Re: [ros-users] surfels and rviz

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: [ros-users] surfels and rviz

Radu B. Rusu
Administrator
Stephane,

We recently addressed the surfels discussions internally, and couldn't come up to a conclusion on whether they make
sense for processing or not. While they are certainly useful for rendering (they were developed for that purpose in
Computer Graphics), we're not sure if they would complement or heavily intersect with the other primitives that we've
been using. We already have points with normals and support size and other features, and we've been using them for
years, except they are called differently.

In terms of adding another point type to PCL, sure, that is possible, but please use pcl-users@, and let's discuss it
asap, so we get it in 1.0. We're more than happy to extend the family of supported types, especially if people want to
use them.

As far as PointCloud2 support goes, diamondback is locked and we cannot make any changes. For RViz again, I'm not sure
exactly what they would bring new... I guess the visualization would be a bit prettier and you would have non-uniform
point sizes? :)

Cheers,
Radu.
--
http://pointclouds.org

On 02/02/2011 10:41 AM, Stéphane Magnenat wrote:

> Hello,
>
> In the continuation of our work with the kinect, we are interested at
> looking into combining multiple scans using surfels. Surfels [1] are
> oriented surface elements (point+normal+radius), and typically hold
> color information and view statistics, to provide intelligence to the
> update process. Recent works [2,3] tend to indicate that surfels are
> well-suited for RGB-D sensors such as the kinect.
>
> We are interested in adding surfel support to rviz/ros. To that end, it
> would be nice to agree on a naming convention for surfels, maybe to add
> a corresponding type to pcl/point_types.hpp, and to add support in rviz.
>
> Are there other people around interested in surfels? Are there some best
> practice on how to extend the PointCloud(2) support in rviz?
>
> Thank you, kind regards,
>
> Stéphane
>
>
> [1] Pfister et al., Surfels: surface elements as rendering primitives
> http://www.cg.inf.ethz.ch/Downloads/Publications/Papers/2000/p_Pfi00.pdf
>
> [2] T Weise et al., In-hand Scanning with Online Loop Closure
> http://www.mmp.rwth-aachen.de/publications/pdf/weise-onlineloopclosure-3dim09.pdf
>
> [3] Henry et al. - RGB-D Mapping: Using Depth Cameras for Dense 3D
> Modeling of Indoor Environments
> http://www.seattle.intel-research.net/RGBD/RGBD-RSS2010/papers/henry-RGBD10-RGBD-mapping.pdf
>
_______________________________________________
[hidden email] / http://pointclouds.org
https://code.ros.org/mailman/listinfo/pcl-users
Reply | Threaded
Open this post in threaded view
|

Re: [ros-users] surfels and rviz

Stéphane Magnenat
Hi,

> We recently addressed the surfels discussions internally, and couldn't
> come up to a conclusion on whether they make sense for processing or
> not. While they are certainly useful for rendering (they were developed
> for that purpose in Computer Graphics), we're not sure if they would
> complement or heavily intersect with the other primitives that we've
> been using. We already have points with normals and support size and
> other features, and we've been using them for years, except they are
> called differently.

What do you mean by "support size"? Do you refer to per-point radius?
(such as PCL::PointWithRange?)

> As far as PointCloud2 support goes, diamondback is locked and we cannot
> make any changes. For RViz again, I'm not sure exactly what they would
> bring new... I guess the visualization would be a bit prettier and you
> would have non-uniform point sizes? :)

If I understood correctly, PointCloud2 is fairly agnostic to the content
of the points, right? So the question lies more at the levels of PCL
mapping and RViz display, isn't it?

I will discuss the other points on PCL-user.

Have a nice day,

Stéphane

--
Dr Stéphane Magnenat
http://stephane.magnenat.net
_______________________________________________
[hidden email] / http://pointclouds.org
https://code.ros.org/mailman/listinfo/pcl-users
Reply | Threaded
Open this post in threaded view
|

Re: [ros-users] surfels and rviz

Radu B. Rusu
Administrator
On 02/03/2011 12:46 AM, Stéphane Magnenat wrote:

> Hi,
>
>> We recently addressed the surfels discussions internally, and couldn't
>> come up to a conclusion on whether they make sense for processing or
>> not. While they are certainly useful for rendering (they were developed
>> for that purpose in Computer Graphics), we're not sure if they would
>> complement or heavily intersect with the other primitives that we've
>> been using. We already have points with normals and support size and
>> other features, and we've been using them for years, except they are
>> called differently.
>
> What do you mean by "support size"? Do you refer to per-point radius? (such as PCL::PointWithRange?)

Yes. Either radius or the number of k-neighbors (first k).

>> As far as PointCloud2 support goes, diamondback is locked and we cannot
>> make any changes. For RViz again, I'm not sure exactly what they would
>> bring new... I guess the visualization would be a bit prettier and you
>> would have non-uniform point sizes? :)
>
> If I understood correctly, PointCloud2 is fairly agnostic to the content of the points, right? So the question lies more
> at the levels of PCL mapping and RViz display, isn't it?

Yup, you're right. You can shove anything into a PointCloud2 :) I think the entire ROS ecosystem could fit there :). I
guess what I wanted to say is that RViz needs to be modified to understand this new type, or simply just use the XYZ
information. RVE could be going along those ways.

In fact, this gave me an idea... if we add proper support for "point + support size" in PCL, we could update our PCL
visualizer library to use it, just to see how it works.


Cheers,
Radu.
--
http://pointclouds.org

_______________________________________________
[hidden email] / http://pointclouds.org
https://code.ros.org/mailman/listinfo/pcl-users