scanbro

This project grew out of the desire to provide an efficient method for scanning one or more documents into a searchable, space-efficient PDF. The name stems from the detail that I currently use it with a Brother scanner. In principle however, any printer supported by SANE can be used.

The script is distributed under the MIT License.

Installation

There is no installation necessary, just make sure you have Python 3 installed.

Currently the scripts make use of SANE, Tesseract, and Ghostscript, so these are also dependencies. ImageMagick can also be useful if you want to extend the script with your own filters.

On Archlinux (or Manjaro), you need the following packages:

tesseract
tesseract-data-deu
tesseract-data-eng
ghostscript
imagemagick
unpaper
texlive-core
evince

You may install the binary to a directory of your choosing with a name of your choosing.

Usage

I prefer to define aliases of the quality settings I want to have for scanning. A set of examples can be found in the aliases.sh file. In whichever directory I am in, I can then scan from the flatbed:

$ scan-xs -s flatbed 2019-01_My_Filename
:: Scan from Brother MFC-J5730DW
-> scanimage --device-name brother4:net1;dev0 --format tiff -x 210 -y 297 --mode Black & White --resolution 300
:: Read file 2019-01_My_Filename.tiff
:: Transform [2019-01_My_Filename.tiff ...] => [2019-01_My_Filename.tesserac
-> tesseract 2019-01_My_Filename.tiff 2019-01_My_Filename.tesseract -l eng+d
:: Transform [2019-01_My_Filename.tesseract.pdf ...] => 2019-01_My_Filename.
-> gs -dNOPAUSE -dSAFER -dQUIET -dBATCH -sDEVICE=pdfwrite -dCompatibil
OLD_BW.tesseract.gs.pdf -dPDFSETTINGS=/ebook -dEmbedAllFonts=false -dC
ageResolution=100 -dGrayImageResolution=100 -dMonoImageResolution=100
orConversionStrategyForImages=Gray 2019-01_My_Filename.tesseract.pdf
-> rm 2019-01_My_Filename.tesseract.pdf
-> mv 2019-01_My_Filename.tesseract.gs.pdf 2019-01_My_Filename.pdf
-> rm 2019-01_My_Filename.tiff

Here, scan-xs is an alias to scanbro -m bw -ccavg low. The output of the above command will be 2019-01_My_Filename.pdf, provided no error occurred. We didn't have to specify -s flatbed, but it keeps the filename simpler for demonstration purposes, as we don't need to do batch scanning.

For other usage examples, have a look at the help. For evaluating the quality settings you would like to use, the --gs-benchmark option is quite useful. Note that it is highly recommended to scan with a resolution of at least 300 DPI. This not only provides the best OCR results from Tesseract, but Ghostscript's downsampling is much more effective the higher the DPI of the input file.

Enjoy!

Known Issues

• Dryrun mode does not always work reliably.
• When specifiying the output name initially, should prompt before overwriting.
• When --clean is specified 3 times, existing files are overwritten, not deleted. This means that it is possible for more files to be in the output that were actually scanned.