You are not logged in.

> The forum rules have been updated. Please take a moment to read them.


#1 2012-09-18 23:31:06

leifs
Member
From: Ørsta Norway
Registered: 2009-09-06
Posts: 464
Website

This should not be possible

I'm struggling with a coastal sphere, and in the process I experience a phenomena that should not be there.
See the image below: how is it possible for APG 3.0b1 to find the CP #10 in both pictures ????
The first is in the sea, dark blue, and the other is in the hills, brown.
The first is to the left of a rank of green links, the second is on the opposite side of the rank.

I know there are several ways to remove very erroneous CP's , but why are they there in the first place ???

leifs


Uploaded Images

Last edited by leifs (2012-09-18 23:47:17)


Olympus OM-D E-M5, Panasonic 8mm f3.5 fisheye, Olympus 12mm f2.0, Leica 25mm f1.4, Zeiss 50mm f1.4, Canon FD 85mm f1.8, Canon 100-400mm f/4.5-5.6 L
Seitz VRdrive2
Intel i7 980X, 48GB RAM, Win7 64bit, SSD RAIDs

Offline

 

#2 2012-09-18 23:41:44

klausesser
Member
From: Düsseldorf, Germany
Registered: 2006-05-22
Posts: 6436
Website

Re: This should not be possible

leifs wrote:

I'm struggling with a coastal sphere, and in the process I experience a phenomena that should not be there.
See the image below: how is it possible for APG 3.0b1 to find the CP #18 in both pictures ????
The first is in the sea, dark blue, and the other is in the hills, brown.
The first is to the left of a rank of green links, the second is on the opposite side of the rank.

I know there are several ways to remove very erroneous CP's , but why are they there in the first place ???

leifs

#18 is on the same place in both pictures. But #10 isn´t - it´s red, and so that has no meaning. Simply delete it.

best, Klaus


If you want something you´ve never had,
then you´ve got to do something you´ve never done.

Offline

 

#3 2012-09-19 08:48:03

ThomasV
Member
Registered: 2012-08-27
Posts: 246

Re: This should not be possible

Hello Leif,

The control point detection algorithm is not based on color. Moreover, as Klaus explains, the fact that this point is marked red informs you that this point has a large error regarding the optimization chosen by Autopano. Therefore, it just influenced a little the overall result.
You can then safetly remove this point to improve the rms on this link.

Regards,
Thomas

Offline

 

#4 2012-09-19 09:37:31

leifs
Member
From: Ørsta Norway
Registered: 2009-09-06
Posts: 464
Website

Re: This should not be possible

ThomasV wrote:

the fact that this point is marked red informs you that this point has a large error regarding the optimization chosen by Autopano.
You can then safetly remove this point to improve the rms on this link.
Regards,
Thomas

I know, I know.
Visual inspection tells me "it should not be there".
I have to inspect every link to remove them and in large panos there are a lot of links !

leifs


Olympus OM-D E-M5, Panasonic 8mm f3.5 fisheye, Olympus 12mm f2.0, Leica 25mm f1.4, Zeiss 50mm f1.4, Canon FD 85mm f1.8, Canon 100-400mm f/4.5-5.6 L
Seitz VRdrive2
Intel i7 980X, 48GB RAM, Win7 64bit, SSD RAIDs

Offline

 

#5 2012-09-20 03:31:12

tived
Member
From: Dane in Western Australia
Registered: 2008-07-11
Posts: 835

Re: This should not be possible

Hi Leif,

I look for all the points with large error and delete them, then optimise

Henrik

Offline

 

Board footer

Powered by PunBB
© Copyright 2002–2005 Rickard Andersson