Simulation of the spatial pattern of scenic spots combining optimal scale and deep learning

Abstract

With the development of deep learning and big data technology, artificial neural network methods are used to simulate new areas with high potential to develop tourist attractions. They break through the limitation of the lifespan development of domestic tourist attractions and improve the credibility of results caused by the sample size and scale effect. This study applied the data for 906 scenic spots in Northwest China by a geographic detector model and deep learning technology to explore the dominant factors explaining their spatial distribution under the optimal research spatial scale and to simulate new areas with a high potential for development as tourist attractions. The main conclusions of this study were as follows. 1) The results were more reliable for Northwest China under a research spatial scale of 150 km × 150 km. 2) The leading factors affecting the development of tourist attractions in Northwest China were normalized difference vegetation index (NDVI), distance from the city, population, and transportation accessibility. 3) The results identified areas that are highly suitable for tourist attractions, showed regional maturity in either the natural environment or social development. This study can act as a reference for further exploration and the application of artificial intelligence technology in scenic spots.