Vision system Real-time PXI or cRIO with CVS-1456

[tiho_bg]

Hi everybody

I would like to use hardware from NI. My task is to control piezo-actuators, step-motors and digital camera. My question is the following:

Could you explain the advantages and disadvantages of PXI module (2.2GHz) with integrated camera module (1394 firewire) and cRIO-9074 module with NI CVS-1456 Compact Vision System module. I would like to work in real-time mode. I'm interrested in the speed of vision system because one of the modules is with high speed and integrated vision system but the other cRIO will be connected with CVS-1456 via ethernet port. Which solution is faster in real-time mode.

Thank you in advance

[Matt_S.]

Hi Tiho_bg,

Both systems have advantages.

The PXI system will provide you higher speed processing if it were only doing machine vision processing. If you start to add other processing to what the PXI controller is doing (like calculating motion trajectories, etc.) then you may see lower performance on the PXI controller than you would see on the NI CVS 1456 that would only be doing machine vision and communication back to the host (the cRIO in your setup). It would need to be a good deal of 'extra processing' you are doing though, the PXI controller you mention is much higher speed than the NI CVS 1456.

The PXI system would also allow for some extremely tight synchronization with other boards you add to the system (including the motion controller you plan to use).

Regarding advantages of the cRIO and the CVS, the main advantage is their rugged nature. These systems are both fanless, and have no moving parts. For many industrial applications this is a requirement. Also, the CVS provides an FPGA for digital I/O so you can synchronize lighting, triggers, and your camera.

A few questions for you.

A) What is the resolution, frame rate, and pixel depth of this camera? If it is a VGA camera with 8-bit pixel depth at 30 fps, then either system could easily acquire the images.

B) What image processing are you doing? Filters, convolutions, and other basic functions are fast on both of these systems. Pattern recognition, optical character recognition, and code reading tend to be slower algorithms and may require a faster processor if you're using high resolution images or have a fast frame rate.

I hope this helps, and please feel free to ask for any clarification if it is needed.

Cheers,

-Matt S.