Solar diagrams

A forum dedicated to the Draft, Arch and BIM workbenches development.
User avatar
microelly2
Posts: 4319
Joined: Tue Nov 12, 2013 4:06 pm
Contact:

Re: Solar diagrams

Postby microelly2 » Wed Nov 23, 2016 1:41 pm

saso wrote:did someone say shadows ?! viewtopic.php?t=9663 ;)
I did not, I stopped work on lights because my experiments did not work like in the open inventor docs.

I use the calucation of the direction of the normal vector to compute the light intensity and cat put it as color map onto the (bspline) surface.
but for a good rendering FreeCAD is not best place.
https://www.youtube.com/watch?v=AAnbG-HnIGM
User avatar
saso
Posts: 1279
Joined: Fri May 16, 2014 1:14 pm
Contact:

Re: Solar diagrams

Postby saso » Wed Nov 23, 2016 3:51 pm

microelly2 wrote:
saso wrote:did someone say shadows ?! viewtopic.php?t=9663 ;)
I did not, I stopped work on lights because my experiments did not work like in the open inventor docs.

I use the calucation of the direction of the normal vector to compute the light intensity and cat put it as color map onto the (bspline) surface.
but for a good rendering FreeCAD is not best place.
https://www.youtube.com/watch?v=AAnbG-HnIGM
Yes, most of your work is totally amazing to me :) Such studies however don't really require high quality renderings, more important is the quality of the results data (the numbers), when we talk about solar studies and such. But my post about the shadows was actually just to show that shadows are possible and as I understood Yoriks example, it is not a rendering but a "live" shadow directly in the 3d view? :|
Last edited by saso on Wed Nov 23, 2016 4:02 pm, edited 1 time in total.
User avatar
microelly2
Posts: 4319
Joined: Tue Nov 12, 2013 4:06 pm
Contact:

Re: Solar diagrams

Postby microelly2 » Wed Nov 23, 2016 3:59 pm

saso wrote:[ But my post about the shadows was actually just to show that shadows are possible and as I understood Yoriks example it is not a rendering but a "live" shadow directly in the 3d view? :|
Yes, it should be according to open inventor theory, but I do not get it on my computer
User avatar
saso
Posts: 1279
Joined: Fri May 16, 2014 1:14 pm
Contact:

Re: Solar diagrams

Postby saso » Wed Nov 23, 2016 4:11 pm

microelly2 wrote:
saso wrote:[ But my post about the shadows was actually just to show that shadows are possible and as I understood Yoriks example it is not a rendering but a "live" shadow directly in the 3d view? :|
Yes, it should be according to open inventor theory, but I do not get it on my computer
By the way, in your work on point clouds did you look in to creating point clouds from pictures? There seem to be some quite usable open libraries but lacking of good and easy to use open tools (gui), http://ccwu.me/vsfm/ is for example best I could found, but with not the most open license... ? :roll:
User avatar
microelly2
Posts: 4319
Joined: Tue Nov 12, 2013 4:06 pm
Contact:

Re: Solar diagrams

Postby microelly2 » Wed Nov 23, 2016 9:51 pm

saso wrote: By the way, in your work on point clouds did you look in to creating point clouds from pictures? There seem to be some quite usable open libraries but lacking of good and easy to use open tools (gui), http://ccwu.me/vsfm/ is for example best I could found, but with not the most open license... ? :roll:
when I have finished my version of the nurbs workbench (create models from point clouds)
I will continue my work on the reconstruction workbench and improve pattern recognition
to get geometrical objects (lines, circles, planes, spheres etc.) from images
for calculating 3D information from multiple images there are already good algorithms
I will remember your link and look how to reuse it, thank you
User avatar
saso
Posts: 1279
Joined: Fri May 16, 2014 1:14 pm
Contact:

Re: Solar diagrams

Postby saso » Wed Nov 23, 2016 11:11 pm

microelly2 wrote:
saso wrote: By the way, in your work on point clouds did you look in to creating point clouds from pictures? There seem to be some quite usable open libraries but lacking of good and easy to use open tools (gui), http://ccwu.me/vsfm/ is for example best I could found, but with not the most open license... ? :roll:
when I have finished my version of the nurbs workbench (create models from point clouds)
I will continue my work on the reconstruction workbench and improve pattern recognition
to get geometrical objects (lines, circles, planes, spheres etc.) from images
for calculating 3D information from multiple images there are already good algorithms
I will remember your link and look how to reuse it, thank you
Point clouds have generally be done with laser scans, lasers however are expensive so people have come up with this technique (https://en.wikipedia.org/wiki/Photogrammetry) to create point clouds (also big and dense ones) from images. VisualSFM as I understand is more just a gui for some of the open libraries for doing this... Here is one video how to set it up and use it with some nice example of scanning terrain with a drone https://www.youtube.com/watch?v=wBKidr0e-XA There are of course also comercial solutions for this like for example Autodesk ReCap https://www.youtube.com/watch?v=5lQVH3hSh0c
User avatar
yorik
Site Admin
Posts: 11398
Joined: Tue Feb 17, 2009 9:16 pm
Location: São Paulo, Brazil
Contact:

Re: Solar diagrams

Postby yorik » Wed Nov 23, 2016 11:26 pm

The example from the other post was just to show that coin can do it. But indeed there are several problems, I could never make it work on the Intel, only on the Nvidia chip.

In any case, it requires encapsulating he whole scene in a SoShadowGroup node. Which cannot currently be done in python. So we need to modify the FreeCAD 3D view to add space for that node. Then with pivy we can tweak it as it needs. But didn't look much at how to do that yet
User avatar
microelly2
Posts: 4319
Joined: Tue Nov 12, 2013 4:06 pm
Contact:

Re: Solar diagrams

Postby microelly2 » Thu Nov 24, 2016 9:13 am

saso wrote: Point clouds have generally be done with laser scans, lasers however are expensive so people have come up with this technique (https://en.wikipedia.org/wiki/Photogrammetry) to create point clouds (also big and dense ones) from images. VisualSFM as I understand is more just a gui for some of the open libraries for doing this... Here is one video how to set it up and use it with some nice example of scanning terrain with a drone https://www.youtube.com/watch?v=wBKidr0e-XA There are of course also comercial solutions for this like for example Autodesk ReCap https://www.youtube.com/watch?v=5lQVH3hSh0c
it's still a dream for me,
I use a kinect sensor to get 3d data. this is good enough for the moment and easy to use.
the software to use it is open source
for working with 2d images there are some methods in opencv too
User avatar
saso
Posts: 1279
Joined: Fri May 16, 2014 1:14 pm
Contact:

Re: Solar diagrams

Postby saso » Thu Nov 24, 2016 2:34 pm

microelly2 wrote:
saso wrote: Point clouds have generally be done with laser scans, lasers however are expensive so people have come up with this technique (https://en.wikipedia.org/wiki/Photogrammetry) to create point clouds (also big and dense ones) from images. VisualSFM as I understand is more just a gui for some of the open libraries for doing this... Here is one video how to set it up and use it with some nice example of scanning terrain with a drone https://www.youtube.com/watch?v=wBKidr0e-XA There are of course also comercial solutions for this like for example Autodesk ReCap https://www.youtube.com/watch?v=5lQVH3hSh0c
it's still a dream for me,
I use a kinect sensor to get 3d data. this is good enough for the moment and easy to use.
the software to use it is open source
for working with 2d images there are some methods in opencv too
Maybe Python Photogrammetry Toolbox http://184.106.205.13/arcteam/ppt.php is a good way for you to make some quick tests https://www.youtube.com/watch?v=xsEdu6Xq6KU I believe it uses the same libraries as VisualSFM
User avatar
saso
Posts: 1279
Joined: Fri May 16, 2014 1:14 pm
Contact:

Re: Solar diagrams

Postby saso » Thu Nov 24, 2016 2:41 pm

yorik wrote:The example from the other post was just to show that coin can do it. But indeed there are several problems, I could never make it work on the Intel, only on the Nvidia chip.
Do you maybe know if it is just the older Intel chips, they are quite well known to have bad support for graphic stuff? If that is the only issue I would say it should not represent a reason not to implement it, shadows would be nice to have, even if just "false" shadows (unreal sun direction). Any info for AMD chips? I will try to find some time to test this a bit more, if I will know how to :roll: