Leslie,

On 06/14/2012 04:37 PM, Lesliecasas wrote:

> Hi everybody, i am trying to registrate two cloud points, the first one is

> from a laser scanner of aprox 160kdpi and the other comes from a 3d

> reconstruction using stereo vision. For example, the ones i am using : the

> first has 65000 points and the second 4000.

>

> I've read the paper for FPFH and it takes a neighborhood of points of the

> same size, does it work in this case, when the neighborhood on both cases

> doesn't mean the same dimensions?

What do you mean by "same dimensions"? Is the data still 3D? Is it still in the same coordinate system / metric units?

What is different? Just the point density?

> If not i should interpolate one in order to match the big one? or i may use

> instead of points, surfaces?

> Pls can you help me. how to preprocess my cloud points?

If we're only talking about point density, I would recommend some density "regularizers", such as MLS or VoxelGrids, etc.

Cheers,

Radu.

_______________________________________________

[hidden email] /

http://pointclouds.orghttp://pointclouds.org/mailman/listinfo/pcl-users