VFH signatures for a set of points

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

VFH signatures for a set of points

Icky
Hey,

I was trying to understand the vfh signature described here http://pointclouds.org/documentation/tutorials/vfh_estimation.php#vfh.

I took a sample cloud and used the pcl vfh algorithm described in the article mentioned which gave me this vector as a result:

99.981308 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 99.981308 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 99.981308 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 97.009056 0 0 0.015782828 0.0039457069 0 0.02367424 0.047348488 0.059185613 0.055239905 0.059185613 0.027619947 0.055239905 0.04340278 0.019728534 0.035511363 0.02367424 0.047348488 0.039457072 0.04340278 0.015782828 0.067077026 0.019728534 0.047348488 0.02367424 0.019728534 0.039457072 0.015782828 0.04340278 0.027619947 0.039457072 0.039457072 0.015782828 0.02367424 0.02367424 0.015782828 0.031565655 0 0.027619947 0.011837121 0.011837121 0.015782828 0.031565655 0.031565655 0.04340278 0.027619947 0.0039457069 0.011837121 0.0039457069 0.015782828 0.027619947 0 0 0.015782828 0.0078914138 0.02367424 0.0078914138 0.0039457069 0.011837121 0.015782828 0.0039457069 0.031565655 0.015782828 0.02367424 0.0078914138 0.019728534 0.0039457069 0.011837121 0 0 0.0039457069 0.019728534 0.011837121 0 0.011837121 0.0039457069 0.0078914138 0.019728534 0.0078914138 0.02367424 0 0 0.0078914138 0.0078914138 0.015782828 0.015782828 0.015782828 0.015782828 0.011837121 0.011837121 0.0078914138 0.019728534 0.0078914138 0.019728534 0.0078914138 0.015782828 0.0078914138 0.0039457069 0.015782828 0.011837121 0.019728534 0.027619947 0.027619947 0.011837121 0.0039457069 0.0039457069 0.0078914138 0.0078914138 0.0078914138 0.0078914138 0.02367424 0 0.0039457069 0 0.0039457069 0.035511363 0 0.019728534 0.027619947 0.019728534 0.04340278 0.059185613 0.074968442 0.13415407 0.1025884 0.11837123 0.13020836 0.13809977

Can u help me understand how that is possible? A histogram is usually normalized to 1 so values like 99.98 as seen in the vector should not be possible!? I got from the article that there are several histograms represented but could not find the resemblance in the vector I have. Can anyone help me clarify this?

Thanks in advance.

Reply | Threaded
Open this post in threaded view
|

Re: VFH signatures for a set of points

aichim
Administrator
Hi,

What you see is correct. Please have a look at Radu's original paper, and you will understand the feature descriptor better. The first few entries are the viewpoint component of the histogram and they are normalized to 100 for each of the \alpha, \beta, \gamma angles. The rest are the usual FPFH descriptor.

Cheers,
Alex
---
http://www.openperception.org



On Feb 6, 2013, at 9:02 AM, Icky <[hidden email]> wrote:

> Hey,
>
> I was trying to understand the vfh signature described here
> http://pointclouds.org/documentation/tutorials/vfh_estimation.php#vfh
> <http://pointclouds.org/documentation/tutorials/vfh_estimation.php#vfh>  .
>
> I took a sample cloud and used the pcl vfh algorithm described in the
> article mentioned which gave me this vector as a result:
>
> 99.981308 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
> 0 0 0 0 0 0 0 0 0 0 0 99.981308 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 99.981308 0 0 0 0 0 0 0 0 0 0 0
> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
> 0 0 97.009056 0 0 0.015782828 0.0039457069 0 0.02367424 0.047348488
> 0.059185613 0.055239905 0.059185613 0.027619947 0.055239905 0.04340278
> 0.019728534 0.035511363 0.02367424 0.047348488 0.039457072 0.04340278
> 0.015782828 0.067077026 0.019728534 0.047348488 0.02367424 0.019728534
> 0.039457072 0.015782828 0.04340278 0.027619947 0.039457072 0.039457072
> 0.015782828 0.02367424 0.02367424 0.015782828 0.031565655 0 0.027619947
> 0.011837121 0.011837121 0.015782828 0.031565655 0.031565655 0.04340278
> 0.027619947 0.0039457069 0.011837121 0.0039457069 0.015782828 0.027619947 0
> 0 0.015782828 0.0078914138 0.02367424 0.0078914138 0.0039457069 0.011837121
> 0.015782828 0.0039457069 0.031565655 0.015782828 0.02367424 0.0078914138
> 0.019728534 0.0039457069 0.011837121 0 0 0.0039457069 0.019728534
> 0.011837121 0 0.011837121 0.0039457069 0.0078914138 0.019728534 0.0078914138
> 0.02367424 0 0 0.0078914138 0.0078914138 0.015782828 0.015782828 0.015782828
> 0.015782828 0.011837121 0.011837121 0.0078914138 0.019728534 0.0078914138
> 0.019728534 0.0078914138 0.015782828 0.0078914138 0.0039457069 0.015782828
> 0.011837121 0.019728534 0.027619947 0.027619947 0.011837121 0.0039457069
> 0.0039457069 0.0078914138 0.0078914138 0.0078914138 0.0078914138 0.02367424
> 0 0.0039457069 0 0.0039457069 0.035511363 0 0.019728534 0.027619947
> 0.019728534 0.04340278 0.059185613 0.074968442 0.13415407 0.1025884
> 0.11837123 0.13020836 0.13809977
>
> Can u help me understand how that is possible? A histogram is usually
> normalized to 1 so values like 99.98 as seen in the vector should not be
> possible!? I got from the article that there are several histograms
> represented but could not find the resemblance in the vector I have. Can
> anyone help me clarify this?
>
> Thanks in advance.
>
>
>
>
>
> --
> View this message in context: http://www.pcl-users.org/VFH-signatures-for-a-set-of-points-tp4025929.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

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