Missing peaks in waveform peak detection.vi

[Jaspal]

I am also attaching my VI here. It is dirty and messy and I am sure many of you will identify better ways of simplifying it.

I am happy to take any suggestions onboard please.

Jas

[ZJ.Gu]

Hi Jaspal ,

I have a look at your data samples. I have a try like this (detrend->peak detection with a suitable threshold). 

Zhijun Gu

[Jaspal]

Hi Zhijun,

Thank you for the post. It indeed is the way forward. I did implement the above in the real time data acquisition and WA detrend and WA Multiscale peak detection doesnt seem to work.

I am sure Ihave not set some parameter righ. I did have a look at the detailed help and seem to be missing something here. Can you please suggest what is going on. I will paste a jpg of the code to give na idea of what I am doing.

I can detrend the signals using filters and detect peaks using waverforms peak detection. But would like to implement Wavelet peak detection beacuse of its obvious advantages.

Thanks again,

Jas

[ZJ.Gu]

Hi,

Could you pleast attach the VI here?

BTW, what you are doing now seems to be "Heart Rate Variability Analysis".  NI has released a startup kit named "Biomedical Startup Kit 2.0" which includes HRV Analysis.

You could refer to it : http://zone.ni.com/devzone/cda/tut/p/id/9037

http://decibel.ni.com/content/docs/DOC-4754

[Jaspal]

Hi Zhijun,

Yeah, you are right. I am implementing HRV analysis and using this information to setup a test sytem for biofeedback analysis.

I am attaching the VI here.

[ZJ.Gu]

Hi,

It is difficult to say which parameter suitable for your use case. From you VI, I couldn't find anything wrong. I could not run you VI because of lack of some subVIs.

Could you please attach your data samples which don't work with your current settings?

[Coq_Rouge]

 I hope you did not learn that programming style in a Labview class. Your program may work. But this is not the proper way to do it

[Jaspal]

Hi Zhijun,

the data samples are the same as the one posted earlier.That is the general trend of the signal. The WA Multiscale Peak detection VI seems to detect peaks but the timing information of the raw data and peak data is not right. (If you see the peak plot time scale on front panel)

I am attaching the sub VI,s. not sure if that helps.

Message Edited by Jaspal on 06-30-2009 05:00 AM

Message Edited by Jaspal on 06-30-2009 05:01 AM

[Jaspal]

Hi Coq Rouge,

You are right. I did. I so knew that Labview experienced would look at my code and go, what is this!!!

I have just thought of the flowchart and implemented as I went along. Didn't really think about timing and so many other details.

Would you suggest me to start somewhere or point me to some information that I can use to improve what I am doing here.

Thanks

Message Edited by Jaspal on 06-30-2009 04:59 AM

[ZJ.Gu]

Hi,

The previous data samples could work with WA Multiscale Peak Detection. What you said about the time info may be originated from filter. Filtering will result in some time shift between original raw signals and filtered signals.

I don't know what's the problem you ran into?

From the screenshots I attached above, the peak locations are right.

帖子被Zhijun Gu在 06-30-2009 05:02 AM

时编辑过了