If you want to crawl webpages, you really shouldn't be using the webbrowser control. Use the httpWebRequest class and make your request, get the html string, and you can loop through the links and DOM objects of the HTML string using MSHTML, so you give it to mshtml and it will turn that HTML string into a nice object you can loop through (so you don't need to try and parse the links out using string manipulation, you just loop through everything as they are all turned into objects thanks to mshtml).
Of course, this way, all that javascript and what have you will not run, and you won't have to waste bandwidth and time loading all those images and having all those elements drawn to the screen when they are not needed.
Get me? Let me know if you need more help.