Cleaning noisy data from a plot
显示 更早的评论
Hi;
I have a regular linear plot from a sensor, but there are some spikes that shouldn't exist actually. I want to eject points that cause these spikes and interpolate the interspaced part. Is there any apps or tools that will help me on it, or shall I build an algorithm which will find these points and eject them?
Thank you in advance.
采纳的回答
Star Strider
2015-8-14
If you have the Signal Processing Toolbox, the findpeaks function (with appropriate name-value pair argument choices and parameters) could help you identify the spikes and their locations.
Keep your original independent variable vector, since you’ll need it for the interpolation. Then use the location index vector provided by findpeaks to eliminate them. Then interpolate.
To illustrate:
x = 1:100; % Create Data
y = 1 + 0.1*x.*rand(1,100); % Create Data
spikes = randi(100, 1, 5); % Create Data
y(spikes) = randi(50, 1, 5); % Create Data
[spk,loc] = findpeaks(y,'MinPeakHeight',10); % Find ‘spikes’
xorg = x; % Save Original ‘x’
yorg = y; % Save Original ‘y’
x(loc) = []; % Eliminate ‘spikes’ From Data
y(loc) = []; % Eliminate ‘spikes’ From Data
yint = interp1(x, y, xorg); % Interpolate Missing Data
figure(1)
subplot(2,1,1)
plot(xorg, yorg)
grid
title('Original Data')
yl = get(gca, 'YLim')
subplot(2,1,2)
plot(xorg, yint)
grid
title('Interpolated Data Without Spikes')
set(gca, 'YLim',yl)
更多回答（0 个）
类别
在 帮助中心 和 File Exchange 中查找有关 Data Import and Analysis 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!
Translated by 
