Segmentation fault (core dumped) with sample

[ErfolgreichCharismatisch]

I am using WSL under Windows 10 and MobaXterm on Windows 10.

In MobaXterm I start

bin/dso_dataset files=../sequence_13/images.zip calib=../sequence_13/camera.txt gamma=../sequence_13/pcalib.txt vignette=../sequence_13/vignette.png preset=0 mode=0

and I get

for 3 seconds to then get

loading data from ../sequence_13/images.zip!
loading calibration from ../sequence_13/camera.txt!
loading gammaCalib from ../sequence_13/pcalib.txt!
loading vignette from ../sequence_13/vignette.png!

=============== PRESET Settings: ===============
DEFAULT settings:
- no  real-time enforcing
- 2000 active points
- 5-7 active frames
- 1-6 LM iteration each KF
- original image resolution
==============================================
PHOTOMETRIC MODE WITH CALIBRATION!
got 1800 entries and 1800 files!
Reading Calibration from file ../sequence_13/camera.txt ... found!
found ATAN camera model, building rectifier.
Creating FOV undistorter
Input resolution: 1280 1024
In: 0.349153 0.436593 0.493140 0.499021 0.933271

Found fx=0.349153, fy=0.436593, cx=0.493140, cy=0.499021.
 I'm assuming this is the "relative" calibration file format,and will rescale this by image width / height to fx=446.915840, fy=447.071232, cx=630.719200, cy=510.497504.

Out: 0.400000 0.530000 0.500000 0.500000 0.000000
Output resolution: 640 480

Rectified Kamera Matrix:
  256     0 319.5
    0 254.4 239.5
    0     0     1

Reading Photometric Calibration from file ../sequence_13/pcalib.txt
Reading Vignette Image from ../sequence_13/vignette.png
Successfully read photometric calibration!
got 1800 images and 1800 timestamps and 1800 exposures.!
ImageFolderReader: got 1800 files in ../sequence_13/images.zip!
using pyramid levels 0 to 3. coarsest resolution: 80 x 60!
START PANGOLIN!
X11 Error: GLXBadFBConfig
Pangolin X11: Indirect GLX rendering context obtained
Initialization: keep 21.7% (need 2000, have 9201)!
INITIALIZE FROM INITIALIZER (2047 pts)!
SPARSITY:  MinActDist 2.100000 (need 2000 points, have 2047 points)!
OPTIMIZE 2047 pts, 2047 active res, 0 lin res!
Initial Error           A(539373.606970)=(AV inf). Num: A(0) + M(0); ab 0.000000 0.000000!
STEPS: A 0.0; B 2.7; R 12.4; T 33.7.    REJECT 0 (L -1.00, dir nan, ss 1.0):    A(678364.091956)=(AV 8.686). Num: A(1124) + M(0); ab -0.000020 -0.228851!
STEPS: A 0.0; B 1.2; R 3.8; T 12.8.     ACCEPT 1 (L -1.60, dir 0.97, ss 1.0):   A(590887.398140)=(AV 7.222). Num: A(1416) + M(0); ab -0.000030 -0.122949!
STEPS: A 0.0; B 0.3; R 0.9; T 2.3.      ACCEPT 2 (L -2.20, dir 0.52, ss 1.0):   A(531791.727600)=(AV 6.847). Num: A(1418) + M(0); ab -0.000032 -0.094246!
STEPS: A 0.0; B 0.0; R 0.8; T 2.0.      ACCEPT 3 (L -2.81, dir 0.89, ss 1.0):   A(518311.553136)=(AV 6.810). Num: A(1397) + M(0); ab -0.000031 -0.091141!
STEPS: A 0.0; B 0.5; R 0.2; T 0.6.      ACCEPT 4 (L -3.41, dir 0.55, ss 1.0):   A(512591.349404)=(AV 6.785). Num: A(1392) + M(0); ab -0.000029 -0.136543!
LOG 8: 6.773 fine. Res: 1392 A, 0 L, 0 M; (0 / 0) forceDrop. a=-0.000029, b=-0.136543. Window 8 (2)
Segmentation fault (core dumped)

Is that a dso or MobaXterm problem?

[ErfolgreichCharismatisch]

Hello?

[jskinn]

Hi, this could well be a DSO issue, it is hard to say. Unfortunately, segmentation faults are not big on contextual information. My approach to debugging this would be to add some noisy prints to DSO (every frame) to try and work out what state it is in when the segfault occurs, and if it can be reproduced consistently.

[shuhul]

Were you able to solve the segfault issue?