Well, my assumption that you could "turn down" the image coming from the camera. I'd used MAX to look at one of my AXIS cameras, saw that it had a "JPEG Mono" mode, so I figured I could choose this mode when I configured IMAQdx and decrease the image size from a U32 RGB to a U8 Grey.
Wrong! First, there seemed to be no Property Node (nor Camera Configuration) that would set the mode to give me an 8-bit Image. I could manually set the camera to send no color information, getting a B&W image on my camera, but still getting a U32 image. However, the IMAQ Cast Image function should be very quick in getting the image size converted to a U8 Grayscale.
So let's go back to your VI. In your While loop, you (a) acquire an Image, (b) display it, and (c) "process the pixels" (do some computations on the pixel values). All three of these events occur in sequence, so the loop time is the sum of the times for all three steps. However, because the Camera runs on a "frame-rate" clock, this time, ideally, should be less than 1/Frame Rate so as not to miss any frames.
One way to speed up the loop process is to adopt a Producer/Consumer Design -- as the Images come in, put them on a Queue and pass them to a Consumer loop for display and image processing. This will allow Image Acquisition (which takes camera time) and Image Processing (which takes PC time) to largely occur in parallel, clocked not by the sum of their respective speeds, but by the slower of their individual speeds. Indeed, if you don't need to view every frame, you can drop the occasional Display to speed up the Consumer loop.
Give that a try, and see if this gets the speed of your Image Acquisition and Processing up to acceptable levels. I'm guessing that it will.
Bob Schor
