Translating 3D Spatial Data into Tactile Braille: A Depth Camera-Based Approach for Assisting Visually Impaired Individuals

This study aims to empower visually impaired individuals to safely perceive their environment by translating depth sensing data into navigable braille. Utilizing an Intel RealSense D435i depth camera, our system captures RGB-D data to generate detailed 3D point clouds. These point clouds undergo real-time spatial mapping and object localization using a Jetson Orin Nano, which processes the data to create a simplified top-down environmental representation. This representation is then converted into a format compatible with braille. The processed data is displayed on a Canute 360 multi-line refreshable braille display, allowing users to tactilely interact with a detailed layout of their surroundings. We will incorporate computer vision and natural language processing to provide semantic descriptions of the environment, enhancing the utility and comprehensiveness of the spatial data for visually impaired users.