I'm not sure if you're wanting to extract the entire tag or the image name path by itself. Here's what I would do.

for($file){
/(<img src=\"(?:\/\w+)+\/image2002\w{6}607.gif\">)/;
$tag=$1;
}
[download]

Then do whatever you'd like with $tag...push it onto an array etc....You say about six characters...the \w{6} will match exactly and only exactly six characters...you can specify a range or characters in the {} so {4,7} would match at least 4 but no more than 7 characters. If you're not concerned with the entire tag...just the path to the image change the location of the outer ()...which capture the match into $1...to be like so:

/<img src=\"((?:\/\w+)+\/image2002\w{6}607.gif)\")>/;

...or if you just want the name of the file:

/<img src=\"(?:\/\w+)+\/(image2002\w{6}607.gif)\">)/;

Hope this helps... -Jason