Improve pytesseract quality

[lindsaypoirier]

I'd like to look into some options for improving the OCR quality (thresholds, blurs, inversions, and morphological operations).