Intelligent approach to mining cultural tourism potential areas based on YOLOv4: insights from Macau

Abstract

Based on historical background and geopolitical factors, Macau is closely related to Portuguese cities; Macau also currently attaches great importance to the creation of a Sino-Portuguese platform and cultural tourism to drive better economic development. This study explores a method for identifying and mining cultural tourism potential areas developed under the YOLOv4 framework and cosine similarity. The main conclusions of this study are as follows: (1) While the YOLOv4 method shows good performance and can effectively identify the architectural texture of Portuguese cities, it may face certain difficulties if applied to newly built urban areas in Macau. (2) Through the cosine similarity method, it is found that the architectural texture slices of the three island cities of Portugal and Macau have certain similarities. The city with the highest level of similarity with Macau is Lisboa, followed by Porto and Evora, while Guimaraes has the lowest level of similarity with Macau. This finding serves as a certain reference for the subsequent construction of the cultural tourism characteristics of the region. (3) Comparing the characteristics of Macau city obtained through the cosine similarity algorithm with the city slices that have a high level of similarity in Portugal, we find the following: “street networks”, “street blocks”, squares, and open space form in the Macau Peninsula, while “street blocks” and “large-scale building complexes” form in the outlying islands. These types of spaces have different scales and can be developed in the future in conjunction with cultural tourism activities on different scales. This study applies machine learning methods to the development of urban cultural tourism potential areas, which provides new perspectives and methods for urban tourism development planning.

Keywords

• tourism potential area
• machine learning
• cosine similarity
• portuguese city
• macau