This work aims to analyze existing image recognition techniques applied to the identification of braille characters and implement them in an interface that allows the system's portability. Despite being a widely used tactile writing system by individuals with blindness or low vision, there are cases in which people with visual impairments may not adapt to the use of the braille system. For this reason, this work seeks to employ preexisting image recognition algorithms applied to the identification of braille characters and integrate them with a microcontroller, providing more portability for visually impaired individuals to use the technology in their daily tasks. This work includes a brief analysis of image pre-processing techniques to extract the relief of a braille character from a surface. It demonstrates the operation of a convolutional neural network model used for character classification and integrates the previous processes into a Raspberry Pi 4 device coupled with a camera. The project showed that the currently available technology allows for the diversification of assistive technologies, which are crucial for ensuring the inclusion of people with visual impairments.