Matlab 2-D color plot without imagesc

22 ビュー (過去 30 日間)
Ongun
Ongun 2015 年 10 月 8 日
コメント済み: Mike Garrity 2016 年 2 月 24 日
Greetings,
I have a 2-D data structured as (x,y,z) where x and y are the coordinates vectors and z is the data vector I am interested in (in my case z is a particular component of the electromagnetic field).
I would like to be able to have a 2-D plot where z values are addressed as colors instead of height so that the plot is not 3-D. Therefore, the appropriate command is not "plot3" and the data is scattered at first so I interpolate it to have a uniform grid. My aim is to have a color plot just like "imagesc" does but since I do not have an array "imagesc" does not work in my case. Which function should I use and is there an elementary solution?
Thanks in advance,
Ongun Arisev
P.S: My working code can be found on the following address: http://pastebin.com/K7gH9C5z

回答 (2 件)

Mike Garrity
Mike Garrity 2015 年 10 月 8 日
What you have is 2D scatter data with colors. Let's make up a simple example and look at your options:
rng default
x = randn(1,1000);
y = randn(1,1000);
c = cos(x) .* cos(y) + rand(1,1000)/10;
The simplest option is to use the scatter command. This will just give you a colored point at each location with no interpolation between them.
scatter(x,y,50,c,'filled','MarkerFaceAlpha',.75)
Your next option, as Image Analyst said, is to interpolate it onto a uniform grid. There are a lot of ways to do this in MATLAB. One good one is scatteredInterpolant.
F = scatteredInterpolant(x',y',c');
[xq,yq] = meshgrid(linspace(-3,3,100));
cq = F(xq,yq);
h = pcolor(xq,yq,cq);
h.EdgeColor = 'none';
As you can see, it's a bit wonky in the areas where it has extrapolated past the edges of the data you gave it. You can tell it not to extrapolate.
And the third option to consider is to create a triangulation.
dt = delaunayTriangulation(x',y');
h = trisurf(dt.ConnectivityList,x',y',zeros(1000,1),c');
h.FaceColor = 'interp';
h.EdgeColor = 'none';
view(2)
There are a couple of other options, but one of those will probably be your best bet. I would probably need to know more details before I could pick the best of those three.
Does that help?
  7 件のコメント
Walter Roberson
Walter Roberson 2015 年 10 月 9 日
'filled' is a new option as well. Leave it out.
Mike Garrity
Mike Garrity 2016 年 2 月 24 日
I cleaned this answer up a bit, and turned it into a post on the MATLAB Graphics blog.

サインインしてコメントする。


Image Analyst
Image Analyst 2015 年 10 月 8 日
How about using imshow()?
  3 件のコメント
Image Analyst
Image Analyst 2015 年 10 月 8 日
Interpolate it to a uniform grid, like you said. Then call it. You can use scatteredInterpolant() then, with the image you get, call imshow(). I don't have any demo for that process but I'm sure a smart engineer like you can figure it out from the examples in the help. Post your code if you need more help with it.
Ongun
Ongun 2015 年 10 月 9 日
Okay thanks for the tips, I will post again if needed.

サインインしてコメントする。

カテゴリ

Help Center および File ExchangeOrange についてさらに検索

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by