Curve at the peak values

For the evaluation of measurement results, a curve should be laid through the maximum values (polynomial or spline). Is there a direct possibility here or must the maximum measured values be read out separately?

Many thanks!

[dpb]

findpeaks should do just fine with such a base spectrum to return the peaks, locations "automagically" for the fitting routine.

[dpb]

[OP's Answer moved to Comment...dpb]

Thanks, with

[pkt,lct] = findpeaks(ChannelE,Frequency,'Threshold',0.8);

this works quite well. What's the best way to create a fitting curve through the points? With polyfit or lsqcurvefit I was not very successful...

Many thanks!

[dpb]

A smoothing spline would likely be the ticket there -- or a N peak mixed Gaussian or somesuch depending upon whether there is some sort of expected shape based on whatever it is the data represent and what the bounding curve is for--altho there's a dichotomy in that the two major peaks with the example "fit" are treated as one while the others with similar spacing are separated. The only difference appears to be the two are of essentially the same magnitude.

Answer 1

[Mepe]

Thanks, with

[pkt,lct] = findpeaks(ChannelE,Frequency,'Threshold',0.8);

this works quite well. What's the best way to create a fitting curve through the points? With polyfit or lsqcurvefit I was not very successful...

Many thanks!