Extracting 2 peak values and their respective location from a set of data points

Question

Jake Bowd 2020-5-15

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/525670-extracting-2-peak-values-and-their-respective-location-from-a-set-of-data-points

回答： Image Analyst 2020-5-15

I am aiming to extract two data points from a waveform within a for loop of n number of repeititions.

I would ideally like it where I can put the range for which I want the two peaks and their locations to be extracted from, i.e. first peak to be extracted from 20-30 and the second peak to be extracted from 60-80. I have included the code I have so far but I feel there must be a much more easier way of doing it:

%% Find Peak Controls
Thold=0.65; %This adjusts the cutoff for the find peaks function 0.65
width=23; %This determines the minimum width between peaks 30
for i=1:nwalks 
[Maxpeaks]=findpeaks(Results.Knee_Add_L_Torque(i).waveforms(1:100),'SortStr','descend');
Threshold=Maxpeaks(1,1)*Thold; %Determines the threshold for the findpeaks function
[Results.Knee_Add_L_Torque(i).MaxPeaks, Results.Knee_Add_L_Torque(i).Peaklocationpercent]=findpeaks(Results.Knee_Add_L_Torque(i).waveforms,'MinPeakHeight',Threshold,'MinPeakDistance',width);
[Results.FirstEKAMMaxpeaks(i), Results.FirstEKAMLocation(i)]=findpeaks(Results.Knee_Add_L_Torque(i).waveforms(23:35),'SortStr','descend');
[Results.SecondEKAMMaxpeaks(i), Results.SecondEKAMLocation(i)]=findpeaks(Results.Knee_Add_L_Torque(i).waveforms(70:89),'SortStr','descend');
end

2 个评论
显示无隐藏无

Mehmed Saad 2020-5-15

编辑：Mehmed Saad 2020-5-15

you want to do it without for loop?

Jake Bowd 2020-5-15

No, I will be using a for loop.

My question is how can i set 2 boundaries to extract data from?

for Results.Knee_Add_L_Torque(i).waveforms I would like to extract the peak between 23-35 and also between 70-89.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Mehmed Saad 2020-5-15

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/525670-extracting-2-peak-values-and-their-respective-location-from-a-set-of-data-points#answer_432622

编辑：Mehmed Saad 2020-5-15

在 MATLAB Online 中打开

I know a workaround(There might be some direct method)

s = sin(2*pi*10*(0:0.01:1-0.01));
findpeaks(s)

Now replace all elements which are not required for peakdetection with NaN. ( i.e. 1:22, 36:69 and 90:100)

ind = 1:100;
s(~(ind>22&ind<36 | ind>69&ind<90)) = NaN;
findpeaks(s)

Edit:

You can also use arrayfun (or cellfun) to directly do that

[peak_s,ind_s]=arrayfun(@(x,y) findpeaks(s(x:y)),[23 70],[35 89],'uni',0);
[Results.FirstEKAMMaxpeaks(i), Results.SecondEKAMMaxpeaks(i)] = deal(peak_s{:});
[Results.FirstEKAMLocation(i), Results.SecondEKAMLocation(i)] = deal(ind_s{:});

where s is equal to your variable Results.Knee_Add_L_Torque(i).waveforms(1:100). You can replace s with Results.Knee_Add_L_Torque(i).waveforms

2 个评论
显示无隐藏无

Jake Bowd 2020-5-15

Thank you for your response.

I have tried this approach and I am receiving an error message of 'cannot call or index into a temporary array'.

I have checked it and it would appear the second section is still trying to call on two peaks, rather than just one.

[peak_Results.Knee_Add_L_Torque(i).waveforms(1:100),ind_Results.Knee_Add_L_Torque(i).waveforms(1:100)]=arrayfun(@(x,y) findpeaks(Results.Knee_Add_L_Torque(i).waveforms(1:100)(x:y)),[23 35],[70 89],'uni',0);

Mehmed Saad 2020-5-15

在 MATLAB Online 中打开

Try this and yes arrayfun will call findpeaks twice

[peak_s,ind_s]=arrayfun(@(x,y) findpeaks(Results.Knee_Add_L_Torque(i).waveforms(1:100)(x:y)),[23 35],[70 89],'uni',0);
[Results.FirstEKAMMaxpeaks(i), Results.SecondEKAMMaxpeaks(i)] = deal(peak_s{:});
[Results.FirstEKAMLocation(i), Results.SecondEKAMLocation(i)] = deal(ind_s{:});

For a single call use the first approach

请先登录，再进行评论。

Answer 2

Image Analyst 2020-5-15

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/525670-extracting-2-peak-values-and-their-respective-location-from-a-set-of-data-points#answer_432647

在 MATLAB Online 中打开

What if you just set the signal outside the ranges of interest to the minimum value and then found the remaining peaks, which would be inside the ranges of interest.

signal = Results.Knee_Add_L_Torque(i).waveforms(1:100);
minValue = min(signal);
signal(1 : 19) = minValue;
signal(31 : 59) = minValue;
signal(81 : end) = minValue;
[Maxpeaks]=findpeaks(signal,'SortStr','descend');