Cnn-Based Intelligent Recognition and Digital Dissemination of Rural Cultural Symbols in Anhui, China

Abstract

In rural areas, cultural communication is essential to human contact and societal development. As a crucial component of China's socialist building and revolutionary past, red culture has great significance and a broad impact. The study begins by outlining the significance of regional symbol identification and the context of its use in cultural heritage. Historic city landscape planning is essential to promoting regional identities and spreading cultural value in the face of rapid advancements in heritage preservation and cultural tourist integration. It has been difficult for scholars to define the link between the two, however. The rural cultural elements and dialect characteristics of various locations are reflected in several research on dialect maps in terms of spiritual civilization. In the context of material civilization, several academics have focused on the regional variability of changes in land use and behavior. The limits of the present conventional design approaches are also discussed, along with the need to introduce new technology to enhance them. Using the well-known historic Rural Cultural city of Anhui, China, as an example, this study introduces a multi-label deep learning approach to explore cultural perceptions in tourism heritage settings.  (1) A framework including artifacts, production, traditional music and living culture was constructed utilizing social media big data and an enhanced ResNet-50 model, combining ArcGIS spatial analysis and diversity indexes. (2) The main component of heritage landscapes is artifact culture, which has a "material-dominated, intangible-weak" structure; (3) the intensity of rural cultural perception is unevenly distributed, with core areas demonstrating higher recognition and diversity; (4) diversity indices indicate that specialized locations reveal marked Rural Cultural singularity, while comprehensive venues present stronger Rural Cultural balance, suggesting a need for enhanced integration across locations. The multi-label classification model's accuracy of 92.35% was shown by the findings, confirming its potential. This work expands the application of multi-label deep learning in tourist heritage studies and offers helpful suggestions for international historic sites that cope with mass tourism.