Fill in region between data points

Hello all,

Trying to shade a region between my data points. The methods I've tried using don't quite work. I've tried convhull() and boundary() with a shrink factor and can't get the bounded region I'm looking for.

The lower edge of my points get lost in the boundary. See the image. I don't want the blue area in the fill.

Likewise if I had another set of points (in the red) I would want it to follow the top and bottom edge and not try to shortcut the bottom edge.

Any help would be great. Heres my code. The variables x and y are [M x N] N being each set of points on a particular curve.

 xx = reshape(x,[size(x,1).*size(x,2),1]);
 yy = reshape(y,[size(y,1).*size(y,2),1]);
 k = boundary(yy,xx,.1);
 fill(xx(k),yy(k),C(ii,1:3))
 plot(xx,yy,'ko','linewidth',2) 
 % also tried this: -----------------------
 k = convhull(x,y);
 fill(x(k),y(k),C(ii,1:3))
 plot(x,y,'ko','linewidth',2)

[KSSV]

We need points to work on.....attach your data..

[Ryan]

Hi KSSV, just attached the x and y data.

Answer 1

[Star Strider]

Try this:

figure
plot(x, y, 'o')
hold on
patch([x(:,1)' fliplr(x(:,3)')], [y(:,1)' fliplr(y(:,3)')], 'b', 'FaceAlpha',0.3)
hold off

I first ran:

[colmax,mxidx] = max(y);

to determine which columns were the upper and lower curves, and used those results in the patch call. The code then takes columns 1 and 3 of ‘x’ and ‘y’, turns them into row vectors and flips column 3. This defines the area for patch to fill, since it needs a closed area. See the documentation on patch (link) for details.

Experiment to get the result you want.

[Ryan]

Hi Star Strider,

That works well for data that doesn't cross each other (see my second image I posted), but I think I see what I need to do. I need to find the upper and lower surface and then use patch. Thanks for the help!

[Star Strider]

As always, my pleasure!

To use patch to create the second image, The idea would be to interpolate the original data to create a common grid for both, get the maxima and minima to define the limits of the patch object, and then plot those.

Example —

x1 = sort(rand(1,30));                                              % Create Data
y1 = 1 - 5*(x1-0.5).^2;                                             % Create Data

x2 = sort(rand(1,40));                                              % Create Data
y2 = sin(2*pi*x2);                                                  % Create Data

xi  = linspace(min([x1(:); x2(:)]), max([x1(:); x2(:)]), 50);       % Interpolation Vector
y1i = interp1(x1, y1, xi, 'linear','extrap');                       % Interpolate ‘y1’
y2i = interp1(x2, y2, xi, 'linear','extrap');                       % Interpolate ‘y2’
upper = max([y1i; y2i]);                                            % Upper Limit
lower = min([y1i; y2i]);                                            % Lower Limit

figure
plot(x1, y1, '-p',   x2, y2, '-p')
hold on
patch([xi fliplr(xi)], [upper fliplr(lower)], [0.5 0.2 0.1], 'FaceAlpha',0.2)
hold off

You will have to experiment with this approach with your own data.