Publication: The mapping of historical streets on social media: a case study based on image recognition and semantic recognition
In the age of social media, the interaction of online and offline activities has produced a more diverse form of interaction between people and space. Weibo is the largest social media in China, where users share their insights by uploading photos and text. Historical streets carry important urban cultural imprint while its protection is facing the problem of enhancing attractiveness under the rapidly urban development. This study collected the Baidu Street View image and geo-tag Weibo photo of Hengfu historic conservation. Images were recognized through the machine learning algorithm, in order to realize the accurate measurement of the image elements. Building and greenery were the focus of attention of the crowd, and some streets have improved people's contact with greening through the design of vertical greening. Second, we combined the campaign of the hot events such as ' Leave fallen leaves' in social media, with the semantic recognition of the text of users ' Weibo. This paper analyze how the streets interacted with the online crowd with the help of social media, and analyze why some streets are more attractive while others lack attention and record, and put forward suggestions for urban design of historical street in future.
social media, machine learning, historical street, visible greenery