Enhancing Object Detection and Navigation for Visually Impaired Individuals: A YOLOv3-Based Approach with GPS Integration

This study focuses on enhancing the life quality of visually impaired individuals by innovatively upgrading the functionalities of Electronic Travel Aids (ETAs). Our research focuses on creating a unique approach that caters to their navigation challenges by integrating object recognition, Global Positioning System (GPS) technology, and communication infrastructures in ETAs. We have utilized the engineered YOLOv3 algorithm for real-time multi-object detection and a GPS-powered location system that accurately informs caregivers about their loved one's exact position ensuring enhanced safety for them in unknown places or times. This approach has been validated through rigorous mechanisms such as Precision, Recall, F1-Score, Intersection over Union (IoU), and an impressive Mean Average Precision (mAP) score of 99.42%, with an average distance difference of 6.1 meters from the last location. Implementing this novel ETAs design can significantly augment outdoor navigation safety while improving independence levels among visually impaired individuals.