Multilingual Braille-to-Text Conversion and Text-to-Speech System using Raspberry Pi

In spite of the fact that Braille is an important channel of communication for the visually impaired, conventional systems require specialized training and expensive devices that are hard to reach. In this article, the system proposes in a few sections an automatic system of Braille-to-text and speech conversion using the Raspberry Pi 4 and image processing techniques. Previously, systems have relied on mechanical scanners, which are expensive and not portable, whereas the method presented captures an image of Braille text, which is then processed using grayscale conversion, Gaussian blur to reduce noise, and Canny edge detection to extract Braille dots. Contour detection is performed, and bounding boxes are drawn to separate the Braille cells before sorting and mapping them to a predefined dictionary. Special rules are applied for spacing and capitalization, and the extracted text is translated optionally through the Google Translation Application Programming Interface (API). At last, text was converted to speech by Google text-to-speech (GTTS) and outputting from speaker. Experimental results exhibit an efficient accuracy in recognition of Braille text but also a few severe limitations, such as a single-line processing constraint. Future work will focus on real-time processing, multilingual capabilities, and the incorporation of deep learning for improved accuracy. This system finds application in assistive technology, education, and portable Braille translation devices producing outputs for both hearing and visually impaired users. The benefits of the solution include cost-effectiveness, portability, and dual-output functionality.