#regex #question

I was going through some HTML markup/ content where my <h1>,<hx> tags have no anchors/ ID's. pain in the butt to do that manually if i want anchors, It occurred to me: why don't I just process the markup and use the &#x003C;h1&#x003e; (..contents ...) &#x003C;/h1&#x003e;
but the (...contents...) atom needs filtering. E.g. "this just in", maybe I want "this_just_in"

That's not really possible w/ #PCRE is it?

Maybe this is something #sed can do? Im not super fmlr w/ all gnutils