How to improve rendering of the figure after big data display (plot3)

Hi,

I have rendering problem after plotting multiple plot3 within for loop. I would like to post-process gcode data and plot it using plot3. However, as it has many points and layers(slice), I need to do it in for loop unless somebody knows better way. In addition, I do not want to plot one slice in once as it has movement included as well and I do not want to visualise it. Therefore I need to cut points from layers to smaller arrays. It gives even more plot within for loop. Afterwords, it takes ages to rotate...even with my Nvidia Quadro M4000. Any idea how it is possible to improve rendering of the plot? Can I for example create a big matrix with smaller matrices inside, different dimensions and then plot all in once (cell function?)? Once again I do not want to have connection between matrices. Anyway, will it helps plotting it once all or it does not change anything? Thanks for any advice.

[Jan]

Without seeing the code it is hard to guess, what you are doing exactly. It is useful to mention absolute numbers instead of the vague term "many". Maybe you mean 100, or a million. I do not know why it matters, that you use a for loop and how plot3 creates "layers(slices)". You can draw points without a connection for an array input also, but you did not explain how the output should look like. It is not clear what a "connection between matrices" is.

Please edit the question and post more details: Code, screen shots, absolute numbers and timings.

Answer 1

[Kamil]

Hi again,

Thanks for response.

It is difficult to say number because it depends on the model. This one has 188 layers. Below will find the code. I know that it is still messy but the most problem I have with plot3 and visualisation.

I put two outputs. First is with movements that I want to avoid and second without them. The second is my goal. This is only one layer. I have like 188 as I mentioned above.

Thanks for any ideas.

name='C:\222cce167bc24a5d.gcode';

% read lines of the gcode fid = fopen(name); tline = fgetl(fid); A(1)=cellstr(tline); fclose(fid);

index1 = find(cellfun('length',regexp(A,';LAYER')) == 1); % read and save layers

for iL = 2:length(index1)

% read lines between task such as print inner wall, outer wall etc.

    index = find(cellfun('length',regexp(A(index1(iL-1):index1(iL)),';')) == 1); 

    B = A(index1(iL-1):index1(iL));  % Read line in once

    xk = zeros(25,5);
    yk = zeros(25,5);
    k=1;
    n=1;

    for ik = 3:length(index)
        for in=index(ik-1)-1:index(ik)-1
            GG=char(B(in));            % Read lines of code
            indexE = strfind(GG,'E');  % find extrusion
            indexX = strfind(GG,'X');  % fnd X values points
            indexY = strfind(GG,'Y');  % find Y values points
            indexG0 = strfind(GG, 'G0'); % find G0 values points

            Eempty = isempty(indexE);
            Xempty = isempty(indexX);
            GOempty = isempty(indexG0);

            if Xempty == 0
                if GOempty == 0 % I need only G1 
                    n=n+1;
                    k=1;
                    xk(k,n)=str2num(GG(indexX+1:indexY-1)); % get x values
                    yk(k,n)=str2num(GG(indexY+1:indexY+6)); % get y values

                else
                    xk(k,n)=str2num(GG(indexX+1:indexY-1)); % get x values
                    yk(k,n)=str2num(GG(indexY+1:indexY+6)); % get y values

                end
                k=k+1;
            end
        end

        if n == 1
        else
            for i = 1:n
                ck = xk(:,i);
                ck(ck==0) = []; % exclude zero values
                cy = yk(:,i);
                cy(cy==0) = []; % exclude zero values
                z(1:length(ck))=Z(iL-1); % z values
                em=isempty(ck);
                if em == 1
                else
                    plot3(ck,cy,z) % plot
                    hold on
                end
                z=0;
            end
            n=1;
            k=1;
            clear xk yk
        end
    end
end

Answer 2

[Kamil]

It seems that putting a 'nan' value between points does the job. I will keep testing it. It speed up the visualisation. I will try to change the code to do it all layers in once and plot it as one big matrix. I hope it will work.