Navigation aid for visually impaired people based on AI and AR technologies

Purpose
The number of blind and visually impaired people in Germany is increasing. Finding an object such as a door handle is sometimes a great challenge for these people. By developing a prototype, the present work investigates the hypothesis that the navigation process to a searched object in the immediate surroundings can be partially automated with the help of a smartphone app and thus made easier for the visually impaired person.

Design/methodology/approach
The object recognition of the prototype is based on the neural network of the Google Cloud Vision API, and the navigation on an idea of the three-dimensional environment of the visually impaired person acquired through the augmented reality framework ARCore from Google.

Findings
The evaluation shows that the recognition of, for example, a door handle and the navigation to this door handle are possible if sufficient criteria such as texture-rich surfaces and good lighting conditions are given.

Originality/value
This work presents a novel approach to assist visually impaired individuals by integrating advanced object recognition and navigation technologies into a user-friendly smartphone application.