PHPUnit matcher/assertions for GD image resources

Fund package maintenance!

v4.1.0 2024-04-16 11:24 UTC

This package is auto-updated.

Last update: 2024-07-16 12:50:56 UTC



Trying to assert images with PHPUnit? This project provides a constraint and the required assertions that allow you do to so.

It supports comparing files on disk as well as image resources in memory.


Add this package to your Composer dev-dependencies:

composer require --dev meyfa/phpunit-assert-gd

Compatibility table


The assertions are available as a trait, so you can easily use them in your test case class:

use AssertGD\GDAssertTrait;

class ExampleTest extends PHPUnit\Framework\TestCase
    // this trait adds the assert methods to your test case
    use GDAssertTrait;

    public function testSomething()
        $this->assertSimilarGD('./tests/expected.png', './tests/actual.png');

Plain comparisons

Use assertSimilarGD if you expect 2 images to be exactly equal. Use assertNotSimilarGD if you expect there to be differences.

$this->assertSimilarGD('./tests/img.png', './tests/same.png');
$this->assertNotSimilarGD('./tests/img.png', './tests/other.png');

Threshold values

Provide a number between 0 and 1 to set the error threshold. For example, a value of 0.2 would allow for at most 20% difference.

$this->assertSimilarGD('./tests/img.png', './tests/similar.png', '', 0.2);

Parameter types

Instead of file paths, you can pass in GD image resources. This eliminates having to write something to disk prior to the comparison.

$img = imagecreatetruecolor(10, 10);
$this->assertSimilarGD('./tests/empty-10x10.png', $img);

Manual constraint

If you need to configure mock objects or do other, more complex matching calls, use isSimilarGD to obtain a constraint object (similar to what would be returned by equalTo, isTrue, etc.).


Difference calculation

By default, this library calculates the difference between two images by comparing the RGBA color channel information at each pixel coordinate of the source image and the test image, and averaging the difference between each pixel to calculate the difference score.

This will work for the majority of cases, but may give incorrect scoring in certain circumstances, such as images that contain a lot of transparency.

An alternative calculation method, which scales the RGB color channels based on their alpha transparency - meaning more transparent pixels will affect the difficulty score less to offset their less observable difference on the image itself - can be enabled by adding a new ScaledRgbChannels instance to the 5th parameter of the assertSimilarGD or assertNotSimilarGD methods.

use AssertGD\DiffCalculator\ScaledRgbChannels;

public function testImage()
        new ScaledRgbChannels()

Custom difference calculators

If you wish to completely customise how calculations are done in this library, you may also create your own calculation algorithm by creating a class that implements the AssertGd\DiffCalculator interface.

A class implementing this interface must provide a calculate method that is provided two GdImage instances, and the X and Y co-ordinate (as ints) of the pixel being compared in both images.

The method should return a float between 0 and 1, where 0 is an exact match and 1 is the complete opposite.

You may then provide an instance of the class as the 5th parameter of the assertSimilarGD or assertNotSimilarGD method to use this calculation method for determining the image difference.