Hi Guys,

I have a little problem, was wondering if anyone had any experience with this?

I am using imagemagick to remove whitespace from images, however some images are corrupt and the server hangs and eventually crashes because imagemagick doesnt know what to do, even though I have set the time out(doesnt seem to work).

Currently I am using command line to find images in a directory and apply imagemagick command to them.

Is there any way I can find images that are greater than 10x10 pixels for example? (All the corrupt images seem to be 0x0 pixels).

I am currently using this command:

find /directory/ -name "*.jpg" -exec mogrify -trim -fuzz 1% {} \;

Have you tried mogrify -verbose xxx.jpg ?

For a legit image, it should return something like:

$mogrify -verbose 1.jpg
1.jpg JPEG 100x171 100x171+0+0 8-bit DirectClass 8.99kb 
1.jpg JPEG 100x171 100x171+0+0 8-bit DirectClass 8.95kb

Also identify(1) is part of imageMagick toolkit. It returns size of the image (3rd field), so you could easily do something like

identify image.jpg | awk '{
   print gensub(/x/," ","",$3)}' |  #print space instead of 'x'
   while read w h ; do  #read in width and height 
       if [ $w = 0 ] || [ $h = 0 ] ; then 
          echo zero image 
       fi
   done

Hi Agama,

Many thanks for the reply. I have just tried that on an image I know is corrupt and it does the same, just hangs and uses 100% of one of my cpu's. Had to kill the process.

Doesnt throw an error or anything!

I have the latest build so its all up to date.

Any other ideas what I could possibly try?

thanks for your help.

mike

---------- Post updated at 06:41 PM ---------- Previous update was at 06:33 PM ----------

Hi Mirni,

Thanks for the response. That has confused me massively!

So by looking at this it looks like it finds images that are 0x0 pixels and echos...?

How can I incorprate my -trim syntax in to this and make it recursive on a directory?

Sorry if I sound stupid!

cheers,

Mike

OK, let me be more specific. The code i posted above is a test for size of an image. You can save it in a script, that takes one argument -- image filename. Let's call it "trimNonzero.sh". The contents of this script would be:

#!/bin/bash
# takes an image as argument and processes only if non-zero size

identify $1 | #get image info
  awk '{print gensub(/x/," ","",$3)}' |  #print size with space instead of 'x' 
    while read w h ; do  #read in width and height
         if [ $w = 0 ] || [ $h = 0 ] ; then            
            echo "$1 is zero image: width $w, height $h"
        else
            echo "processing image $1 ..."
            mogrify -trim -fuzz 1% $1        
        fi
    done

chmod u+x trimNonzero.sh

./trimNonzero.sh image.jpg

find $dir -name "*.jpg" -exec ./trimNonzero.sh {} \;