An Indoor Navigation App using Computer Vision and Sign Recognition

Area of Science:

• Computer Vision
• Human-Computer Interaction
• Assistive Technology

Background:

• Visually impaired individuals face significant challenges with indoor navigation due to the lack of visual cues.
• Existing navigation aids often do not adequately address the complexities of indoor environments.

Purpose of the Study:

• To develop and describe an accessible iOS application for indoor wayfinding for people with visual impairments.
• To enhance the real-time localization and direction-giving capabilities for visually impaired users.

Main Methods:

• A computer vision-based localization approach combining visual-inertial odometry (VIO) with informational sign detection and map constraints.
• Development of an iOS application providing turn-by-turn directions.
• Improvements to sign detection and range estimation algorithms for accurate localization.

Main Results:

• The developed app provides real-time, turn-by-turn navigation assistance.
• The localization approach effectively integrates VIO with environmental cues like Exit signs.
• Enhanced algorithms improve the accuracy of sign detection and distance estimation.

Conclusions:

• The created iOS app offers a viable solution for indoor navigation challenges faced by the visually impaired.
• The integrated localization method provides a robust foundation for accessible wayfinding systems.
• Further development can significantly improve independent mobility for visually impaired individuals.