Blur video frame with NI PCI-1411 and JAI CV M-300 camera (urgent problem)

[Veres]

Hello,

I am new with Vision and I need some advice with a problem, which I encountered.

I have a computer with integrated  NI PCI-1411 board, which is connected to a JAI CV M-300 camera. I am measuring a steel part which is positioned in a dark chamber. The part is put on a glass board and is illuminated from underneath with light.

The part is moving. I have to measure this part when it reaches the camera region. I have made a test with not-moving part and everything is perfect. Now, when I move the part I get blur frames from video (with these no part measurement is possible). I need specify, that the camera has no trigger, so it makes video all the time (this video is later inspecting with Vision Builder).

I have tried to change parameters in MAX, but no involvement. I have found no proper icd/file for the camera.

I also could use a SONY DXC-950P camera, if this is better.

How can I eliminate the blur from video frames? Or what am I doing wrong? Is the camera not good enough? Camera + trigger would get better images.

I have attached the video frame with moving object (static) and the blur video frame with moving object (dynamic).

Please give me a answer or ideas as soon as possible.

Thanks.

[BruceAmmons]

You are using an interlaced camera.  It takes half the picture (every other line), waits a split second, then takes the other half of the picture (the other lines).  When it puts the two images together, you can see the two positions it was in when the images were taken.

The best solution is to use a non-interlaced or progressive scan camera, which takes the entire image at one time.  The Sony might be this type of camera, since it has a P in its name.  I didn't look up the specs.

If you can't get the right kind of camera, your only other option is to use fields instead of frames.  One field is the half image taken at one time.  The image will be half the height of the original image, though, which can mess things up.

Another possibility is the reduce the resolution of the image.  If you take every other pixel in both directions, you will have a very clear image that is half the size of the original.  You can do this with Extract in LabVIEW, not sure if it is available in VBAI.

Bruce

[AnalogKid2DigitalMan]

Maybe you should review the manual, link here:

http://1stvision.com/cameras/JAI/CV-M300.html

The camera does indeed have a trigger and you can set your exposure times and other paramters via DIP switches.

I would try using a high speed strobe for backlighting that also triggers the camera using a short exposure.

-AK2DM

[Veres]

Thanks for the information's so far.

Does anyone know how to resize/scale the acquired frames without distortion on width-direction?

Does anyone have a sub-VI I can integrate in Vision Builder? I would really appreciate it.

Thanks.

[BruceAmmons]

The ratio is exactly two to one.  Your options are to double the resolution in one direction using IMAQ Expand, or eliminate every other pixel in the other direction using IMAQ Extract.  For measurements and the like, you might be better off leaving it as is and just remembering to double the Y measurements.  If you are measuring diagonally, that won't work so well.

You would probably be better off getting a progressive scan camera (or a digital camera, which are pretty cheap nowadays.)  The image quality will be better with a digital camera, also, because the analog signals are converted onboard the camera instead of travelling through cables.

Bruce