Lols that's funny I thought that I am a string referred to the datatype. Couldn't figure out why it was repeated several times. I'm going back to my coffee.
You can match that with a simple
'~(?:image/[^\s]*)~' and a preg_match_all assuming they do not contain spaces.
mlseim is getting at you cannot just extract it like that if you are html scraping. You'll need to parse for the tags that are enclosing them, but with well designed html you can use a dom parser like DomDocument or simplexml to pull it up.
header('HTTP/1.1 420 Enhance Your Calm');