How can image processing be improved?

Question

PHP

I'm trying to read captcha with tesseract.
I process the image before sending it to tesseract.

original

After processing

Result: U4E2\.
I.e. does not recognize - 1

The code itself:

$image = new \Imagick('test.jpg');
$image->SharpenImage($radius = 5, $sigma = 1);
$image->gaussianBlurImage($radius = 2, $sigma = 1);
$brightness = 120;
$saturation = 100;
$hue = 60;
$image->modulateImage($brightness, $saturation, $hue);

Tell me, what else can be applied to recognize with a greater chance?

=========
According to the documentation, you can use whitelist, thereby restricting Tesseract to only look for letters/numbers, not symbols.
I tried it like this (as in the documentation):

echo (new TesseractOCR('img.png'))
    ->whitelist(range('a', 'z'), range(0, 9))
    ->run();

But for some reason he ignores these requirements ..