Comment by rafram

I’m not aware of any separate OCR in OpenCV. Some builds include an interface to Tesseract, which might be what you’re thinking of. Tesseract certainly benefits from preprocessing (conversion to grayscale, posterization) with OpenCV.