If you have Black and White Text, that means that your background is something else.
You might be able to use Canny Edge Detection to detect the characters, and then fill them in.
You might be able to use an adapted Global Thresholding technique, where you find two thresholds, one for White Text and the other for Black Text. When you convert your image to grayscale, and assuming your background is gray since you didn't provide an image, anything between these two thresholds would be background and can be removed. Everything else would be foreground and would be your Text.
If your text was an actual color other than White, you would have to try to isolate it before any grayscale conversion.