Abstract
The incorporation of computer vision into tourism information systems has the potential to transform the way tourists interact with their surroundings, offering improved navigation and richer experiences. This research presents a revolutionary signpost navigation system for mobile applications that enhances the tourist experience in scenic areas. The proposed system leverages advanced object recognition and image classification methods to provide real-time location guidance. An Intelligent Ant Colony-tuned Bottleneck Residual Network (IntACO-Bottleneck-ResNet) is utilized for effective image recognition, enabling the system to identify and categorize landmarks in its environment. A single-stage target detection method is then employed to accurately localize targets based on the recognized images. For this study, a specific dataset was compiled, comprising photographs from popular tourist locations, along with signage and landmark images. This dataset, containing both static and dynamic images under various lighting and weather conditions, was annotated for model training and testing. The system’s performance was measured in terms of image recognition accuracy and localization precision. The results demonstrated that the IntACO-Bottleneck-ResNet-based system achieved a high image identification accuracy of 97%, alongside 96% precision, 95% recall, and a 94% F1-score. These findings underscore the potential of this method for real-time, context-aware tourism navigation, significantly enhancing user experience and location-based services in mobile tourism applications.
Keywords
Get full access to this article
View all access options for this article.