Implementing a Simple CAPTCHA Generator and Recognizer in PHP

The article explains why many web pages use CAPTCHAs and questions their security, then shows a simple PHP program that both generates and cracks such CAPTCHAs.

CAPTCHA Generation – A 4‑character code is randomly built from alphanumeric characters, drawn onto a 60×30 image with random background colors, and output as PNG data.

Binary Conversion (Binarization) – The generated image is read pixel by pixel; black pixels are marked with "*" and others with "0" to create a binary matrix.

Segmentation – Columns containing only zeros are filtered out, the remaining columns are grouped into character blocks, and empty rows are removed to isolate each character.

Feature Extraction – For each isolated character, a string representation is built and compared with a predefined set of feature patterns for letters, digits and symbols.

Recognition – The similarity between the extracted pattern and each stored pattern is calculated using similar_text; the character with the highest similarity is selected as the recognized symbol.

PHP Implementation

/**
 * Simple CAPTCHA recognizer
 * @author zhjx922
 */
class vCode {
    // character feature patterns
    private $_wordKeys = array(
        'A' => '000**00000****000**00**0**0000****0000****0000************0000****0000****0000**',
        // ... (other patterns omitted for brevity) ...
        '9' => '00****000**00**0**0000****0000**0**00***00***0**000000**0*0000**0**00**000****00'
    );

    /**
     * Generate CAPTCHA image
     * @author 武老师
     */
    public function make($verCode = '') {
        if (empty($verCode)) {
            $baseChars = 'ABCDEFGHIJKLMNPQRSTUVWXYZabcdefghigklmnopqrstuvwxyz0123456789';
            $verCode = '';
            $codeCharLenth = 4;
            for ($i = 1; $i <= $codeCharLenth; $i++) {
                $verCode .= $baseChars{mt_rand(0, strlen($baseChars) - 1)};
            }
        }
        $font_size = 20;
        $width = 60;
        $height = 30;
        $img = imagecreate($width, $height);
        $bgR = mt_rand(50, 200);
        $bgG = mt_rand(50, 200);
        $bgB = mt_rand(50, 200);
        $background = imagecolorallocate($img, $bgR, $bgG, $bgB);
        $black = imagecolorallocate($img, 0, 0, 0);
        imagestring($img, 5, 9, 8, $verCode, $black);
        ob_start();
        imagepng($img);
        $image = ob_get_contents();
        ob_end_clean();
        return array('image' => $image, 'code' => $verCode);
    }

    /**
     * Convert image string to binary matrix
     */
    public function getImage($imageString) {
        $im = imagecreatefromstring($imageString);
        list($width, $height) = getimagesizefromstring($imageString);
        $image = array();
        for ($x = 0; $x < $width; $x++) {
            for ($y = 0; $y < $height; $y++) {
                $rgb = imagecolorat($im, $x, $y);
                $rgb = imagecolorsforindex($im, $rgb);
                $image[$y][$x] = ($rgb['red'] == 0 && $rgb['green'] == 0 && $rgb['blue'] == 0) ? '*' : 0;
            }
        }
        return $image;
    }

    // ... (methods remove, removeByLine, getWordString, match, show omitted for brevity) ...
}

$vCode = new vCode();
$codeImage = $vCode->make();
$imageString = $codeImage['image'];
$image = $vCode->getImage($imageString);
$vCode->show($image);
$newImage = $vCode->remove($image);
$code = '';
foreach ($newImage as $img) {
    $vCode->show($img);
    $code .= $vCode->match($img)['key'];
}
echo "生成的验证码为:{$codeImage['code']}
";
echo "识别的验证码为:{$code}
";

The script prints the original binary image, the segmented characters, and finally displays both the generated CAPTCHA code and the code recognized by the algorithm, demonstrating that a simple pattern‑matching approach can successfully break the example CAPTCHA.