Sensing and Navigation of Wearable Assistance Cognitive Systems for the Visually Impaired

This article develops a wearable vision-based assistance system to provide situational awareness for blind and visually impaired (BVI) people in indoor scenarios. The system is built upon nonintrusive wearable devices, including an RGB-D camera, an embedded computer, and haptic modules. First, the depth map and color images of the scene are obtained from an RGB-D camera, which provides 3-D environmental information. The modular work modes are then designed for different tasks, such as navigation and multitarget recognition. Then, the cognition results are summarized and presented to the user through verbal or haptic feedback. Our system is evaluated by a pilot test to validate its effectiveness of improving the navigation capabilities and multitarget recognition capabilities for the BVI in indoor environments. We present study results with different tasks, including navigation, object localization, face recognition, and text reading. The experiments prove that the system can meet the needs of the BVI in daily use.