AI-Based Smart Blind Stick for Real-Time Obstacle Detection and Object Recognition with Voice Assistance

Abstract

Visually impaired persons are faced with a challenge in terms of navigation due to a lack of awareness of their surroundings. In this regard, this project aims to develop a Smart Blind Stick using artificial intelligence for real-time obstacle detection and object identification with voice guidance. In this project, a camera module and ultrasonic sensor are employed for real-time obstacle detection and distance calculation. In addition, a deep learning algorithm based on YOLOv8 is implemented for real-time object detection and classification. Using such a technique, objects such as pedestrians, vehicles, and obstacles can be identified. The identified objects are processed on a laptop and converted into a voice message using an NLP-based text-to-speech module. The proposed solution is a portable, cost-effective solution for visually impaired persons to navigate in both indoor and outdoor environments.