Build a Web Spider

This project is to create a tool to automate a web browser to spider a website and search for specific graphical content. The project will accept a series of domains to search and a series of images that it is searching for. It will open a web browser on a desktop and navigate to a web page. It will record all hyperlinks on that page which it can spider. Then, it it will look at the screen visually (as in a screen shot) to see if the screen contains a series of images. It will need to be able to scroll the screen and click on any tabs or dialogs that may be present on the screen. If an image is matched (the image may be sized differently than the one that is being looked for) then a screen shot will be taken of the web browser and that screen shot will be saved in a folder for the domain. The file name will contain the URL that was searched and the name of the image found. The images that are being searched for may be contained within another image or anywhere on the screen that the browser displays. For an example, perhaps I wanted to search the Nike site for any page that contained the Nike logo. This logo could be contained in many different forms throughout the site. It could be on a shoe as part of another graphic. The program search the entire Nike site and take a screen shot of any page where the logo was found and save a screen shot of it. As it spiders a site, it should randomly vary its time for moving from page to page. It should be able to recursively save the pages that have been searched already as to not return to them again.