[Rtk-users] need help with geometry

[Simon Rit]

Hi,
The z spacing of projection stack is not used, it can be set to anything.
What you do to set the spacing of the reconstructed volume is correct
except you probably want to set the z spacing to a similar value.
In the file you sent, the origin is not set. You must origin for the
coordinate of the first pixel of your projection, such that point 0,0 in
your projection corresponds to the coordinate of the projection of the
center-of-rotation. For example, I have set it with your dataset such that
the projection is centered and gives promising results. Please find
attached the corrected mhd and I executed the following commands to obtain
a decent result (I think, I'm not used to (real) bugs!):

rtksimulatedgeometry --sdd 214.119 --sid 91.770 -o geo.xml -n 148 --arc
198.45
rtkfdk -p . -r proj_sub_simon.mhd -o fdk.mha --dimension 256,256,256 -g
geo.xml --spacing 0.00901029,0.00902951,0.00901029

then I tried with larger spacing to cover the whole dataset (on my laptop I
don't want to increase the number of pixels)

rtkfdk -p . -r proj_sub_simon.mhd -o fdk.mha --dimension 256,256,256 -g
geo.xml --spacing 0.02

Simon

On Sat, Mar 21, 2015 at 4:41 PM, 1 1 <infrombox at yandex.ru> wrote:

> Hello RTK users. I am newbie in RTK but its interesting to me because of
> open source code! I wonder if i could get it worked
> But for this moment i can't reconstruct my test projections.
> I use SkyScan tomograph (cone beam) and after acquisition process i have
> set of tifs and .log file which contains meta info.
>
> In my test i rotate bug on a table - gantry rotation, one step: 0.45
> degrees, total rotation: 199.35 degrees (444 steps) and all other info
> presents in meta file.
>
> In the following sequence i use every 3rd projection
> rtksimulatedgeometry -o geometry.xml --n 148 --arc=198.45 --sid=214.119
> --sdd=91.770
> rtksubselect --geometry=geometry.xml --out_geometry=sub_geometry.xml
> --out_proj=sub_proj.mhd -p tifs -r .tif$
> rtkfdk -p . -r sub_proj.mhd -o fdk.mha -g geometry_sub.xml --verbose
> --dimension 1000,500,1000 --spacing 0.009,1,0.009
>
> but fdk.mha contains blank black cube. i play with geometry params all day
> long, but have nothing results =) especially with spacing... as i
> understand in auto-generated mhd file the spacing values based on dpi info
> in tif, but what is 1 for z ? should i change it ?
> so now values:
> in auto-generated proj_sub.mhd: 0.00901029 0.00902951 1
> on rtkfdk call which i set: --spacing 0.00901029,1,0.00902951
> i want to set the same spacing, because i want the same spacing in volume,
> but i am not sure that i do it in a right way.
> i tried different variations of spacing, but nothing changes in result...
> could you explain me with spacing there a bit
>
> but problem may be not in/not only in spacing,
>
> these tomography scans reconstructs in another software, so i have no
> doubt that RTK can do it, i just have set right geometry.
>
> pair mhd/raw generated by rtksubselect:
> https://drive.google.com/file/d/0B9EY8AokXE8zSHZqaUJ5UmtHa1E/view?usp=sharing
> (163 MB)
> meta file:
> https://drive.google.com/file/d/0B9EY8AokXE8zaWw0aFdRWUQ0d2c/view?usp=sharing
> i can share original tifs if needed.
>
> offtop: seems files 00.mhd, 00.raw at
> http://wiki.openrtk.org/index.php/RTK/Scripts/ForwardProjection are not
> accessible.
>
> thanks
> _______________________________________________
> Rtk-users mailing list
> Rtk-users at public.kitware.com
> http://public.kitware.com/mailman/listinfo/rtk-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/rtk-users/attachments/20150323/ffdd0884/attachment-0009.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: proj_sub_simon.mhd
Type: application/octet-stream
Size: 328 bytes
Desc: not available
URL: <http://public.kitware.com/pipermail/rtk-users/attachments/20150323/ffdd0884/attachment-0009.obj>