goscanpdf

goscanpdf performs a batch scan with scanimage, converts the uncompressed PNM files to PDF files in the background and concatenates these PDF files to a single, timestamp-named PDF file. It has been tested with a Fujitsu ScanSnap iX500 on a Raspberry Pi 3 running Linux, but might work with other devices as well.

Pages with a dark pixel ratio less than 0.0003 are sorted out. The "dark" threshold for a pixel is 50%.

The resulting PDF file is rsync'ed via SSH to goscanpdf-target:${prefix}scaninput. There will be no password prompt, you must use key based authentication insted.

Installation

Arch Linux users can install goscanpdf from the AUR.

Others can compile it themselves, e.g. for Linux on a Raspberry Pi:

export GOOS=linux
export GOARCH=arm
go build path/to/goscanpdf.go

Setup

goscanpdf uploads scanned documents via SSH. Configure your goscanpdf-target in ~/.ssh/config:

Host goscanpdf-target
    HostName example.com
    User alice

In order to speed up scanimage, you can remove unused backends from /etc/sane.d/dll.conf.

You can use insaned to run goscanpdf when a button on the scanner is pressed.

Configuration

Usage of goscanpdf:
  -cores number
    	Maximum number of simultaneous single-core convert workers. You might want to spare one CPU core for scanimage, which typically produces about 50% of goscanpdf's CPU load but can't be parallelized. (default 3)
  -dpi resolution
    	scan resolution in dots per inch (default 200)
  -prefix string
    	a string which is prepended to the PDF filename

System requirements

The next page is not scanned until the latest page has been passed to a worker thread. Thus there are up to cores+1 uncompressed PNM files in the working directory.

Binaries

• graphicsmagick
• netcat
• openssh
• pdfunite (from poppler)
• rsync
• scanimage >= 1.0.25 (from sane-utils)
• sh (a shell)

Exit values and RaspberryPi LED flash counts:

• 0 - output file successfully created
• 1 - system error
• 2 - network error
• 3 - no scanner found
• 4 - zero pages scanned