Abstract
This study aims to optimize the evaluation and decision-making of ethnic tourism resources through the utilization of deep learning algorithms and Internet of Things (IoT) technology. Specifically, emphasis is placed on the recognition and feature extraction of Mongolian decorative patterns, providing new insights for the deep application of cultural heritage and visual design. In this study, the existing DL algorithm is improved, integrating the feature extraction algorithm of ResNet + Canny + Local Binary Pattern (LBP), and utilizing an intelligent decision method to analyze the intelligent development of indigenous tourism resources. Simultaneously, the DL algorithm and IoT technology are combined with visual design and convolutional neural networks to perform feature extraction and technology recognition. Visual design offers an intuitive representation of tourism resources, while fuzzy decision-making provides a more accurate evaluation in the face of uncertainty. By implementing an intelligent decision-making system, this study achieves a multiplier effect. The integration of intelligent methods not only enhances the accuracy of tourism resource evaluation and decision-making but also elevates the quality and efficiency of the tourism experience. This multiplier effect is evident in the system's capacity to manage substantial datasets and deliver prompt, precise decision support, thus playing a pivotal role in tourism resource management and planning. The findings of this study demonstrate that optimizing intelligent development technology for rural tourism through IoT can enhance the efficacy of intelligent solutions. In terms of pattern recognition accuracy, AlexNet, VGGNet, and ResNet achieve accuracies of 90.8 %, 94.5 %, and 96.9 %, respectively, while the proposed fusion algorithm attains an accuracy of 98.8 %. These results offer practical insights for rural tourism brand strategy and underscore the utility of applying fuzzy decision systems in urban tourism and visual design. Moreover, the research outcomes hold significant practical implications for the advancement of Mongolian cultural tourism and provide valuable lessons for exploring novel paradigms in image analysis and pattern recognition. This study contributes beneficial insights for future research endeavors in related domains.