This system seeks to provide a support tool for the visually impaired community, allowing them to preserve, share, and consult braille in a more accessible way. It also represents a valuable resource for educators, students, and people interested in learning the braille system, by facilitating the interpretation and automated translation of physical documents into digital formats. It presents the development of a computational algorithm of traditional computer vision for the identification of characters according to the standard established by the Spanish National Organization of the Blind (ONCE, by its acronym in Spanish) of distance between points and composition of the cells of the tactile reading-writing system. The system was implemented on an embedded hardware platform based on a Raspberry Pi 4 Model B, in charge of processing the image captured by an OV5647-based wide-angle camera. To facilitate the detection of the point, dark field illumination was used. Image processing was carried out using the OpenCV library in Python, with tasks such as contrast enhancement, noise elimination, morphological analysis, segmentation of lines and cells, identification of dots by region in the cell clustering techniques and finally, translation into grade 1 braille, i.e., with no contraction, generating as output the transcription of the text in alphanumeric characters.