I wonder then if anyone can tell me what I'm doing wrong, because I'm 
getting extremely different results every time.  I have a template for the 
camera orientations which I'm setting via pto_var, but would like to 
fine-tune that via control points.  I'm running the following script on a 
set of 8 images:

#! /bin/sh
FOV=125.2430276873036


pto_gen --projection=0 --fov=$FOV -o project1.pto "$@"
pto_var -o project1.pto --set  y0=-2.0,r0=-91.0,p0=4.5,y1=142.7,r1=-88.5,p1=
3.8,y2=88.7,r2=-87.0,p2=0.4,y3=135.0,r3=-88.0,p3=-3.3,y4=179.0,r4=-89.0,p4=-
1.8,y5=225.0,r5=-92.0,p5=-3.0,y6=266.8,r6=-95.0,p6=-0.0,y7=312.0,r7=-93.7,p7
=1.2 --opt y,p,r project1.pto
cpfind --prealigned -o project1.pto project1.pto
cpclean -o project1.pto project1.pto
linefind -o project1.pto project1.pto
autooptimiser -l -n -o project1.pto project1.pto > /dev/null
pano_modify --projection=2 --fov=AUTO --center --canvas=AUTO --crop=AUTO -o 
project1.pto project1.pto
pto2mk -o project.mk -p pano project1.pto
make -j 12 -f project.mk all


On Wednesday, October 22, 2014 1:48:18 PM UTC-7, Spencer Egart wrote:
>
> I'm stitching images via script, and have noticed that every time I run it 
> I get a slightly different final pano.  I'm making sure to delete any old 
> intermediate .pto files beforehand, so it's not that.
>
> Are any of the tools in Hugin (autooptimiser, cpfind, etc.) based on 
> random/non-deterministic algorithms?  If so is there a way to seed the RNG 
> or in some other way ensure reproducible results? 
>

-- 
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
--- 
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/hugin-ptx/5c2947a4-82a8-4c85-b33e-a85f414b360b%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to