Plot envelope of a grassy plot

9 visualizzazioni (ultimi 30 giorni)
André
André il 11 Giu 2016
Commentato: André il 14 Giu 2016
Hi all, I have a graph which looks very grassy. What I want is the following (see picture). I have a grassy graphy (blue line), but I want the envelope which clearly shows the peaks of the graph (red line - this line is drawn in paint). A group of peaks may added to 1 group of peak. I have tried the following in matlab, without success: - smooth; - basic tool of curve fitting; - Hilbert transform; - polyfit; - Toolbox Ezyfit; butterworth function.
The problem is that I either get too many peaks, or the envelope does not really fit.
Any suggestions how to deal with this? Thanks in advance!

Risposte (3)

Chad Greene
Chad Greene il 11 Giu 2016
Do you have the Signal Processing toolbox? If so, the envelope function makes is pretty easy to use.
  1 Commento
André
André il 11 Giu 2016
Thanks for your answer. I tried the envelope function, also with 'peak' to detect the peak envelope. However, it is still too grassy ..but it is sort of in the right direction ..

Accedi per commentare.


Image Analyst
Image Analyst il 11 Giu 2016
Another option is sgolayfilt(). They all give different looks and it just depends on what "look" you're looking for.

Image Analyst
Image Analyst il 11 Giu 2016
Maybe you can use findpeaks(). There are lots of options for specifying which peaks you want to detect and which you want to ignore. Then, once you've found the peaks, use linspace() to draw lines in between the peaks to fill out the array so that you have a value at every index.
  6 Commenti
André
André il 13 Giu 2016
I tried the following with findpeaks():
  1. findpeaks(y); because I want find the peaks of y and then proceed as you suggested
  2. findpeak(x,y); then I get the error message as above. In addition, I did not provide an X (only x and y as the file attached).
For point 1, I am able to detect the peaks, but there is a shifting in the x-axis and y-axis.
How to proceed from here?
André
André il 14 Giu 2016
I also tried stem() instead of plot() and then envelope, but without success still.

Accedi per commentare.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by