Multiple Valley Detection Query

[TriboBoy]

Hello There,

I am having trouble detecting multiple valleys within a bell curve type shape (see pic). The valleys are moving horizontally and I need to detect each of their positions in real time. The other issue is that the dips increase and decrease in magnitude thus making it difficult to use the dip detect function as it keeps jumping from one to the other.

any help would be greatly appreciated,

thanks.

[odessy27]

Are you using peak detector vi?

[Marshall_B]

Hi TriboBoy,

From what I understand you are trying to detect the valleys of your bell shaped curve as and when they arise.  What have you tried so far?

If you have not done so already, you may want to take a look at the examples in the example finder.  While in LabVIEW go to Help>>Find Examples... and search for "Valleys" in the search tab.  This will bring up 3 examples for finding peaks/valleys of a waveform.  I would suggest taking a look at the "Advanced Peak Detector.vi" example.

I hope this helps but feel free to post any more questions you may have.

Regards

[TriboBoy]

Hello,

Yes I have been using the peak detection.vi and have gone through the examples, but as the valleys keep moving left and right and up and down, the thresholds keep changing and the output from the peak detector keeps jumping.

So far I have created an idealised waveform and divided through, getting rid of the curve. I have then inverted the result and squared it giving nice sharp peaks. But I am still having trouble detecting the peaks as they are moving around too much.

Thanks again for the help

[Marshall_B]

Hi TriboBoy,

Have you tried setting the threshold level above your data so that the algorithm used within the Peak Detection and Display VI considers all of the valleys occurring below it?

Below is a screen shot of the resulting graph using the Peak Detection and Display VI example in Example Finder:

The blue threshold line corresponds to the valley data points being considered and has been set to a value higher than any expected data points.  

It is also worth noting that the width used should be as low as possible (limited to a value of 3) so that the maximum number of data points are considered.  

Let me know if this helps you at all.

Regards

[Marshall_B]

Hi TriboBoy,

Have you been able to track the locations of all the valleys in your waveform using the information I provided?

Regards