I don't think it's going to be able to help you if they're scans. From the README:

Tabula only works on text-based PDFs, not scanned documents. If you can click-and-drag to select text in your table in a PDF viewer (even if the output is disorganized trash), then your PDF is text-based and Tabula should work.