of safety, hierarchy,
st
urbanization in China has led to the emergence of and equity by analyzing visual representations of four
numerous central towns in the Pearl River Delta (PRD) internet-centric cities. Gavric et al. (2011) analyzed Berlin’s
region in Guangdong province, China. These towns, tourism characteristics and tourist routes using Flickr data,
characterized by strong economic development and while Kavaratzis and Ashworth (2005) linked urban spatial
the early stages of small city formation, have become perception to the development of urban image.
significant focal points for planning and control efforts In China, research on city image began later. Before the
(Tang et al., 2020). However, the townscape often exhibits rise of social media, most studies relied on questionnaire
more “commonalities” than “differences,” and most surveys and cognitive maps to explore urban and rural
studies on the image have primarily focused on urban or landscape image perception. For example, Gu and Song
rural levels (Lin, 1999; Liu,1994; Zhang,2013). Research (2011) emphasized the application of city image research
specifically addressing the image of towns is limited, with in urban design and other fields. Shen (2004) reflected on
existing studies predominantly concentrating on historical Lynch’s five elements, proposing that city image includes
and tourist towns (He et al., 2020). non-spatial elements, such as culture, alongside spatial
To address this gap, this study explores the following elements. Liu and Li (2017) and Liu et al. (2018) applied
research questions: urban landscape iconographic methods to summarize city

Volume 7 Issue 2 (2025) 2 https://doi.org/10.36922/jcau.5733