How can I convert a file format FIG to PCD?

Hi.

I'm new to matlab. I did the following to edit point cloud data.

ptCloud=pcread(filename)
pcshow(ptCloud)

After editing, the data is saved as a FIG file, how can I convert it to a PCD file?

Answer 1

[Benjamin Thompson]

There is a function called pcwrite. The MATLAB documentation has a good article on it.

[Walter Roberson]

fig files contain the complete information about the graphics objects, including the size of the figure, the position of the axes, the format to be used for the axes tick marks. Most of that information cannot be saved to pcd. If you are working with a fig file you need to findobj() the pointCloud from the figure, and extract the data from it, and pcwrite()

[Motoki Hatsuda]

Thank you Walter and Benjamin.

I want to save this point cloud data as pcd files. I'd like to know the specific code.

[Walter Roberson]

Untested.

fig = openfig('NameOfFig.fig');
outnamebase = "pcout_";
pcviewers = findobj(fig, 'tag', 'pcviewer');
for K = 1 : length(pcviewers)
    x = pcviewers(K).XData;
    y = pcviewers(K).YData;
    z = pcviewers(K).ZData;
    C = pcviewers(K).CData;
    xyz = [x,y,z];
    ptc = pointCloud(xyz, 'Color', C);
    outname = outnamebase + K + ".pcd";
    pcwrite(ptc, outname);
end

This writes all point clouds that are in the figure. It has to reconstruct them as fig files store Scatter objects rather than pointCloud objects.

[Motoki Hatsuda]

Sorry for the late reply.

I ran the code and got the following error. Do you know how to solve this problem?

The first error statement reads "Invalid argument for the position 1. The size of oh the input 'xyzPoints' must be an M-by-3 list of points or an M-by-N-by-3 array for an organized point cloud."

[Walter Roberson]

xyz = [x(:), y(:), z(:)];

[Motoki Hatsuda]

The first error statement reads " 'Color' must correspond to the number of input points. "

[Walter Roberson]

Could you show size(x) and size(C)?

I have seen some cases where the XData was a 2d matrix instead of a vector, but I have not figured out how that could be, as all the pointCloud code I have looked at expects it to be a vector.

What is class(C) and min() and max(C)?

[Motoki Hatsuda]

Thanks for the quick reply.

How can I check? Do I type in the code or check the fig file?

[Walter Roberson]

put a breakpoint on the line

      ptc = pointCloud(xyz, 'Color', C);

and run the code. When it gets there, in the command window type in

whos x C
min(C(:)), max(C(:))

and show us the results

[Motoki Hatsuda]

OK?

[Walter Roberson]

Try this

fig = openfig('NameOfFig.fig');
outnamebase = "pcout_";
pcviewers = findobj(fig, 'tag', 'pcviewer');
for K = 1 : length(pcviewers)
    x = pcviewers(K).XData;
    y = pcviewers(K).YData;
    z = pcviewers(K).ZData;
    C = pcviewers(K).CData;
    xyz = [x(:),y(:),z(:)];
    if size(C,2) == 1
        ptc = pointCloud(xyz, 'Intensity', C);
    else
        ptc = pointCloud(xyz, 'Color', C);
    end
    outname = outnamebase + K + ".pcd";
    pcwrite(ptc, outname);
end

[Motoki Hatsuda]

Wow! You are amazing!

I have one question.

I made a pcd file after deleting point clouds, but looking at the contents of 'pcout_1.pcd', there are still 1,832 point clouds.

Does this mean that point clouds disappear visually, but point clouds remain in terms of internal information?

[Walter Roberson]

There are 1832 points in one point cloud.

You might want to change the colormap to highlight more of the points. There are a lot of dark blue points in the front right.

[Motoki Hatsuda]

OH, I see.

fig = openfig('NameOfFig.fig'); outnamebase = "pcout_"; pcviewers = findobj(fig, 'tag', 'pcviewer'); for K = 1 : length(pcviewers) x = pcviewers(K).XData; y = pcviewers(K).YData; z = pcviewers(K).ZData; C = pcviewers(K).CData; xyz = [x(:),y(:),z(:)]; if size(C,2) == 1 ptc = pointCloud(xyz, 'Intensity', C); else ptc = pointCloud(xyz, 'Color', C); end outname = outnamebase + K + ".pcd"; pcwrite(ptc, outname); end

↑ I would like to know more about what each line of this code does.