When you have a plotted signal, how can you isolate one part of it, or more parts?

When you have a plotted signal, how can you isolate one part of it, or more parts?

Here is the signal I have, i need to isolate the peak:

Answer 1

[Mark Sherstan]

Use the findpeaks function. Post your data if you need more help with the function.

[Valentina Malconi]

I can't use findpeak function unfortunately.

Also, how can I identify the two lowest points right next to the peak? So where the peak starts and where it ends.

they're not the minimum points of the function so I can't find them that way.

[Mark Sherstan]

Can you use the islocalmin and islocalmax functions? Post your .mat file so we can help you more than just giving a list of functions.

[Valentina Malconi]

This is the signal, we need to isolate the peak.

Thank you very much

[Mark Sherstan]

Here is the peak isolated:

load('signals.mat')

% Find max point
[Y idx] = max(Signal);
x = 1:length(Signal);

% Plot results
figure(1)
hold on
plot(x,Signal)
plot(x(idx),Signal(idx),'*r')
legend('Signal','Max')
hold off

Can you not use the find peaks function because of tool box limitations or because it is an assignment question?

[Valentina Malconi]

Thank you very much.

I have another question.

I have to remove the peak from the signal because it is noise, and to do that I have to mark the two points where the peak starts and ends.

Any idea how to do that? :)

[Mark Sherstan]

Why cant you use the find peaks function? This would solve all your problems.

[Mark Sherstan]

load('signals.mat')

% Find max point
[Y idx] = max(Signal);
x = 1:length(Signal);

figure(1)
hold on
plot(x,Signal)
plot(x(idx),Signal(idx),'*r')
legend('Signal','Max')
hold off

% Find all the min values and there index
idxMin = islocalmin(Signal);
idxMin = find(idxMin);

% Find the min values closest to the spike and set to nan
leftCutoff = idxMin(idxMin<idx); 
rightCutoff = idxMin(idxMin>idx);

Signal(leftCutoff(end):rightCutoff(1)) = NaN;

figure(2)
plot(x,Signal)

[Valentina Malconi]

Thank you very much!

[Mark Sherstan]

If this works for you please accept the answer so that it can also help out other people in the future.