Indoor Localization for Visually Impaired Travelers Using Computer Vision on a Smartphone

Area of Science:

• Computer Vision
• Human-Computer Interaction
• Assistive Technology

Background:

• Visually impaired travelers face significant challenges with indoor wayfinding due to reliance on visual cues.
• Global Positioning System (GPS) is inaccurate for indoor localization, limiting navigation options.

Purpose of the Study:

• To develop a computer vision-based indoor localization system for visually impaired individuals.
• To create a smartphone application for real-time indoor navigation support.

Main Methods:

• The approach combines computer vision, inertial sensors, and 2D maps for localization.
• It utilizes existing infrastructure like informational signs (e.g., Exit signs).
• No new physical infrastructure is required for the system.

Main Results:

• The system runs as a real-time app on a conventional smartphone.
• It demonstrated feasibility with five blind travelers in an indoor environment.
• Achieved localization accuracy of approximately 1 meter after algorithm convergence.

Conclusions:

• The computer vision approach offers a viable solution for indoor localization for the visually impaired.
• The system does not require users to aim their phone at specific signs, enhancing usability.
• This technology has the potential to support future full-featured wayfinding applications.