Is it possible to do virtual screen scraping in Perl, or Python? Suppose I have a login page, where once I enter the username or password, it takes me to another authentication page where I have to write what a Captcha reads. Now with Mechanize or a similar tool I can automate the first step. However, for the 2nd step, is it possible to capture a screenshot of the captcha page (virtual, since we are not really loading the page in a browser) through Perl? Once it is done, perhaps I can automate a captcha reading tool (Google has one), which will attempt to read it. (All such captcha pages will have the captcha image in a fixed place within a fixed size box, so I can use Imagemagick to crop that part of the screenshot and feed to the Google tool. It will take a few trial and error runs to find out which portion of the screenshot contains the captcha). So is it possible?
You don't need to emulate or do anything with screen at all. Just trace where CAPTCHA requests its image data and download it yourself - you'll have ready image file.