Recently I solved a small problem and found it funny enough to write a post.
I had a PDF document with numbers in it (ones and zeros). I needed the numbers for a program, but they were embedded as a scanned picture instead of text. Copying them by hand would be boring and error-prone. I wouldn't want any typos, seeing as the numbers themselves were supposed to be part of an error-correcting code.
Then I thought: perhaps Perl could do this for me! I came up with this:
Running the script produces:
How does it do that? The script divides the image -- read pixel-by-pixel via ImageMagick -- into squares and counts the black pixels in every square. The character "0" has obviously more black than "1"; the threshold was found by experimenting. An empty square has nearly no black pixels at all, and depicts a zero in this example. Calculating a hex value for every row is simple.
I ended up having to write only slightly more characters than the image contained! :)