MATLAB®

Representing a Matrix as a Surface

Functions for Plotting Data Grids

MATLAB^® graphics software defines a surface by the z-coordinates of points above a rectangular grid in the x-y plane. The plot is formed by joining adjacent points with straight lines. Surface plots are useful for visualizing matrices that are too large to display in numerical form and for graphing functions of two variables.

MATLAB can create different forms of surface plots. Mesh plots are wire-frame surfaces that color only the lines connecting the defining points. Surface plots display both the connecting lines and the faces of the surface in color. This table lists the various forms.

Back to Top

Mesh and Surface Plots

The mesh and surf commands create 3-D surface plots of matrix data. If Z is a matrix for which the elements Z(i,j) define the height of a surface over an underlying (i,j) grid, then

mesh(Z)

generates a colored, wire-frame view of the surface and displays it in a 3-D view. Similarly,

surf(Z)

generates a colored, faceted view of the surface and displays it in a 3-D view. Ordinarily, the facets are quadrilaterals, each of which is a constant color, outlined with black mesh lines, but the shading command allows you to eliminate the mesh lines (shading flat) or to select interpolated shading across the facet (shading interp).

Surface object properties provide additional control over the visual appearance of the surface. You can specify edge line styles, vertex markers, face coloring, lighting characteristics, and so on.

Back to Top

Visualizing Functions of Two Variables

The first step in displaying a function of two variables, z = f(x,y), is to generate X and Y matrices consisting of repeated rows and columns, respectively, over the domain of the function. Then use these matrices to evaluate and graph the function.

The meshgrid function transforms the domain specified by two vectors, x and y, into matrices X and Y. You then use these matrices to evaluate functions of two variables. The rows of X are copies of the vector x and the columns of Y are copies of the vector y.

To illustrate the use of meshgrid, consider the sin(r)/r or sinc function. To evaluate this function between -8 and 8 in both x and y, you need pass only one vector argument to meshgrid, which is then used in both directions.

[X,Y] = meshgrid(-8:.5:8);
R = sqrt(X.^2 + Y.^2) + eps;

The matrix R contains the distance from the center of the matrix, which is the origin. Adding eps prevents the divide by zero (in the next step) that produces Inf values in the data.

Forming the sinc function and plotting Z with mesh results in the 3-D surface.

Z = sin(R)./R;
mesh(X,Y,Z)

Emphasizing Surface Shape

MATLAB provides a number of techniques that can enhance the information content of your graphs. For example, this graph of the sinc function uses the same data as the previous graph, but employs lighting and view adjustment to emphasize the shape of the graphed function (daspect, axis, view, camlight).

surf(X,Y,Z,'FaceColor','interp',...
	'EdgeColor','none',...
	'FaceLighting','phong')
daspect([5 5 1])
axis tight
view(-50,30)
camlight left

See the surf function for more information on surface plots.

Back to Top

Surface Plots of Nonuniformly Sampled Data

You can use meshgrid to create a grid of uniformly sampled data points at which to evaluate and graph the sinc function. MATLAB then constructs the surface plot by connecting neighboring matrix elements to form a mesh of quadrilaterals.

To produce a surface plot from nonuniformly sampled data, first use griddata to interpolate the values at uniformly spaced points, and then use mesh and surf in the usual way.

Example — Displaying Nonuniform Data on a Surface

This example evaluates the sinc function at random points within a specific range and then generates uniformly sampled data for display as a surface plot. The process involves these tasks:

• Use linspace to generate evenly spaced values over the range of your unevenly sampled data.

• Use meshgrid to generate the plotting grid with the output of linspace.

• Use griddata to interpolate the irregularly sampled data to the regularly spaced grid returned by meshgrid.

• Use a plotting function to display the data.

1. First, generate unevenly sampled data within the range [-8, 8] and use it to evaluate the function.

   x = rand(100,1)*16 - 8;
   y = rand(100,1)*16 - 8;
   r = sqrt(x.^2 + y.^2) + eps;
   z = sin(r)./r;
   

2. The linspace function provides a convenient way to create uniformly spaced data with the desired number of elements. The following statements produce vectors over the range of the random data with the same resolution as that generated by the -8:.5:8 statement in the previous sinc example.

   xlin = linspace(min(x),max(x),33);
   ylin = linspace(min(y),max(y),33);
   

3. Now use these points to generate a uniformly spaced grid.

   [X,Y] = meshgrid(xlin,ylin);
   

4. The key to this process is to use griddata to interpolate the values of the function at the uniformly spaced points, based on the values of the function at the original data points (which are random in this example). This statement uses a triangle-based cubic interpolation to generate the new data.

   Z = griddata(x,y,z,X,Y,'cubic');
   

5. Plotting the interpolated and the nonuniform data produces

   mesh(X,Y,Z) %interpolated
   axis tight; hold on
   plot3(x,y,z,'.','MarkerSize',15) %nonuniform
   

   

Back to Top

Parametric Surfaces

The functions that draw surfaces can take two additional vector or matrix arguments to describe surfaces with specific x and y data. If Z is an m-by-n matrix, x is an n-vector, and y is an m-vector, then

mesh(x,y,Z,C)

describes a mesh surface with vertices having color C(i,j) and located at the points

(x(j), y(i), Z(i,j))

where x corresponds to the columns of Z and y to its rows.

More generally, if X, Y, Z, and C are matrices of the same dimensions, then

mesh(X,Y,Z,C)

describes a mesh surface with vertices having color C(i,j) and located at the points

(X(i,j), Y(i,j), Z(i,j))

This example uses spherical coordinates to draw a sphere and color it with the pattern of pluses and minuses in a Hadamard matrix, an orthogonal matrix used in signal processing coding theory. The vectors theta and phi are in the range -π ≤ theta ≤ π and -π/2 ≤ phi ≤ π/2. Because theta is a row vector and phi is a column vector, the multiplications that produce the matrices X, Y, and Z are vector outer products.

k = 5;
n = 2^k-1;
theta = pi*(-n:2:n)/n;
phi = (pi/2)*(-n:2:n)'/n;
X = cos(phi)*cos(theta);
Y = cos(phi)*sin(theta);
Z = sin(phi)*ones(size(theta));
colormap([0 0 0;1 1 1])
C = hadamard(2^k); 
surf(X,Y,Z,C)
axis square

Back to Top

Hidden Line Removal

By default, MATLAB removes lines that are hidden from view in mesh plots, even though the faces of the plot are not colored. You can disable hidden line removal and allow the faces of a mesh plot to be transparent with the command

hidden off

This is the surface plot with hidden set to off.

Back to Top

Line Plots of 3-D Data

Coloring Mesh and Surface Plots