Is there a way to check duplicate images with different names using php? [closed]

Question

Is there a way to check duplicate images with different names using php ? I want to delete all duplicates.

Answer 1

You can compare and check it by sha1_file hash of a file

It returns 40 character hex number

Answer 2

I suppose a somewhat simple solution would be to do a checksum on the images using md5().

Open a directory, loop through the files generating md5s, compare md5s, delete duplicates.

EDIT: Here's a script using hash_file()

<?php

$dir = "/full/path/to/images";
$checksums = array();

if ($h = opendir($dir)) {
    while (($file = readdir($h)) !== false) {

        // skip directories
        if(is_dir($_="{$dir}/{$file}")) continue;

        $hash = hash_file('md5', $_);

        // delete duplicate
        if (in_array($hash, $checksums)) {
            unlink($_);
        }
        // add hash to list
        else {
            $checksums[] = $hash;
        }
    }
    closedir($h);
}

Answer 3

I spent a lot of time looking for the best solution in php, but failed, read my 5 steps to heaven (or just get step #5).

1. hash_file does not work as desired, for example in a folder of 11000 pictures with different names I know that there are only 800 unique, hash_file () found only 30 matches.

2. I could not install a third-party library like http://libpuzzle.pureftpd.org/project/libpuzzle/php on Windows + Openserver.

3. Tried to compare by dominant color or pixel-by-pixel ImageColorAt() , creating "digital stamp of image". It works very slow, manycoding and in final very bad - changing size or merge/rotate images are elusive.

4. Checked Github to find readytogo solution, but there are no any solution on PHP (why? It was surprise for me).

5. Finally, I found the shareware desktop program http://www.mindgems.com/products/VS-Duplicate-Image-Finder/VSDIF-Tutorials.htm?postinstall=1 which worked just super (fast! it works in multithreading and loads CPU to 100%, 8gb and 11000 images compared in just ~30 secs) and has all the necessary functions, exceptions and filtering. In those 11000 images dir this program founded all visual similar images, showing me groups and instances, allowing to move selected with autofilters and etc. The main disadvantage is money, but there are torrents ;)