Is there a method in order to compute some sort of edge detection / convex hull of a signal x[n] (in orange here). Here of course the orange signal is not convex nor concave, but only piecewise...

(I would like to know if this could be an alternative way of computing the envelope of a signal, alternative to the traditional method input -> abs() -> low pass filter)

enter image description here

  • 1
    $\begingroup$ What do you mean by edge detection? Are you talking about running image processing on a plot of the signal? $\endgroup$
    – nispio
    Oct 30 '13 at 18:35
  • $\begingroup$ @nispio In the 1-D context, the start/stop of a signal is what is usually meant as its 'edge'. $\endgroup$ Oct 31 '13 at 15:15
  • $\begingroup$ @nispio @user4619 I don't really know what I mean by edge detection ;) I used this word in order to speak about the orange curve in the previous example... I wanted to know if there is another method for envelope than input -> absolute value > low pass $\endgroup$
    – Basj
    Oct 31 '13 at 22:01

You can use the Hilbert transform to get the analytic signal. The absolute value of the analytic signal is the envelope. It's available in SciPy and many other software suites, though it's usually poorly named.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.