An approach to the use of stereo vision system and AI for the accessibility of the visually impaired

As result of the daily challenges faced by visually impaired people and which, in a negative way, affect the quality of life of debilitated patients, this article proposes a device developed to offer an assistive, optimized and effective solution. The integration of computer vision with compact processing systems and cameras in Stereo configuration, using Python and OpenCV as tools to deploy the mathematical model and of IA Stereo BGM, enables disparity calculation, obstacle detection, and tactile feedback through vibration. Additionally, still using OpenCV as a computer vision tool, the YOLO AI model incorporates a touch-activated image caption feature, which combined with the information generated by the aforementioned disparity calculations makes it possible, through a text-to-speech engine, to make this function even more complete. This technology provides assistance and contributes to the prevention of accidents.