International Research Journal of Education and Technology

Visually impaired people face many difficulties in their daily lives. They often rely on the help of others. Several techniques have been developed to help the visually impaired. Among the various technologies used to assist the visually impaired, computer vision-based solutions have emerged as one of the most promising options due to their affordability and accessibility. We have proposed a system to provide wearable visual aids for the visually impaired. Its function deals with the identification of objects and helps them to cope with their daily activities of shopping for their things and to navigate to their surroundings on their own. The Raspberry Pi is used on the OpenCV platform to implement artificial vision using the Python language. The system is portable and it can be carried as a model like an ID card and can be carried anywhere. A built-in battery connected to the system powers the device and certain sensors attached to the device detect objects and obstacles in front of blind people and alert them through vibration. It also consists of a camera module that captures and identifies the object and gives voice feedback to blind people. With the use of this smart system, blind people can easily identify the people in front of them, and hear audio feedback with object name and their positions, it acts as a virtual eye for them, understands their surroundings, and helps them to survive confidently.

Key Words: Data set, YOLO Algorithm, Blindness, Machine learning, Python