how mark the highest maximums of a graph?

hello, I try to mark the highest maximums of a graph that I make using data from an excel. but I can't.

the code i use is this:

% Representación Gráfica

plot(t_al,m_al)

xlabel("tiempo"), ylabel("nivel")

datetick('x',10)

grid on

%findpeaks(select,Fs,'MinPeakDistance',0.005)

I get an error with the select function and I don't know what to do

Answer 1

[Steven Lord]

Let's plot some sample data.

x = 0:360;
y = sind(x).*(x-60).*(x-180).*(x-210)/(360.^2);
plot(x, y)
hold on

Identify the local maxima and plot a second "line" highlighting them with circles.

L = islocalmax(y);
plot(x(L), y(L), 'o')

[Victoria Pilar Quesada García]

thanks, but what I need would be for the program to mark the maximums that are greater than a certain "y" or maximums that are at a distance 'x' between them.

would you know how to do the second?

[Bjorn Gustavsson]

You're latest request becomes a very messy task. For that I would definitely write a function. Because now you have to handle selection of maxima in a "not at all obvious manner" what about high peaks that are way above the threshold but a bit close in x, how to handle a slightly lower peak that is just outside the x-separation to the highest peak but within that x-separation to a sub-peak close to the highest one, should that third peak be included or excluded. This is too much hassle to just toss something together on a whim without rather careful and precise specifications.

Answer 2

[Bjorn Gustavsson]

You might get something from a simple solution like this (if your curve is not too noisy):

[m_al_max,idx_max] = max(m_al);
hold on
plot(t_al(idx_max),m_al_max,'rx')

HTH

[Victoria Pilar Quesada García]

thanks, this one works. One more question, if I wanted to mark, for example, the maximums that are above a certain value. how would it be?

[Bjorn Gustavsson]

That is another can of worms entirely - then you're back to the more general solutions provided by the different versions of findpeaks - since you most likely want peaks, but not noisy local maxima. You could possibly get something with:

m_al_threshold = 3.14;
idxM = find((diff(m_al(1:end-1))>0) & ...
            (diff(m_al(2:end))<0) & ...
            m_al(2:end-1)>m_al_threshold) + 1;
plot(t_al(idxM),m_al(idxM),'co')

But this version gladly picks out all noise-peaks and ignore peaks where m_al have a couple of consecutive identical values. (a proper QD-solution - might get the job done, might make your hands dirty.)

[Victoria Pilar Quesada García]

thank you so much !!!

[Bjorn Gustavsson]

You're welcome.