Bug or feature: IMAQ Particle Filter 3 with ROI connected erases particles outside ROI

[comrade]

Hi,

I just ran into something odd. When using the ROI Descriptor connector on the IMAQ Particle Filter 3 (using LV2013) particles lying outside the ROI are erased no matter what the filter criteria are. The particles within the ROI are scrutinized and accordingly kept or erased. This is regardless of the "Keep/Remove" connector value.

As the help states that the connected ROI "defines the Region of Interest (ROI) within which the particle filter is performed" I would expect the function to keep all particles in the image that are outside the connected ROI untouched. Have a look for yourself with the attached VI.

Am I barking up the wrong tree here?

[annav]

Hi comrade,

I am not sure, what you mean. I have executed your example (thanks for that) with LV 2016 and I don't exactly know what you expect. I have attached two screenshots of what I set and what I get. Do you mean something else?

Best,

Anna

[comrade]

The left ROI constant points to an area around the upper left rectangular particle (the larger one). The 2nd ROI constant covers an area of the image with no particle.

The point is, that with either ROI definition I would expect the lower right particle to survive the filter as neither ROI engulfs it.

However this can only achieved by not setting an ROI at all.

The help text concerning the ROI connector for this function states (as I understand it) that the particle filter is subjected ONLY to the area within the ROI. So it should leave any particle OUTSIDE of the ROI untouched. But it does not. It seems to filter all particles outside the ROI regardless of them fulfilling or not fulfilling the defined filter criteria. Only within the ROI the criteria become active and only particles meeting the criteria are treated as defined by the function.

[annav]

HI,

thanks for your answer.

I tried several things and I think it is indeed a bug. The ROI has the purpose of defining, as it says, a region of interest and not a mask. I have reported this bug.

Best,

Anna

[Jacobson-ni]

This is reported as CAR #616904 for those wishing to track this in the future.

[comrade]

Thanks for quickly addressing this.