Colmap fails on DTU dataset scan30 for 3 input views

[NagabhushanSN95]

Hi, I tried running colmap using imgs2poses.py for 3 input views on the DTU dataset. While it runs for most of the scans, it is failing for scan30. I used the frames 22,25,28. I have attached the colmap_output.txt below. Can you tell me how you were able to run colmap on this scan?

==============================================================================
Feature extraction
==============================================================================

Processed file [1/3]
  Name:            0022.png
  Dimensions:      400 x 300
  Camera:          #1 - SIMPLE_RADIAL
  Focal Length:    480.00px
  Features:        311
Processed file [2/3]
  Name:            0025.png
  Dimensions:      400 x 300
  Camera:          #1 - SIMPLE_RADIAL
  Focal Length:    480.00px
  Features:        262
Processed file [3/3]
  Name:            0028.png
  Dimensions:      400 x 300
  Camera:          #1 - SIMPLE_RADIAL
  Focal Length:    480.00px
  Features:        289
Elapsed time: 0.002 [minutes]

==============================================================================
Exhaustive feature matching
==============================================================================

Matching block [1/1, 1/1] in 0.006s
Elapsed time: 0.000 [minutes]

==============================================================================
Loading database
==============================================================================

Loading cameras... 1 in 0.000s
Loading matches... 0 in 0.000s
Loading images... 3 in 0.000s (connected 0)
Building correspondence graph... in 0.000s (ignored 0)

Elapsed time: 0.000 [minutes]

WARNING: No images with matches found in the database.

PS: I also tried sparse reconstruction with known poses which also failed.

[dunbar12138]

Did you get the same output even when you provide the poses?

[NagabhushanSN95]

I don't remember the exact output. I remember that it failed. I'll check once more and update the output here

[violet257]

Hi, I meet the same problem, colmap fails on 3 views DTU of scan30 and scan110, I wonder whether you solved this and how?

[NagabhushanSN95]

Nope, didn't work for me. I just remembered that I need to add the output. I'll add the output so that the authors can help the future users.

[NagabhushanSN95]

@dunbar12138 here is the output when I provide the poses.

==============================================================================
Feature extraction
==============================================================================

Processed file [1/3]
  Name:            0000.png  (0022.png renamed to 0000.png)
  Dimensions:      400 x 300
  Camera:          #1 - SIMPLE_RADIAL
  Focal Length:    480.00px
  Features:        311
Processed file [2/3]
  Name:            0001.png  (0025.png renamed to 0001.png)
  Dimensions:      400 x 300
  Camera:          #1 - SIMPLE_RADIAL
  Focal Length:    480.00px
  Features:        262
Processed file [3/3]
  Name:            0002.png  (0028.png renamed to 0002.png)
  Dimensions:      400 x 300
  Camera:          #1 - SIMPLE_RADIAL
  Focal Length:    480.00px
  Features:        289
Elapsed time: 0.010 [minutes]
colmap exhaustive_matcher --database_path ../../tmp/database.db

==============================================================================
Exhaustive feature matching
==============================================================================

Matching block [1/1, 1/1] in 0.006s
Elapsed time: 0.002 [minutes]
colmap point_triangulator --database_path ../../tmp/database.db --image_path ../../tmp/images --input_path ../../tmp/sparse/0 --output_path ../../tmp/sparse/0 --Mapper.tri_ignore_two_view_tracks 0 --Mapper.num_threads 16 --Mapper.init_min_tri_angle 4 --Mapper.multiple_models 0 --Mapper.extract_colors 0

==============================================================================
Loading model
==============================================================================

==============================================================================
Loading database
==============================================================================

Loading cameras... 1 in 0.000s
Loading matches... 0 in 0.000s
Loading images... 3 in 0.000s (connected 0)
Building correspondence graph... in 0.000s (ignored 0)

Elapsed time: 0.000 [minutes]

==============================================================================
Triangulating image #1 (0)
==============================================================================

  => Image sees 0 / 0 points
  => Triangulated 0 points

==============================================================================
Triangulating image #2 (1)
==============================================================================

  => Image sees 0 / 0 points
  => Triangulated 0 points

==============================================================================
Triangulating image #3 (2)
==============================================================================

  => Image sees 0 / 0 points
  => Triangulated 0 points

==============================================================================
Retriangulation
==============================================================================

  => Merged observations: 0
  => Completed observations: 0

==============================================================================
Bundle adjustment
==============================================================================

F0220 10:12:56.509768 621392 colmap.cc:1573] Check failed: bundle_adjuster.Solve(&reconstruction) 
*** Check failure stack trace: ***
    @     0x7f7b529301c3  google::LogMessage::Fail()
    @     0x7f7b5293525b  google::LogMessage::SendToLog()
    @     0x7f7b5292febf  google::LogMessage::Flush()
    @     0x7f7b529306ef  google::LogMessageFatal::~LogMessageFatal()
    @     0x55df43977280  RunPointTriangulator()
    @     0x55df4396eeaf  main
    @     0x7f7b50d30083  __libc_start_main
    @     0x55df43972f6e  _start

PS: sorry for late response. Lemme know if you need any other logs.