Tecnologia de processamento de imagem para orientação de deficientes visuais dentro do espaço residencial

Abstract

This work seeks to develop a product of assistive technology based on the Raspberry Pi platform in order to allow greater autonomy inside residential spaces for individuals with blindness. Therefore, it was assembled a wearable obstacle detection and path orienting system that uses a camera to capture images directed to the floor immediately in front of the user, in order to identify QR codes created specifically to generate responses within the written algorithm, so that responses are emitted through vibratory patterns in motors and recorded sentences through headphones. The system was developed in this way in order to allow autonomy in the choice of paths by the user with the aid of a low-cost assistive technology product, easily accessible and maintainable, and flexible in regard to the user's needs and preferences. The system’s capabilities were tested in a fixed setup inside a controlled study room without windows and in a movable setup in a controlled corridor, and the results were satisfactory regarding to the detection distances of paths and obstacles and the swiftness of emission of responses. Finally, a body vest made with bag straps was designed and the developed system was installed into it, to be used as a simple and nonintrusive wearable device.