OpenCV project that translates an image with braille into text using edge detection.
I also wrote a blog post about it.
Not very robust at the moment and can only translate braille that has the proper dimensions and alignment as specified in the US braille standards. Some finagling with tolerance and image size might also be needed prior to translating depending on size and resolution of the image.
There are also different ways to express several symbols such as a period, and as of right now the code only accounts for one of them.
It also doesn't handle indicators to capitalize the entire word (such as ^^) or numbers with more than one digit.
Also, if using an online text-to-braille translator to generate an image, it might be the case that the translator is using improper conversions (the one in the link uses the same symbol for period and apostrophe for example).
Below is an example of how this braille to text translator works.
