Interpolating points of a 3D point cloud

I have two questions:

1. I have a 3D point cloud which isn't continuous. There are a few gaps in it. I want to interpolate this data and fill in those gaps.

(I have attached the point cloud and code to display it on MATLAB here .)

I came across functions like triscatteredInterp and scatteredInterpolant. I did not understand how to proceed with this in my case. Kindly help me here. I am getting a blank output.

This is what I am doing:

 figure,
x = PointCloud(:,:,1);
x_max = max(x(:))
x_min = min(x(:))
y = PointCloud(:,:,2);
y_max = max(y(:))
y_min = min(y(:))
z = PointCloud(:,:,3);
[xi,yi] = meshgrid(x_max:0.01:x_min, y_max:0.01:y_min);
xyz_2D = cat(2,x(:),y(:),z(:));
pointcloud_2D = xyz_2D(~any(isnan(xyz_2D)|isinf(xyz_2D),2),:);
zi = griddata(pointcloud_2D(:,1),pointcloud_2D(:,2),pointcloud_2D(:,3),xi,yi);

surf(xi,yi,zi); xlabel('x') ylabel('y') zlabel('Depth')

2. My second question is: looking at the point cloud, I want to remove those parts which have lesser number of connected pixels. The way we have bwlabel for 2D Images, is there any function in MATLAB which outputs labels for 8 connected objects? This way, I can disregard the one with lesser number of points connected.

[Gaurav]

Hi @Meghana did you found a good approach for your first question? If yes Can you please tell how you solved it.

Thanks

Answer 1

[Fabio Freschi]

Does triscatteredinterp make the job?

[Meghana Dinesh]

Nope, I wasn't able to execute it using this function as well.