Revision history [back]

The grabbing is not limited in size. Basically you just read in the complete data stream bit by bit. Just keep in mind that the larger the resolution, the larger the actual frame transfer will be and the lower your frames per second will get.

If you are planning to immediatly use a region of interest on the captured frame, try looking at your device specification. Most camera's support a built in region of interest functionality, which reduces the amount of data needed to be transfered to the camera.

Also keep in mind, when visualizing these larger frames, as long as the windows is smaller than your screen resolution, everything goes fine. Larger frames will get displayed, but the moment you call mouseactions on that windows, coördinates will get screwed up. Solution here would be to split your frame and display each part sequentially.

EDIT: answer on your second edit question:

You are looking on the wrong term. Basically you want a USB camera wich has an API and development kit connected to it. By using the API you can adapt the camera's properties from software.

Most camera's use something like offset combined with width and heigth to define a region of interest. Also binning (which is in fact skipping pixels) is possible here.

I am using a more expensive camera for my projects over here, from AVT. We got the manta and the guppy camera, which both have an API to select regions of interest.

However, most of these cams are firewire or GigE versions.