Issue with finding the peaks of a graph.

I would like to plot peaks on the autocorrelation graph. however, the findpeaks function does not give me the peaks of the autocorrelation graph. Is there something wrong with my code?

I have a histogram that looks like this:

I use this histogram to plot the autocorrelation:

and when I use the findpeaks function to find and plot the peaks, I get this:

my code is:

MINval = min(CODCT);
MAXval = max(CODCT);
width = 1;
h=hist(CODCT,floor(MINval):width:ceil(MAXval)); %histogram
x = h;
Rxx=autom(x); %autocorrelation
figure()
grid;
[pks,locs] = findpeaks(Rxx);
plot(Rxx); 
hold on
plot(Rxx(locs),pks,'k^','markerfacecolor',[1 0 0]), hold off

Answer 1

[Star Strider]

I cannot figure out what you are doing, and specifically what ‘autom’ returns.

Try this instead:

plot(locs,pks,'k^','markerfacecolor',[1 0 0])

[Amy Wong]

It works! however, do you know why is the first peak not located?

autom is the function to find the autocorrelation

[Star Strider]

Great!

The first ‘peak’ is not actually a ‘peak’ as findpeaks defines peaks. It is instead a global maximum (of which there is only 1 in your data), so if you want to append it to the findpeaks output, this will probably work (I say ‘probably’ because I cannot test it with your code):

[pks,locs] = findpeaks(Rxx);  
[maxRxx,idx] = max(Rxx);
pks = [maxRxx; pks];
locs = [idx; locs];

NOTE — This assumes ‘pks’ and ‘locs’ are column vectors. If they are row vectors, replace the semicolons (;) with commas (,). The max function returns the first occurrence of the maximum value in a vector, and its index if requested. Since you only have 1 maximum, it will return the value and index. Concatenating those values with your existing ‘pks’ and ‘locs’ vector should give you all of them.

[Amy Wong]

It works! however, I do not understand what you are doing. Can you explain to me again?

[Star Strider]

I will do my best.

The findpeaks function returns the values and locations that it identifies as peaks. Its criteria are the ability to determine the derivative of the input waveform so it can identify the peak. In your situation, it cannot identify the maximum, since it starts at the maximum rather than being a ‘peak’ with respect to the values on each side of it. The solution to identifying it is that since it is the maximum, use the max function with 2 outputs, the first being the value of the maximum, and the second being the index if the maximum. Since your findpeaks call requests the values and indices of the peaks, it is then straightforward to concatenate the value and index of the maximum to those vectors to provide all the information you want.

For a full explanation, see the documentation on the max (link) function.

[Amy Wong]

I understand. Thank you for your detailed explaination!

Do you know any other method to find the values and locations of the peaks instead of using the findpeaks function?

[Star Strider]

As always, my pleasure.

There may be some peak-finding routines on the File Exchange. I have not used them, so have no experience with them.

Answer 2

[Steven Lord]

If the issue to which the title of your question refers is the X ticks of the last plot, that's because you called plot with one data input.

Calling plot(q) plots points (1, q(1)), (2, q(2)), (3, q(3)), ....

Calling plot(q, w) plots points (q(1), w(1)), (q(2), w(2)), (q(3), w(3)), ....

If that's not the issue about which you're asking with which you'd like us to help you, please state that issue clearly.

[Amy Wong]

I have edited my question.

My issue is i do not understand why the peaks does not match my graph