Takes a folder of images, binarizes and deskews them and converts them to a PDF for OCRing
It uses the deskew function from wand and the iSauvola algorithm from doxapy.
Usage: python3 -i input_folder -o output_file.pdf
The resulting PDF can then be processed with OCRMyPDF. OCRMyPDF also has its own deskewing algorithm, but I've found that using it increases the file size considerably.