eli-the-bearded / ipevo-tooling Goto Github PK

I have an Ipevo V4K document camera that I want to use for digitizing books controlled from a Linux computer. Ipevo has software for Linux, but it's not great. This is my first stab at a more script friendly interface.

Basic workflow works like this:

0. Install the Ipevo Visualizer package. https://www.ipevo.com/software/visualizer https://web.archive.org/web/20230430235444/https://www.ipevo.com/software/visualizer https://web.archive.org/web/20230430235544/https://ipevo-api-cms.s3-us-west-1.amazonaws.com/software/visualizer/download/Linux/Visualizer_linux_1.1.4.39.zip
1. Install imagemagick (for the rotator tool), img2pdf (creates pdfs), and xdotool (manipulates the Visualizer program). In Ubuntu that's an easy apt-get install, your distro may vary.
2. Copy the files here to your bin directory of choice. My choice is ~/bin/
3. Make and cd into a directory to hold the project files.

   mkdir -p ~/images/first-book
   cd ~/images/first-book
   

4. Run the ipevo-keyboard tool from an xterm (or other X11 terminal)

   ipevo-keyboard
   

5. After start-up, you can set preferences if you like in the Ipevo program. Visualizer will deposit preferences in an xml file in the current directory. A standardized set of preferences will be seeded to the directory: Agrees to the EULA Sets camera to maximum resolution Sets save location to current directory Enables sounds for focus and photo (I find Visualizer is flakey about sound) Without the save location preference, the script fails to find output.
6. I find it helpful to leave the Ipevo window in the upper left and put my terminal in the lower right. The ipevo-keyboard reads and operates on single characters, not complete lines, in a break from typical shell scripts. Usage inside the script:

   Basic operation: return to capture, q to quit.
   
   c to calibrate UI button location
   f to explicitly focus.
   p to take picture without waiting for focus (not stored in list)
   l to list pictures so far
   h for this help
   

   The script controls the Visualizer program using xdotool to simulate mouse movement and clicks plus some hard coded constants (comments in the script has the code I used to figure them out).
7. I take pictures in an "all odd pages", "all even pages", then "backfill reshoots" order. I put the images I want to use in a file called "files", just regular ls order. I view the images in a viewer, my preference is feh and note down the page numbers in a file called "order" file. It's just one page number per line, for matching line in "files". There is a collation tool that works with this, but just putting all the images in page order in the "files" file works too. The sednl script makes it easy to convert 1 3 5 7 9 2 4 6 8 10 into one number per line:

   echo 1 3 5 7 9 2 4 6 8 10 | sednl > order
   

8. Stick metadata that img2pdf uses like 'title', 'subject', etc in files by those names. The subject file content will be passed in as a --subject "Subject here" line.
9. Crop and/or rotate all images if neeed. pnginplacerotate will automate clockwise (cw), counter-clockwise (ccw), and around rotations, eg:

   pnginplacerotate ccw *png
   

   For cropping, I (grudgingly) use gimp, generally configured with a fixed size crop so I can just click and drag the crop rectangle before overwriting.
10. Build the final pdf:

mkpdfbook

Among other flaws of the Visualizer program is that it tries to make a noise after each photo is taken, but it is inconsistent about that. It's a good idea to operate slowly.

Eli the Bearded, 30 April 2023

Next time, I'll include less overlap between pages on the crop:

https://archive.org/details/booklet_202304/page/n11/mode/2up