Ecological Capacity Assessment of Forest Park Tourists by Integrating Attention Mechanisms and Remote Sensing Images

Traditional methods for assessing tourist ecological capacity in forest parks often use single indicators or static models, failing to capture dynamic ecological changes. This study introduces a novel approach leveraging fusion Attention Mechanisms (AM) and Remote Sensing (RS) images to evaluate the ecological capacity of visitors in forest parks. The method uses high-resolution RS images and deep learning to accurately measure the impact of tourist activities on the ecological environment. The proposed method includes an image reconstruction technique that integrates AM with RS data. Performance analysis and validation are conducted, showing an average accuracy of 98.76%, a recall rate of 98.17%, and an F1 score ranging from 96.23% to 98.25% in the Fujian Regional Remote Sensing Image Dataset for Scene classification (FJ-RSIDS). Applying this method to Qilian Mountain Park, the study predicts ecological capacities of 31.43 million for the environment, 104.83 million for space, 115.74 million for tourist psychology, 334,200 for facilities, and 64.84 million for tourism. These predictions closely align with observed values. This research provides a scientific basis and technical support for ecological protection and tourist management in forest parks, contributing significantly to the sustainable development of the tourism industry