MATLAB: Getting the linear portion of a non-linear curve

gradientlinearMATLAB

Hello,

I am trying to extract the linear section of a non-linear curve. (See picture attached – I would like to get only the section between the red lines). I also attach the data for that curve – it is data obtained from mechanical testing.

I have tried taking the 2nd derivative but so far I have not managed to do this.

Would you have any suggestions?

Thank you.

Salomé

Best Answer

The ischange function may be of use to you. Depending on what you're planning to do with this data the detrend function may also be of interest.

Related Solutions

MATLAB: How to separate a 3D shape with repetitions to each repetition

Try this:

clc;    % Clear the command window.
clearvars;
close all;  % Close all figures (except those of imtool.)
workspace;  % Make sure the workspace panel is showing.
format long g;
format compact;
fontSize = 20;
s = load('matrix.mat')
FE = s.FE
x = FE(1, :);
y = FE(2, :);
z = FE(3, :);
subplot(2, 2, 1);
plot3(x, y, z, 'b.');
xlabel('X', 'FontSize', fontSize);
ylabel('Y', 'FontSize', fontSize);
zlabel('Z', 'FontSize', fontSize);
grid on;
title('Original Single Curve', 'FontSize', fontSize);
numPoints = length(x)
% Find distances from starting point.
distances = sqrt((x - x(1)).^ 2 + (y - y(1)).^ 2 +(z - z(1)).^ 2);
subplot(2, 2, 2);
plot(distances);
xlabel('Index', 'FontSize', fontSize);
ylabel('Distance from point 1', 'FontSize', fontSize);
% Find valleys by finding peaks of inverted signal.
[peakValues, indexesOfPeaks] = findpeaks(-distances);
peakValues = -peakValues; % Invert again.
% Add on very first and very last points
indexesOfPeaks = [1, indexesOfPeaks, numPoints];
peakValues = [distances(1), peakValues, distances(end)];
hold on;
plot(indexesOfPeaks, peakValues, 'rv');
grid on;
title('Distances from point #1', 'FontSize', fontSize);
% Extract out each part of the curve
subplot(2, 2, 3:4);
curveCount = 0;
for k = 1 : 2 : length(indexesOfPeaks) - 1
  index1 = indexesOfPeaks(k);
  index2 = indexesOfPeaks(k + 2) - 1;
  fprintf('Extracting points %d to %d.\n', index1, index2);
  xSingles{k} = x(index1:index2);
  ySingles{k} = y(index1:index2);
  zSingles{k} = z(index1:index2);
  plot3(xSingles{k}, ySingles{k}, zSingles{k}, '-', 'LineWidth', 2);
  hold on;
  curveCount = curveCount + 1;
end
xlabel('X', 'FontSize', fontSize);
ylabel('Y', 'FontSize', fontSize);
zlabel('Z', 'FontSize', fontSize);
caption = sprintf('%d curves', curveCount);
title(caption, 'FontSize', 20);
grid on;

MATLAB: Help with linear fit

Here is one way:

x      = 0:0.05:1;                          % define data
y      = sin(x).^2;
x1     = x(10);                             % define interval
x2     = x(20);
xrange = x(find(x == x1):find(x == x2));    % pick out x and y vals in interval
yrange = y(find(x == x1):find(x == x2));
p      = polyfit(xrange, yrange, 1);        % do linear curve fit 
yfit   = polyval(p, xrange);
figure
hold on; grid; box
plot(x,y,xrange,yfit,'-r')

The first element of p is the slope of the linear curve.

Best Answer

Related Solutions

MATLAB: How to separate a 3D shape with repetitions to each repetition

MATLAB: Help with linear fit

Related Question