peak/valley detection problem

[teeeeg]

Hi,

How can i detect the maximum peak to valley on the graph attached.  It is not the difference between the maximum and minimum all the whole data set that i require, I want to know the maximum value of the higher frequency peaks and valleys ignoring the small 'blips' within the data.

The X-axis is controlled by time and the number of samples between each peak varies between data sets.  I have considered using a moving average to smooth the data but cannot do this due variation of samples between peaks.

Using the peak/valley detection VI would work great but due to the small 'blips' in the data set a true maximum value cannot be found.

Can anyone help?

Tom

[Davidek]

Hi Tom

I'm not really sure which values you are after. Could you indicate which peaks/valleys you are interested in in the image you attached?

Best Regards

David

NISW

[teeeeg]

The value i want to know is the difference between each of the coloured lines, ignoring any 'blips' inbetween.

I then can easily find the max value.

Some of the blips have been circled.

Thanks

Tom

[Harold_Timmis]

so you want to find the maximum point on the graph?

because if thats the case you can do this...

Message Edited by Harold Timmis on 10-22-2009 11:05 AM

[aeastet]

Do you have the raw data in a format that we can use? I have a program that I think will work for what you are trying.

[teeeeg]

Please find data attached.  The VI contains a constant with data and a graph.

[Christian_Hartshorne_NIUK]

Hi Tom,

I have been looking into your issue in an attempt to find you a solution. My initial thoughts are that you should apply some form of smoothing to the waveform before subjecting it to the the Peak Detector sub VI to detect the amplitude of the peaks and the troughs. These amplitudes could then be subtracted form each other thereby arriving at the values you are looking for. Can I confirm that this would solve your problem.

I have been looking into a methodology for smoothing your waveform and I found the following example which uses a moving average technique, which I think could be useful. Take a look and let me know what you think.

Best Regards,

Christian Hartshorne

Applications Engineer

National Instruments UK  

[teeeeg]

Hi,

Could you please post the moving average VI in version 8.5?

Thanks

Tom

[Christian_Hartshorne_NIUK]

H Tom,

Here is the Moving Average VI in LabVIEW version 8.5. Hope this helps.

Regards,

Christian Hartshorne

Applications Engineer

National Instruments UK