[Rock-dev] 3d point uncertainty

Javier Hidalgo Carrió javier.hidalgo_carrio at dfki.de
Fri Nov 8 14:59:59 CET 2013


On 11/08/2013 12:33 PM, Sylvain Joyeux wrote:
> On 11/08/2013 10:35 AM, Alexander Duda wrote:
>> To not block the hole process we could also think of creating a new
>> package for vision related types and deprecate the one we have in base
>> (the tooling is already supporting this particular case).
> Doing it this way also has drawbacks, namely that we'll have 
> fragmentation between the components that use the new type(s) and the 
> components that use the old types.
>
> But, in general, we did start talking about using as much of the types 
> directly from the "established libraries": opencv for images, pcl for 
> point clouds, ...
>
Still pcl::Pointcloud does not have uncertainty information.

I agree with it in case of well established libraries. However, I have 
my doubts on how to organize it. Instead of creating optional 
dependencies. One possible suggestion would be to organize the types per 
package set.

  * base would have the basic types: Time, Eigen, Temperature, Joints,
    Angle, RigidBodyState, etc.. as well as current types to support
    backward compatibility.
  * perception (could be a rename for image_processing)  would have
    opencv::Mat, pcl::Pointcloud, cv::detail::ImageFeatures, etc...
  * slam would have dedicated slam types
  * planning would have common planning classes perhaps using some SBPL,
    MoveIt types
  * control would have the correspondent control types.

Question: Is there any inconvenient to organize the types in such a way?

Javier.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.dfki.de/pipermail/rock-dev/attachments/20131108/479c1619/attachment.htm 


More information about the Rock-dev mailing list