I'm using a cell-averaging constant false-alarm rate (CA-CFAR) detector to detect peaks like the first one in the following plot:

enter image description here

The magenta line is the adaptive threshold, which is the local standard deviation multiplied by a contrast factor. What other peak-detection techniques could I try in this context?

  • $\begingroup$ Well, what have you tried and what doesn't work? $\endgroup$ – panthyon Oct 19 '15 at 16:49
  • 2
    $\begingroup$ Yes, generally "fishing" questions like this tend to be closed as too broad. Can you give a little more information about what you'd prefer to happen? If there's a specific issue you're trying to get around, that may help us to help you. CA-CFAR seems like a good approach, without any other considerations to take account of. $\endgroup$ – Peter K. Oct 19 '15 at 18:17
  • $\begingroup$ I have tried using a constant threshold and an adaptive one (CFAR). Constant thresholds don't work very well under unknown (and varying) interference levels, which is why I looked into CFAR. I was wondering whether I am missing any other standard detection technique. $\endgroup$ – geschema Oct 20 '15 at 8:05
  • $\begingroup$ It would be lot more effective and likely to be answered if a more detailed, mathematical, description were generated. For instance: what is a "peak"? Is it the true absolute peak in a group of numbers? Or a data point that is surrounded by 5 (?) lower points on each side. Typically a mathematical statement of the problem is necessary before an analysis can be done. Are there outliers to be tossed? Is this a signal+noise? What generates the noise? Do the "peaks" correspond to fractures in a window or variances of manufacturing machinery? $\endgroup$ – rrogers Oct 21 '15 at 14:37

Your Answer

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

Browse other questions tagged or ask your own question.