Segmentation Fault on When Using Ceres

[resindraburiza]

Hi,
First of all, i want to thank you for this wonderful work. I can run it without problem when I specify

-c 0

for not running optimization. But when I want to do the optimization, I got segmentation error fault. When I do the backtrace, I get this:

Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". Core was generated by './runLine3Dpp_vsfm -i ../testdata/ -m ../testdata/vsfm_result.nvm'. Program terminated with signal SIGSEGV, Segmentation fault. #0 0x00007f8df05f5bdd in std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >::_M_assign(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) () from /usr/lib/x86_64-linux-gnu/libstdc++.so.6

My system is as follows:
Ubuntu 16.04
Boost 1.58
Cuda 8.0
Eigen 3.3.3
Ceres 2.0.0
OpenCV 3.4

Have you ever encounter this error? Thank you for your help.

[manhofer]

Hi Thanks! :-) I have not tried with Ceres 2.0.0 yet, it is possible that there is something different in this version, which is incompatible with my older code. Or it could be the case that CERES was not build as a shared lib? Make sure that CERES is definitely built as a shared lib, not a static one (you can specify this in the CERES CMakeLists.txt), and then try again. Otherwise I am not quite sure what the issue could be... But anyways, the optimization does not change the results by a very large margin. You should be fine without it. Best, Manuel

…

On Tue, Sep 25, 2018 at 9:39 AM Aji Resindra Widya ***@***.***> wrote: Hi, First of all, i want to thank you for this wonderful work. I can run it without problem when I specify -c 0 for not running optimization. But when I want to do the optimization, I got segmentation error fault. When I do the backtrace, I get this: Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". Core was generated by './runLine3Dpp_vsfm -i ../testdata/ -m ../testdata/vsfm_result.nvm'. Program terminated with signal SIGSEGV, Segmentation fault. #0 0x00007f8df05f5bdd in std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >::_M_assign(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) () from /usr/lib/x86_64-linux-gnu/libstdc++.so.6 My system is as follows: Ubuntu 16.04 Boost 1.58 Cuda 8.0 Eigen 3.3.3 Ceres 2.0.0 OpenCV 3.4 Have you ever encounter this error? Thank you for your help. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#19>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ANWX6kE2CICrFJ4qrxqrt0VlT0VMyfztks5ued21gaJpZM4W4CdQ> .

-- Dipl.-Ing. Dr. Manuel Hofer Sternäckerweg 101/4 8041 Graz AUSTRIA

[resindraburiza]

Hi Manuel,
thank you for your fast response.

I'm sure I compiled CERES as shared lib because it has these in output terminal
-- Installing: /usr/local/lib/libceres.so.2.0.0 -- Installing: /usr/local/lib/libceres.so.2 -- Installing: /usr/local/lib/libceres.so
more over I also tried older version of CERES (1.12) but still got segmentation fault.

But anyways, the optimization does not change the results by a very large
margin.

how big the difference between optimized reconstruction and non-optimized reconstruction?

[manhofer]

Hi Okay, then I am not really sure what the problem is... I can't really quantify the difference between unoptimized vs. optimized in exact numbers, but it really is not a very large difference. For most cases I would say it really is negligible. Best, Manuel

…

On Tue, Sep 25, 2018 at 12:25 PM Aji Resindra Widya < ***@***.***> wrote: Hi Manuel, thank you for your fast response. I'm sure I compiled CERES as shared lib because it has these in output terminal -- Installing: /usr/local/lib/libceres.so.2.0.0 -- Installing: /usr/local/lib/libceres.so.2 -- Installing: /usr/local/lib/libceres.so more over I also tried older version of CERES (1.12) but still got segmentation fault. But anyways, the optimization does not change the results by a very large margin. how big the difference between optimized reconstruction and non-optimized reconstruction? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#19 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ANWX6mhJVCUdNlacovX1dxR44Oe7ZJG6ks5uegR9gaJpZM4W4CdQ> .

-- Dipl.-Ing. Dr. Manuel Hofer Sternäckerweg 101/4 8041 Graz AUSTRIA

[resindraburiza]

Hi again.

I will try compiling on another machine and tell whether it works or not.

For most cases I would say it really is negligible.

The reconstruction result without optimized is actually good I think visually. Thank you.

I have another question but it's not related to the segmentation fault. I hope it's okay to ask it here:
How do you structure the matched 2D line segments? Is it possible if I want to extract the result of line segment matching?

Again, thank you very much.

[manhofer]

Hi Let me know if it works! You can have a look at the "Output" section of the readme file. The txt file that contains the result gives you the 3D lines, plus all 2D residuals. Which is basically the line matching result. Best, Manuel

…

On Thu, Sep 27, 2018 at 12:01 PM Aji Resindra Widya < ***@***.***> wrote: Hi again. I will try compiling on another machine and tell whether it works or not. I have another question but it's not related to the segmentation fault. I hope it's okay to ask it here: How do you structure the matched 2D line segments? Is it possible if I want to extract the result of line segment matching? Again, thank you very much. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#19 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ANWX6q3Sjo5Labb-a5WX3PMOl1K9a0yZks5ufKHngaJpZM4W4CdQ> .

-- Dipl.-Ing. Dr. Manuel Hofer Sternäckerweg 101/4 8041 Graz AUSTRIA

[resindraburiza]

Hi,

thank you for pointing out to the readme file. I've just read it. In my understanding, the data in that .txt file is the match that can be triangulated into 3D line. Am I correct? What if I need raw matches for each line segment from each image? Is it possible to access it?

Thank you very much.
Best regards,
Resindra

[manhofer]

Hi I am afraid that information is not contained in the output. Best, Manuel

…

On Thu, Sep 27, 2018 at 1:33 PM Aji Resindra Widya ***@***.***> wrote: Hi, thank you for pointing out to the readme file. I've just read it. In my understanding, the data in that .txt file is the match that can be triangulated into 3D line. Am I correct? What if I need raw matches for each line segment from each image? Is it possible to access it? Thank you very much. Best regards, Resindra — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#19 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ANWX6gnklIeqbimoCD0rEGPIvegA_fZgks5ufLeAgaJpZM4W4CdQ> .

-- Dipl.-Ing. Dr. Manuel Hofer Sternäckerweg 101/4 8041 Graz AUSTRIA

[MojEKi]

Dear Dr Hofer
could you please Check your Email: hofer@icg.tugraz.at;
I am looking forward to receiving your response!
Best regards,
Mojdeh

[manhofer]

Hi I am afraid that email address does no longer exist, since I left Graz University of Technology several years ago. Please use this mail address instead: man.hofer@gmail.com <mailto:man.hofer@gmail.com> Best, Manuel

…

Am 15.02.2020 um 10:42 schrieb MojEKi ***@***.***>: Dear Dr Hofer could you please Check your Email: ***@***.*** ***@***.***>; I am looking forward to receiving your response! Best regards, Mojdeh — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=ADKZP2RPOACT6OU4GYJITWDRC62J5A5CNFSM4FXAE5IKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEL3GJDA#issuecomment-586572940>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADKZP2R2PTTET3K6K4LC7PDRC62J5ANCNFSM4FXAE5IA>.