Record video signal

Hi, I want to make a system where i can record video of the infrared camera onto something (for example a SD card or so). I was wondering if that would be possible by using only an Arduino, a camera and a SD card (and maybe some video shield or so). I dont know much about the type of signal video is, so its hard to me to judge the feasability of this project. Any advise is welcome.

The end result should portable and should be able to record video (maybe 500x500 pixels or so) for preferably 1 week (yes, 7x24h). Alternatively, 2 pictures per second would be ok. The camera is located in the open forest, so battery charging should be not problem with a solar panel i guess.

I know i can do this without the arduino using a multimedia player (Archos, or any other thing that allows to record video), but they are quite expensive.

EDIT: the system will be used to detect what preditors eat our eggs in our bird nests

the two biggest obstacles with "only an arduino" is memory and speed

you dont have enough memory to buffer whats coming in at any decent resolution, and you dont have enough speed to cram it on the SD card before you run out of memory

would it be possible with extra hardware? I dunno depending on how crazy you get anything is possible

Yes, extra hardware is not a problem. I just want to get something that works, and that is cheap :)

Well forget the arduino for this, it has no role in your project.

There will be more elegant solutions, but the simplest solution I can see would be a cheap laptop running motion detection software on the video stream and saving snapshots at a defined rate when motion is detected. I'm sure you would be able to do that with open source software. To power the PC over a week I would disconnect all the peripherals you don't need, and connect it to the biggest deep cycle lead/acid battery you can find. An old UPS battery or golf cart battery would probably do. If you can get any power from a solar collector that would extend the battery life but you would need to calculate the relative currents to decide whether that was worth the effort.