Machine Vision

cancel
Showing results for 
Search instead for 
Did you mean: 

setup using fastRearm

I am looking for information as to the definition and functionality of various directive entries which are used in a camera *.icd file to control a camera. In particular, I have an IMAQ PCI-1424 card with a proprietary camera which simply provides vertical synch, horizontal synch, clock pulses and data. I am trying to operate at  240 Hz, and the download time from the camera is 2.1 msec/ frame (leaving ~ 2msec for the transfer from the card buffer into RAM on the host computer. The frame size is only 16384 words (pixels) at two bytes per word. How should the transfer be setup? Does the FastRearm attribute allow the transfer to occur faster? What other things have to be done to assure that this frame can be loaded into memory without losing frames?  I understand that the use of buffers in the 1424 can allow one frame to be captured, and then download that frame while the next frame is being captured, which would basically alllow for 4 msec for the transfer to RAM. We are having difficulty even getting close to this, and are unable to even get 120 Hz, and are puzzled as to why this is the case. We ar thinking that maybe the *.icd isn't being done right, but have no info on the meaning of several "calls". 
0 Kudos
Message 1 of 2
(3,386 Views)
Charlieboy,
 
Thanks for your post.  Here is some information on Fast Rearm:
 
Fast Rearm is a feature supported by some IMAQ devices that allows the framegrabber to work with cameras that use an excessively short frame-blanking time. Instead of acquiring all valid lines in a frame, only those lines included in the region of interest are acquired. This allows the board to begin rearming itself before the camera finishes sending the current frame. More detailed information follows.

Signals Used in Acquiring an Image
In a normal digital frame acquisition two inputs are used to define valid data. The FVAL signal dictates when a frame is being sent, and the LVAL signal denotes a line of valid pixel data. These inputs are driven by the camera and read by the IMAQ device, and each is used as a gating signal: when the input is active, the data is valid. Most specifications--including the CameraLink standard--use an active high scheme.

Although FVAL and LVAL determine which data sent from the camera is valid, the IMAQ device can selectively choose a region of interest using two internal signals: VROI and HROI. VROI is similar to FVAL in that it defines which lines of video data will be stored to memory. HROI parallels LVAL by specifying which pixels within each line are to be used in an image.

IMAQ Device Behavior During Frame Acquisition
By default, digital IMAQ devices acquire all valid data lines in a frame (see Fig. 1). Acquisition begins when VROI is asserted (usually concurrent with FVAL) and ends when FVAL is deasserted. After each frame is acquired, the framegrabber is given a chance to rearm itself before acquiring the next frame. Rearming involves communicating with the host system via DMA and requires a system-dependent amount of time. (Typically, with versions of NI-IMAQ 3.0 and greater, ~20-30us is required for rearming.)
If FVAL is asserted again before rearming is complete, the framegrabber will not be ready to acquire the incoming frame. It will miss the asserting edge of FVAL and fail to acquire that frame. This process oftentimes repeats itself, so that the acquisition framerate is reduced to half that of the camera's output.

The Effect of Fast Rearm
When Fast Rearm is set (in the camera file), frame acquisition ends with the deassertion of VROI, instead of FVAL (see Fig. 2). Since most acquisitions use a region of interest with fewer than the total number of lines in a frame, this allows the camera to begin rearming itself prior to the end of the frame. (The region of interest may have to be decreased to create a large enough gap for rearming.)
 
I have attached two pictures that show a representation of how fast rearm works.  Unfortunately, I don't have much more information on camera file methods, we normally suggest customers use the Camera File Generator to change functionality of their camera.  Hope all this info helps!
 
Thanks,
 
Scott Savage
National Instruments
Applications Engineering
Download All
0 Kudos
Message 2 of 2
(3,376 Views)