IntegralImageNormalEstimation with NaN points problem

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

IntegralImageNormalEstimation with NaN points problem

maria
Hi I am using IntegralImageNormalEstimation for Organized point cloud data and when the number of points  with NaN value in the data exceeds the 12% of the total number of points, i get for all points normals with NaN value.

I have tried all different options
pcl::IntegralImageNormalEstimation<pcl::PointXYZ, pcl::Normal> nei;
nei.setNormalEstimationMethod (nei.COVARIANCE_MATRIX);
or
nei.setNormalEstimationMethod (nei.AVERAGE_3D_GRADIENT);
or
nei.setNormalEstimationMethod (nei.AVERAGE_DEPTH_CHANGE);

with nei.setNormalSmoothingSize(10.0f);

but results are the same, that is normals for all points have NaN values.

What are the limitations of the algorithm? I am targeting real-time application and this method works quite fast if the number of points with NaN values are less than 12% of the total number of points.

Thank you in advance for any help on this.
Maria

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: IntegralImageNormalEstimation with NaN points problem

Stefan Holzer
Hi,

how are these NaN distributed over the point cloud? Are they clustered
in areas or equally-ish distributed? Could you provide an example for
such a cloud?

Cheers,
Stefan


On 10/11/2013 11:10 AM, maria wrote:

> Hi I am using IntegralImageNormalEstimation for Organized point cloud data
> and when the number of points  with NaN value in the data exceeds the 12% of
> the total number of points, i get for all points normals with NaN value.
>
> I have tried all different options
> pcl::IntegralImageNormalEstimation<pcl::PointXYZ, pcl::Normal> nei;
> nei.setNormalEstimationMethod (nei.COVARIANCE_MATRIX);
> or
> nei.setNormalEstimationMethod (nei.AVERAGE_3D_GRADIENT);
> or
> nei.setNormalEstimationMethod (nei.AVERAGE_DEPTH_CHANGE);
>
> with nei.setNormalSmoothingSize(10.0f);
>
> but results are the same, that is normals for all points have NaN values.
>
> What are the limitations of the algorithm? I am targeting real-time
> application and this method works quite fast if the number of points with
> NaN values are less than 12% of the total number of points.
>
> Thank you in advance for any help on this.
> Maria
>
>
>
>
>
> --
> View this message in context: http://www.pcl-users.org/IntegralImageNormalEstimation-with-NaN-points-problem-tp4030119.html
> Sent from the Point Cloud Library (PCL) Users mailing list mailing list archive at Nabble.com.
> _______________________________________________
> [hidden email] / http://pointclouds.org
> http://pointclouds.org/mailman/listinfo/pcl-users


--
Stefan Holzer, M.Sc., Dipl.-Inf(FH).
Computer Aided Medical Procedures
Technische Universität München
Phone: +49 (0)89 / 289 19400
Boltzmannstr. 3, 85748 Garching

http://campar.in.tum.de/Main/StefanHolzer

_______________________________________________
[hidden email] / http://pointclouds.org
http://pointclouds.org/mailman/listinfo/pcl-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: IntegralImageNormalEstimation with NaN points problem

peter9606
May I ask which method is the fastest ?
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: IntegralImageNormalEstimation with NaN points problem

maria

Hi i haven't use PCL for a while (these messages are from 2013). At that time IntegralImageNormalEstimation was faster compared to other methods for organized point clouds.


On 11/7/2017 5:52 πμ, peter9606 [via Point Cloud Library (PCL) Users mailing list] wrote:
May I ask which method is the fastest ?


If you reply to this email, your message will be added to the discussion below:
http://www.pcl-users.org/IntegralImageNormalEstimation-with-NaN-points-problem-tp4030119p4044748.html
To unsubscribe from IntegralImageNormalEstimation with NaN points problem, click here.
NAML

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: IntegralImageNormalEstimation with NaN points problem

peter9606
Thanks for your quick replying.
Loading...