I used plot3. when I try to rotate the 3D image, it is too slow to rotate for each time.

11 Ansichten (letzte 30 Tage)
hFig = figure('name','3D Plotting_Solve_Pyramid','numbertitle','off'); hold on;
set(gcf,'Renderer','OpenGL'); % might improve performance
hAxes = gca;
for row = 1:hightl
for col = 1:4:widthl
objcolor = double(I2(row,col,:))./255.0;
if objectpoint(row,col,3) >= -2.0 %&& (sum(objcolor)>0.3)
plot3(hAxes, objectpoint(row,col,1), objectpoint(row,col,2), objectpoint(row,col,3), '.', 'MarkerEdgeColor',objcolor);
end
end
end
%Set up the view.
xlabel(' X axis');
ylabel(' Y axis');
zlabel(' Z axis');
grid on;
cameratoolbar show;
axis vis3d;
axis equal;

Antworten (1)

Mike Garrity
Mike Garrity am 20 Aug. 2015
Bearbeitet: Mike Garrity am 20 Aug. 2015
I'm guessing that widthl and hightl are fairly large. This means that you're creating a large number of objects which are each drawing a single point. That's not a very efficient way to do it. You'll be much better off batching them up into a single object. A modern GPU is optimized for drawing large batches of geometry, not for drawing large numbers of tiny batches.
I discussed that in quite a bit of detail in this blog post and this blog post.
It looks like it'd actually be quite simple in this case. It's probably something like:
x = objectpoint(:,:,1);
y = objectpoint(:,:,2);
z = objectpoint(:,:,3);
r = double(x) / 255;
g = double(y) / 255;
b = double(z) / 255;
mask = (z(:)>=-2) & (r(:)+g(:)+b(:) > .3);
scatter3(hAxes,x(mask),y(mask),z(mask),[],[r(mask), g(mask), b(mask)],'.')
On my system, with a Quadro K600 on R2015a, that runs at over 10 million markers per second. You could probably do better with a bit of tweaking, but that's about 120 frames per second for this image:
  5 Kommentare
Houghton
Houghton am 22 Aug. 2015
From scatter3(hAxes,x(mask),y(mask),z(mask),[], ... [r(mask), g(mask), b(mask)],'.')
What does [] mean in this command?
Walter Roberson
Walter Roberson am 22 Aug. 2015
"If S is empty, then the default size of 36 points squared is used."

Melden Sie sich an, um zu kommentieren.

Kategorien

Mehr zu Get Started with Image Processing Toolbox finden Sie in Help Center und File Exchange

Community Treasure Hunt

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

Start Hunting!

Translated by