Reconstruction start crashes Colmap

798 views
Skip to first unread message

johan.wo...@gmail.com

unread,
May 10, 2017, 5:13:29 PM5/10/17
to COLMAP
Colmap 2.1 (commit 308d5c1 without Cudo)
OSX 10.12.4

Loaded 1046 images
Feature extraction: done and OK
Feature matching: sequential (default settings)

When reconstruction is started Colmap crashes immediatly. No time to check the log.
Re-opened the program. 
Reconstruction next image 1x: works
Reconstruction next image an other time: works. Each time 1 images is loaded.
Reconstruction next image: image pair shown on screen
Reconstruction next image: crash.


Followed the advice on the documentation page:
"Manually choose an initial image pair, if COLMAP fails to initialize. ChooseReconstruction > Reconstruction options > Initand set images from the database management tool that have enough matches from different viewpoints."

Image pair selected with >1000 matches.
Identical behaviour as above.

Followed the advice on the documentation page:
"enable guided matching &  or increase the overlap in sequential matching"

Identical behaviour.

Tried: "automatic reconstruction".
Crash at the moment the matching is done. I assume this is agains the same behaviour.

Any idea?

johan.wo...@gmail.com

unread,
May 10, 2017, 5:19:38 PM5/10/17
to COLMAP, johan.wo...@gmail.com
Extra note: same crash when running the Windows 64bit (no Cuda) version.

Op woensdag 10 mei 2017 23:13:29 UTC+2 schreef johan.wo...@gmail.com:

Johannes Schönberger

unread,
May 10, 2017, 7:19:22 PM5/10/17
to col...@googlegroups.com
Can you share the database with me?
--
You received this message because you are subscribed to the Google Groups "COLMAP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to colmap+un...@googlegroups.com.
To post to this group, send email to col...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/colmap/f795ec03-65fb-4c9f-9184-6cf86ad84ae4%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Johannes Schönberger

unread,
May 10, 2017, 7:23:06 PM5/10/17
to col...@googlegroups.com
*No need for the images. The database alone is fine. You can share it privately with me per email. 

Johan Wouters

unread,
May 11, 2017, 1:31:17 AM5/11/17
to col...@googlegroups.com
Johannes,

Just tried to launch it once more this morning: no more crash. Even though it crashed around 10x in a row.

Things that changed: 
1) I downloaded the "south-building" dataset and did the full flow (incl reconstruction) on that one. That worked fine. And then I went for bed.
2) I had a second instance of Colmap running at the same time of the crashins. That instance was running an exhaustive feature matching on a different project. That was completed sometime this morning.

When I open the "previously crashing" project now, the reconstruction starts normally.  : (
I will, later today, check whether Time Machine can still get me my original DB from the time it was crashing. I will check whether I can get it back into crashing mode.

I copied the full "Apple Crash Report". Can that be of any use?


Sidenote: The other instance (as mentioned above) completed the exhaustive feature matching on the same photo set (to compare the results). I started the reconstruction. It only seems to have found roughly 30 of the >1000 camera positions. As can be expected the results not as it needs to be.
Note on the note: I did run the same photo set through Agisoft before: the sparce cloud is very fine there. So the dataset should be good enough to do a reconstruction (I hope).
Since the documentation set does not explain each of the settings in detail, I did not fiddle around with the setting yet to see how to improve this.
Also for this one, if it is useful for you to get extra info, just let me know.


While typing this mail the "now-not-crashing-anymore" project did do some of the reconstruction already. You can see (attached image) is has already found many more camera positions.
Top one = crashing before - sequential feature matching (still running)
Bottom one = exhaustive matching (stopped reconstruction). Start reconstruction is disabled now, since the SW seems to find it OK. : )


Regards
Johan

Inline image 1

To unsubscribe from this group and stop receiving emails from it, send an email to colmap+unsubscribe@googlegroups.com.

To post to this group, send email to col...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/colmap/f795ec03-65fb-4c9f-9184-6cf86ad84ae4%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "COLMAP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to colmap+unsubscribe@googlegroups.com.

To post to this group, send email to col...@googlegroups.com.

--
You received this message because you are subscribed to a topic in the Google Groups "COLMAP" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/colmap/tFK_fgFB07g/unsubscribe.
To unsubscribe from this group and all its topics, send an email to colmap+unsubscribe@googlegroups.com.

To post to this group, send email to col...@googlegroups.com.
Reply all
Reply to author
Forward
0 new messages