A Computer Vision and IoT Based Smart Stick for Assisting Vision-Impaired People

Vision impairment is a worldwide issue, and those without sight suffer more than other individuals with disabilities. The rapid development of Computer Vision and Internet of Things (IoT) technology has made it possible to employ innovative techniques to help visually impaired people. This paper proposes a system that improves the mobility and independence of visually impaired people by helping them navigate. Many existing solutions have struggled with recognizing and categorizing a wide range of objects and obstacles accurately, having low accuracy using camera-based distance measurements and the Global Positioning System (GPS) for navigation. The proposed system recognizes 80 previously trained objects with an additional 8 objects using computer vision and IoT-based techniques. If an object appears within the range of 2 meters, the distance sensor will detect the obstacle, a buzzer will alert, and the camera will capture the image to identify the obstacle using the YOLOv5 algorithm and provide direction to the navigation using headphones. The system makes the decision to instruct the user to move left or right based on audio feedback. The proposed system has improved object identification, distance measurement, and accessibility with satisfactory performances for visually impaired people traveling. Due to Industry 5.0's requirement that humans collaborate with AI and IoT systems, this technology has a huge impact on the growth of smart cities as well as impaired people.