Slice plot of a 3D-data set with color code

Hello,

I have a rather simple question - tried to reread other posts but it simply doesnt work out so far..

I have several thousands of points in a 3D field - each represented by X,Y and Z value. Each point is described by a fourth number. Coloring should be according to the value of the fourth variable.

I have done a 3D-scatter plot but chances are I miss out interesting things in the middle that are blocked out of my view.

Hence, I would have to do a compilation of slices to visualize the inside. Slice planes should go from the bottom of the 3D-figure to the top.

I know of the slice-function in matlab. But my input arguments produce errors I cannot interprete.

Can someone provide a sample code? Sadly, I could not find one before.

I know that it should be something like -> slice (X,Y,Z,V, PlanevectorX, PlanevectorY, PlanevectorZ)

what would be V in my case?

[Arne]

I have thought about this question a bit. Maybe it might be the easiest thing to separate the matrix into several smaller matrices according to their Z-value. For example Z ranges from -428 to -152. Then I divide the giant matrix into smaller ones; first one with Z-values from -425 to -400, next one with Z-values from -400 to -375, and so on.. And then actually plot the scatter3 plots of those - I guess I can separate the matrix with histc but I only did this for vectors.. any idea how to histc a matrix according to a column and then de-index them to get the small matrices?

Best

Answer 1

[Adam]

I'm not familiar with the slice function, although it does look very useful for the work I do also, but I'm pretty sure that V should be your colour data, of the same size as the X, Y and Z data which should be provided using meshgrid e.g.

[X,Y,Z] = meshgrid(-2:.2:2,-2:.25:2,-2:0.1:2);

The parameterisation of meshgrid should match your sampling to give you X, Y, and Z matrices whose dimensions match your colour values.

[Adam]

To reply to your answer below which should be a comment on this answer rather than its own answer ideally...

In my example above you should replace those hard-coded numbers with your x, y and z vectors. Often naming convention seems to just be to use capitals for the meshgrid equivalents so e.g.

[X, Y, Z] = meshgrid( x, y, z );

I'm not sure what the simplest way is to get your V vector into the same form, but each element in the 3d arrays X, Y, and Z will correspond to exactly one triple (x,y,z) in your original vectors.

Do you have a V value for every single (x,y,z) triplet? If so you should just be able to do a reshape or reshape and permute to get V matching up with X, Y, Z

[Arne]

Thank you. But it is not totally clear to me. I have four vectors from my data. X-coordinates, Y-coordinates, Z-coordinates and the property vector (colour vector). The x,y,z you mentioned have nothing to do with my X,Y,Z I guess? It only establishes a grid of a defined dimension in meshgrid??

Yes, I also think V should be the colouring property. Yet, I cant use it in slice as a vector. Matlab tells me V should be a 3d-array. I have no idea how to transform my vector into a reasonable 3d-array.

If your x,y,z has nothing to do with my coordinate-vectors, I also wonder where the function will receive the necessary spatial information from.

[Adam]

The x, y, z in my comment above are now the same as your x, y, z. I changed naming convention from my original answer just to match the way the help names things.

X, Y and Z in my comment answer are the meshgrid outputs that I named x, y, z originally.

(I have now edited my original answer to use consistent variable naming)

[Arne]

Thank you! I understand a little bit more now - I tried to create the meshgrid like you suggested:

[X,Y,Z] = meshgrid(my_x,my_y,my_z);

Unfortunately, I get an error message called: Error using repmat Out of memory. Type HELP MEMORY for your options.

Looks like the vecotors I have are too large for this operation (31147 lines).

Answer 2

[matt dash]

Answer to your updated comment: Use the second output (ind) of histc, which is the index of which bin the rows are in. Then use this to plot subsets of the data.

Example: see how data is binned according to the 4th column (which corresponds to color in the scatter plot)

data = rand(10000,4)
[counts,ind]=histc(data(:,4),[0 0.25 0.5 0.75 1]);
for i = 1:max(ind)
 figure;
 axes;
 idx = ind==i;
 scatter3(data(idx,1),data(idx,2),data(idx,3),16,data(idx,4),'filled')
 set(gca,'clim',[0 1])
end