search

cdcseacave / openMVS_sample

OpenMVS usage example
GNU Affero General Public License v3.0
101 stars 50 forks source link

Always failed to DensifyPointCloud with some default params on openMVS_sample dataset #11

Open CeciliaPYY opened 4 years ago

CeciliaPYY commented 4 years ago

Here is my script,

dataset_out=/dataset/openMVS_sample-master

# Initialize the scene (spherical camera model)
openMVG_main_SfMInit_ImageListing -i $dataset -o $dataset_out/matches -f 3.4

# Compute of the features (you can add -n <THREAD_COUNT> to be faster)
openMVG_main_ComputeFeatures -i $dataset_out/matches/sfm_data.json -o $dataset_out/matches -n 2 

# Compute the image matches (notice the angular mode, more adapted for spherical images)
openMVG_main_ComputeMatches -i $dataset_out/matches/sfm_data.json -o $dataset_out/matches  

# Compute of the camera motion and structure of the scene
openMVG_main_IncrementalSfM -i $dataset_out/matches/sfm_data.json -m $dataset_out/matches -o $dataset_out/reconstruction

#
# MVS (optional)
#

# Convert the scene from OpenMVG to OpenMVS data format
openMVG_main_openMVG2openMVS -i $dataset_out/reconstruction/sfm_data.bin -o $dataset_out/reconstruction/scene.mvs -d $dataset_out/reconstruction/openmvs_images

And here is my DensifyPointCloud-2005241555208B4734.log

15:55:20 [App     ] Build date: May 20 2020, 12:54:10
15:55:20 [App     ] CPU: Intel(R) Core(TM) i5-7360U CPU @ 2.30GHz (2 cores)
15:55:20 [App     ] RAM: 7.79GB Physical Memory 1024.00MB Virtual Memory
15:55:20 [App     ] OS: Linux 4.9.184-linuxkit (x86_64)
15:55:20 [App     ] SSE & AVX compatible CPU & OS detected
15:55:20 [App     ] Command line: /dataset/openMVS_sample-master/reconstruction/scene.mvs --fusion-mode -2
15:55:20 [App     ] Camera model loaded: platform 0; camera  0; f 0.001x0.001; poses 2
15:55:20 [App     ] warning: uncalibrated image '/dataset/openMVS_sample-master/reconstruction/openmvs_images/00000.jpg'
15:55:20 [App     ] warning: uncalibrated image '/dataset/openMVS_sample-master/reconstruction/openmvs_images/00001.jpg'
15:55:20 [App     ] warning: uncalibrated image '/dataset/openMVS_sample-master/reconstruction/openmvs_images/00002.jpg'
15:55:20 [App     ] warning: uncalibrated image '/dataset/openMVS_sample-master/reconstruction/openmvs_images/00003.jpg'
15:55:20 [App     ] Image loaded   4: 00004.jpg
15:55:20 [App     ] warning: uncalibrated image '/dataset/openMVS_sample-master/reconstruction/openmvs_images/00005.jpg'
15:55:20 [App     ] warning: uncalibrated image '/dataset/openMVS_sample-master/reconstruction/openmvs_images/00006.jpg'
15:55:20 [App     ] Image loaded   7: 00007.jpg
15:55:20 [App     ] warning: uncalibrated image '/dataset/openMVS_sample-master/reconstruction/openmvs_images/00008.jpg'
15:55:20 [App     ] warning: uncalibrated image '/dataset/openMVS_sample-master/reconstruction/openmvs_images/00009.jpg'
15:55:20 [App     ] warning: uncalibrated image '/dataset/openMVS_sample-master/reconstruction/openmvs_images/00010.jpg'
15:55:20 [App     ] Scene loaded from interface format (17ms):
    11 images (2 calibrated) with a total of 11.49 MPixels (5.75 MPixels/image)
    257 points, 0 vertices, 0 faces
15:55:20 [App     ] Preparing images for dense reconstruction completed: 11 images (62ms)
15:55:20 [App     ] error: reference image   4 has no good images in view
15:55:20 [App     ] error: reference image   7 has no good images in view
15:55:20 [App     ] Selecting images for dense reconstruction completed: 0 images (1ms)
15:55:20 [App     ] Depth-maps fused and filtered: 0 depth-maps, 0 depths, 0 points (-2147483648%%) (0ms)
15:55:20 [App     ] Densifying point-cloud completed: 0 points (77ms)
15:55:20 [App     ] Scene saved (2ms):
    11 images (2 calibrated)
    0 points, 0 vertices, 0 faces
15:55:20 [App     ] MEMORYINFO: {
15:55:20 [App     ]     VmPeak:   795104 kB
15:55:20 [App     ]     VmSize:   795104 kB
15:55:20 [App     ] } ENDINFO

Could anyone help me to find out why???

CeciliaPYY commented 4 years ago

However, I got quite good result with COLMAP

cdcseacave commented 4 years ago

the SfM reconstruction (the input to MVS) is not correct

CeciliaPYY commented 4 years ago

the SfM reconstruction (the input to MVS) is not correct

Sorry to interrupt, but what do you mean by reconstruction is not correct, I quite follow the MvgMvsPipeline.py 's SEQUENTIAL pipeline.

cdcseacave commented 4 years ago

pls check in Viewer

CeciliaPYY commented 4 years ago

I happened to use docker due to some building errors, which means I didn't use Viewer but CCViewer instead. I also tried to run tutorial_demo.py on my dataset, which from my point of view, two datasets share same attributes, such as large overlap between images, however nothing come out under reconstruction_sequential directory, and the result under reconstruction_global is very strange.