Find images by size: find / file / awk

Question

I've been trying to find png image files a certain height (over 500px). I know that file will return image dimensions. Example:

$ file TestImg1a.png

TestImg1a.png: PNG image data, 764 x 200, 4-bit colormap, non-interlaced   

But I need to use this to find all files in a directory with a height over 500px. I know how to print out all files regardless of height:

find . -name '*.png' | xargs file | awk '{print $7 " " $1}'

But how do I limit the $7 to those results greater than 500?

Solution 1

i know this is a bit overkill but, this will work every time (even if there are spaces in your filename) and regardless of how file displays the information.

find . -name '*.png' -exec file {} \; | sed 's/\(.*png\): .* \([0-9]* x [0-9]*\).*/\2 \1/' | awk 'int($1) > 500 {print}'

and it prints the dimensions of the picture and the file

explaination:

1. find all files named *.png under . and for each do a file on it

2. use sed to print only the filename and dimensions then re-order to print dimensions first

3. use awk to test the first number (height of pic) making sure its greater than 500 and if it is print dimensions and file name, if not do nothing.

Solution 2

exiftool -q -r -ext png -if '$ImageHeight > 500' -p '$Directory/$FileName' .

Solution 3

You can also use identify from ImageMagick:

find . -name \*.png -print0|xargs -0 identify -format '%h %f\n'|
awk '$1>500'|cut -d' ' -f2-

Or in OS X:

mdfind 'kMDItemFSName=*.png&&kMDItemPixelHeight>500' -onlyin .

Solution 4

I feel that something other than shell utilities would be more appropriate, e.g., Perl:

#!/usr/bin/perl

use File::Find;
use Image::Info qw(image_info dim);

find (\&check_height, './');

sub check_height {

  my $info = image_info( $_ );
  my ($width, $height) = dim( $info );
  print $_ . " has height $height\n" if ( $height > 500 );

}

Less dicking around with trying to parse out $7; just get the dimensions directly. Yes, you'll need the Image::Info module, but, on CentOS/RHEL, it's a standard package, so you can just run yum install perl-Image-Info.

View more solutions

Comments

[steve-er-rino]

doesn't work: 96, ./4/45445106_w185.png: 86, ./4/404358x_w185.png: 86, ./4/404341x_w185.png: 80, ./4/475986_w185.png: 621, ./4/481693_w185.png: 667, ./4/42513x_w185.png: 86, ./4/404372x_w185.png:

[Mathias Begert]

@tink, cast $7 to an int before comparison, i.e. int($7) > 500.. in the absence of casting awk is resorting to a literal string compare

[tink]

Steve, where does that "," come from? My "file" doesn't produce that. But as Chandra said: you can explicitly force $7 to become an integer using the method pointed out in his comment.

[h3rrmiller]

your awk statement will only work if the filename has no spaces in it

[steve-er-rino]

Thanks. I did have to make a slight change -- the $1 in the awk argument to $3. But this definitely got it for me.

[steve-er-rino]

While perl is a normally a great solution, it's not in this case, esp since I don't have Image::Info nor the option to install.

[steve-er-rino]

@h3rrmiller requiring the ability for filenames with spaces is not a concern and was not included in the question. But thanks for pointing that out for the future.

[h3rrmiller]

I know it wasn't but you never know :)

[rivimey]

The perl solution is somewhat faster than the find/file/awk one, which is nice, and on ubuntu derivates the image info module is available from apt-get install libimage-info-perl