Point Cloud from DEPTH and RGB image

Hi everyone, my aim is to create a point cloud from depth image and RGB image I obtained from kinect. Successfully I calculated coordinates and visualised depth map as a cloud, but dont know how to add the color information from RGB.

Is there a way how to insert RGB matrix as a color parameter of ptCloud?

%ptCloud = pointCloud(points,'Color',cmatrix);

I found

scatter3

plot in discussion, but in my case it doesnt show correct values.

Any advice?

Thank you

clearvars
%load depth
final{1}=['FinalKamera.hdr'];
linear = hdrread(final{1});
linear = linear(:, :, 1);
%load rgb
rgb = imread('FinalKamerargb.jpeg');
% depth coordinates calculation
points=zeros(512*424,3);
inc=1;
for i=1:424
     for j=1:512
     points(inc,1)=(i-256.4626)*linear(i,j)/(365.3277);  % x array
     points(inc,2)=(j-213.1488)*linear(i,j)/(366.8126);  % y array
     points(inc,3)= linear(i,j); % z array
     inc=inc+1;
     end
end

%ptCloud = pointCloud(points,'Color',cmatrix);
%pcshow(points);

clr = reshape(double(rgb)/255, [], 3);
scatter3(points(:,1), points(:,2), points(:,3), 6, clr, '.')
axis tight vis3d
title('Point Cloud'); xlabel('X'); ylabel('Y'); zlabel('Z');

Answer 1

[Andrej Satnik]

This post is old but for anyone who is searching for the answer this is faster method without loops. Correct me if I am wrong.

%depth is depth image in double format
Sd = size(depth);
[X,Y] = meshgrid(1:Sd(2),1:Sd(1));

%K is calibration matrix
X = X - K(1,3) + 0.5;
Y = Y - K(2,3) + 0.5;

XDf = depth/K(1,1);
YDf = depth/K(2,2);
X = X .* XDf;
Y = Y .* YDf;

XY = cat(3,X,Y);
cloud = cat(3,XY,depth);
cloud = reshape(cloud,[],3) / 1000.0;

% if you can use matlab point cloud library
cloud = pointCloud(cloud);
pcshow(cloud);

Happy coding.

Best Andrej

[muhammad siddique]

@Andrej where in code you read rgb image?

Answer 2

[Preetham Manjunatha]

This link can help to convert RGB-D images to point cloud, provided the camera intrinsic parameters.