If you're looking for a simple way to use a multi-page scanner and get your document into a PDF in Linux without any proprietary or commercial software, then ScanPDF might be the solution. I wrote it to quickly take the Linux SANE scanner system output image files, and process them into usable PDFs. By usable, I mean PDFs that maintain their original scanned resolution, omit blank pages (if you're scanning in duplex mode, for example), preserve color unless the original is greyscale/black and white, in which case they are intelligently down-converted to B/W PDFs to save space.

• Free and open-source software: ASL2 license
• Documentation: http://virantha.github.io/scanpdf/html
• Source: https://github.com/virantha/scanpdf

• Uses SANE/scanadf to automatically scan to multi-page compressed PDFs
• Integrates with ScanBd to respond to hardware button presses
• Automatically removes blank pages.
• Scans in color, and automatically down-converts into 1-bit B/W image for text/greyscale images
• Auto-crops to the proper page size.

The simplest way to use this is:

scanpdf scan pdf <pdffile>

This will first perform the scan, and then the conversion to PDF. If you want to split up the scan and the PDF conversion into two separate invocations (for reasons clarified below), then you can do:

scanpdf --tmpdir=tmp scan
scanpdf --tmpdir=tmp pdf <pdffile>

One reason for the separation might be if you want to keep scanning documents (very quick) while the post-processing (slower) for the PDF conversion is taking place in the background. For instance, if you're using the hardware button on the scanner to initiate scans (as detailed in this document), then you want to return immediately after the scan instead of waiting for the full conversion to PDF has taken place.

You can optionally use the following switches to control if you're putting pages face up or face down in the auto document feeder, if you want to skip the blank page processing, adjust the blank page detection threshold, or add additional post-processing using unpaper:

--dpi=<dpi>                 DPI to scan in [default: 300]
--face-up=<true/false>      Face-up scanning [default: True]
--keep-blanks               Don't check for and remove blank pages
--blank-threshold=<ths>     Percentage of white to be marked as blank [default: 0.97]
--post-process              Run unpaper to deskew/clean up

Right now, I'm assuming this is getting called via ScanBD, so I don't have the option to manually specify the scanner. If you really want to use this standalone, for now, please just set the SCANBD_DEVICE environment variable to your scanner device name before running this script.

$ pip install scanpdf

Requires ImageMagick and SANE to be installed, for the command line tools:

• convert
• identify
• ps2pdf
• scanadf

Also requires epstopdf.

The software is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.