[Sidefx-houdini-list] Imitating Perceptual field of view using Point clouds..

Olex P hoknamahn at gmail.com
Fri Jan 28 05:00:16 EST 2011


Hey,

> Also I had setup a while loop with the point
> cloud iterate node to manually filter them out (within the loop) but I
> couldn't really think of any clever ways to get the average of any arbitrary
> imported point cloud attribute..(which I would need to calculate Perceived
> average position of each point in space)

You will need the pcopenlod() function for that
http://localhost:48626/vex/functions/pcopenlod


On Fri, Jan 28, 2011 at 4:15 AM, bhavesh pandey <pandeybhavesh at gmail.com> wrote:
> Hi Andrew! I took up flocking as a subject because I wanted to explore
> implementations of technical effects in Houdini..
> Which is why I want to avoid the Follow and Interact pop..
> I would definitely have a look at Tiny.otl
> Thanks.
>
> On Fri, Jan 28, 2011 at 1:58 AM, bhavesh pandey <pandeybhavesh at gmail.com>wrote:
>
>> Hi people..I'm trying to implement a flocking system in houdini using point
>> clouds in SOP Solver..Although I can compare the angles of the points and
>> thus simulate something similar to a field of view, I want to optimize my
>> system by using Point Clouds instead of doing a one to one comparison with
>> all the points in the scene. Also I had setup a while loop with the point
>> cloud iterate node to manually filter them out (within the loop) but I
>> couldn't really think of any clever ways to get the average of any arbitrary
>> imported point cloud attribute..(which I would need to calculate Perceived
>> average position of each point in space)
>>
>> To test the cone angle option in the PCOpen node I did a small test. I took
>> 2 points and set their normals to their respective normalized positions so
>> the points had Normal attribute on them. After this I used a transform node
>> (both nodes had recompute normals turned off) and moved around any one
>> arbitrary point (transform node was appended before the vopsop used for
>> calculations) but I dont feel Point cloud data is getting filtered out based
>> off the Cone Angle..
>> for instance:
>> according to the help card:
>>
>> "Points with dot(N, Npoint) > cos(cone) will be queued, and unmatched
>> points will be rejected."
>>
>> now since i had set position data as normals, the dot product of the
>> positions for this point (1.95327, 0.334443, 0.340146) and the nearest point
>> (-3.49418, 0, 1.43283) as indicated by the pcfilter (i'm just looking up for
>> 1 point, so i guess the value would not be averaged out??) was  -6.337703
>> whereas Cos(cone = 90) = 0
>>
>> I am not really able make sense out of it..I'm sure I'm missing something,
>> Could some one point me to the right direction??
>>
>> The Point cloud open node should have some optimized solution for this
>> (filtering based off cone angle) which would help me optimize my VOP
>> network.
>>
>> Thanks.
>>
>> --
>> Kind Regards,
>> Bhavesh Pandey.
>>
>> www.bhaveshpandey.blogspot.com
>>
>>
>
>
> --
> Kind Regards,
> Bhavesh Pandey.
>
> www.bhaveshpandey.blogspot.com
> _______________________________________________
> Sidefx-houdini-list mailing list
> Sidefx-houdini-list at sidefx.com
> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>



More information about the Sidefx-houdini-list mailing list