We have a Spectrum Digital DM365 EVM and we are using DVSDK 3.10.00.19. We are interfacing a 5 MP (2456x2058 pixels) camera to the pixel interface. The camera outputs 12-bit raw Bayer formatted data. Our camera can be setup the way we want prior to connecting to the EVM and there is no need for us to adjust any camera setting during operation. Therefore, from my understanding we do not need a device driver for the our camera. We want to receive full resolution video from our camera and pass it through the VPFE. Our assumption has been to use a suitable demo from the DVSDK and modify the setup of v4l2_capture and dm365_ccdc to tell the VPFE what the resolution is of our camera. However, we've been looking through the DVSDK for days and have been unable to find a suitable starting place for our application. Every example we've seen has an input resolution of 1920x1080 or smaller, and I don't see how to tell the VPFE through the drivers how to deal with an image larger than this size. I know we'll need to do vertical slicing to pass the entire 5MP image through the VPFE; I can't find an example of how to do this with the DVSDK. Could someone point me to a good starting point?
Thanks,
Bryan