What is a good workflow for importing and cropping EPS images?

Method A: Dealing directly with PostScript

Since you're dealing with PostScript images, one option might be to use GhostView; I've done so successfully in the past. As an example, consider tiger.eps.

  1. Open tiger.eps using GhostView.

  2. Show the bounding box if you wish to see the extent of the existing whitespace:

    enter image description here

  3. Position the cursor on to identify the left, bottom, right and top coordinates. In my example, these are l b r t = 15 175 564 743.

  4. Open tiger.eps and find the lines starting with %%BoundingBox in the "preamble":

    %!PS-Adobe-2.0 EPSF-1.2
    %%BoundingBox: 1 2 611 792
    
  5. Update this to incorporate the new bounding box:

    %!PS-Adobe-2.0 EPSF-1.2
    %%BoundingBox: 15 175 564 743
    

    There may also be a %%HiResBoundingBox entry, which you can update accordingly.

  6. Refresh the view in GhostView which now shows the cropped bounding box:

    enter image description here

  7. Include as usual via \includegraphics, since the bounding box has been permanently updated. If you don't want to update the EPS permanently, you can also use

    \includegraphics[...,bb=15 175 564 743, clip=true...]{tiger}
    

    in your document.


Method B: Cropping via PDF

  1. Use the command-line conversion

    epstopdf tiger.eps
    

    which produces tiger.pdf.

  2. Use pdfcrop

    pdfcrop tiger.pdf
    

    which produces tiger-crop.pdf.

  3. Convert back to EPS using

    pdftops -eps tiger-crop.pdf tiger.eps
    

    to overwrite tiger.eps with a now-cropped version:

    enter image description here

    New bounding box resembles

    %%BoundingBox: 0 0 548 576
    

    which is marginally different to the manual way described in Method A (which yielded a width of 564 - 15 = 549; height of 743 - 175 = 568).


You can use epspdftk, the gui interface to Siep Kroonenberg's epspdf.texlua script. It exists both for Windows and Linux and does everything (converting and cropping the resulting .pdf) in one step. I'll take the example of one of my old graphic files, produced with pstricks and pst-eps, that had a (very) badly computed bounding box. The procedure is very simple: launch epspdftk and check Compute tight bounding box, then click on Open to choose the file to be converted:

enter image description here

Once chosen, you may ask to view the .eps file:

enter image description here

Then click on Convert and save…. A second later, you can look at the resulting .pdf in your favourite pdf viewer:

enter image description here

The .texlua script can also be launched from the command line, if you want to do batch conversion.