Joseph Chen's Braille Vision Is a Portable, Raspberry Pi-Powered Text-to-Braille Converter

Maker Joseph Chen and colleagues have shown off a device designed to convert written text into braille on the fly: the Braille Vision.

"The aim of our project, Braille Vision, was to improve media accessibility for the visually impaired, by letting them read text around them that has no braille transcriptions available," Chen explains of the project. "It works by first taking a photo of media such as a poster that is inaccessible to the visually impaired, reading the text from it, before converting this data into a VI accessible format (braille) via a braille display pad."

Invented by Louis Braille in 1824, at just fifteen years of age, Braille is a touch-based writing system that turns the alphabet into embossed dots arranged in a 3×2 cell. The reader runs their fingertip over the dots and by feeling their presence and absence can read by touch — no sight required.

Refreshable braille displays for use with computer systems are nothing new, but the Braille Vision aims to take the concept further by providing a real-time transcription system. Inside the housing is a Raspberry Pi 4 Model B or Raspberry Pi 5 single-board computer connected to a Raspberry Pi Camera Module and a refreshable braille display. When pointed towards writing, the system captures a photo and feeds it through the open-source Tesseract OCR optical character recognition system to turn the image into text — then passes the text to the braille display for the user's consumption.

The braille display itself is based on a 3×2 array of compact solenoids, connected to the Raspberry Pi via an Arduino UNO microcontroller and driven using MOSFETs. As each character is chosen for display, the tips of the solenoids are ejected or retracted accordingly — and the reader can feel the change.

"We think that this project is a great way to start learning about the vast range of practical applications you can achieve with a Raspberry Pi (this being our first project using one)," Chen says, "as well as more general electronics, Arduino, and braille."

The project is documented in full, with source code and 3D print files, on Instructables.