Automatically crop borders?

[denilsonsa]

Is it possible to setup paperless/ocrmypdf/unpaper so that the empty borders around the actual document are trimmed?

Use-case: the (network-connected) scanner can easily scan an A4-sized sheet of paper. If I have smaller page, I could theoretically fiddle with the settings in the scanner to set up the exact page dimensions, but that is difficult to do and very time-consuming. (EDIT: no, this scanner doesn't allow that; it only accepts a few preset sizes.) Instead, it is quicker to just scan everything in the default A4 size, and trim the excess later. My question is… how to trim this excess?

I know I can manually get the PDF from the scanner, use pdfimages -all to extract the JPEG of each page, then use cropgtk or jpegtran to crop the borders manually (hoping to preserve the DPI), then use img2pdf to rebuild the PDF. Clearly, this is a lot of work. Can we get something better? Something a bit more automated?

[denilsonsa]

There are some suggestions at Super User and at Ask Ubuntu, but both suggestions (pdfcrop and pdfjam don't actually crop the page, they just adjust the page bounding box, so the rest of the embedded JPEG becomes off-page.

[denilsonsa]

I thought I could write a Pre-consumption script that would extract the images, crop them, and rebuild the PDF. Unfortunately, there is no pdfimages nor jpegtran in the docker image, so I either need a workaround, a solution using another tool, or I need to update the docker image adding these packages.