tsattler/meshloc_release

meshloc_release's Issues

undefined symbol: __pow_finite while importing meshloc

@tsattler
Hi, thank you for the great work! I've come across an issue while importing meshloc in localize.py. There is no error while building meshloc from this repo, but an error occurs when running localize.py:
Traceback (most recent call last): File "localize.py", line 48, in <module> import meshloc ImportError: /home/weientai18/meshloc_release/meshloc.cpython-38-x86_64-linux-gnu.so: undefined symbol: __pow_finite
I wonder if you have encountered similar issues before?
Here are some outputs when running python setup.py develop:

running develop
/home/weientai18/anaconda3/envs/immatch1/lib/python3.8/site-packages/setuptools/command/develop.py:40: EasyInstallDeprecationWarning: easy_install command is deprecated.
!!
********************************************************************************
Please avoid running setup.py and easy_install.
Instead, use pypa/build, pypa/installer or other
standards-based tools.
See pypa/setuptools#917 for details.
********************************************************************************
!!
easy_install.initialize_options(self)
/home/weientai18/anaconda3/envs/immatch1/lib/python3.8/site-packages/setuptools/_distutils/cmd.py:66: SetuptoolsDeprecationWarning: setup.py install is deprecated.
!!
********************************************************************************
Please avoid running setup.py directly.
Instead, use pypa/build, pypa/installer or other
standards-based tools.
See https://blog.ganssle.io/articles/2021/10/setup-py-deprecated.html for details.
********************************************************************************
!!
self.initialize_options()
running egg_info
writing meshloc.egg-info/PKG-INFO
writing dependency_links to meshloc.egg-info/dependency_links.txt
writing top-level names to meshloc.egg-info/top_level.txt
reading manifest file 'meshloc.egg-info/SOURCES.txt'
adding license file 'LICENSE'
writing manifest file 'meshloc.egg-info/SOURCES.txt'
running build_ext
['cmake', '/home/weientai18/meshloc_release', '-DCMAKE_LIBRARY_OUTPUT_DIRECTORY=/home/weientai18/meshloc_release', '-DPYTHON_EXECUTABLE=/home/weientai18/anaconda3/envs/immatch1/bin/python', '-DCMAKE_BUILD_TYPE=Release']
-- pybind11 v2.11.0 dev1
-- Found AMD headers in: /usr/include/suitesparse
-- Found AMD library: /usr/lib/x86_64-linux-gnu/libamd.so
-- Found CAMD headers in: /usr/include/suitesparse
-- Found CAMD library: /usr/lib/x86_64-linux-gnu/libcamd.so
-- Found CCOLAMD headers in: /usr/include/suitesparse
-- Found CCOLAMD library: /usr/lib/x86_64-linux-gnu/libccolamd.so
-- Found CHOLMOD headers in: /usr/include/suitesparse
-- Found CHOLMOD library: /usr/lib/x86_64-linux-gnu/libcholmod.so
-- Found COLAMD headers in: /usr/include/suitesparse
-- Found COLAMD library: /usr/lib/x86_64-linux-gnu/libcolamd.so
-- Found SPQR headers in: /usr/include/suitesparse
-- Found SPQR library: /usr/lib/x86_64-linux-gnu/libspqr.so
-- Found Config headers in: /usr/include/suitesparse
-- Found Config library: /usr/lib/x86_64-linux-gnu/libsuitesparseconfig.so
-- Did not find Intel TBB library, assuming SuiteSparseQR was not compiled with TBB.
-- Adding librt to SuiteSparse_config libraries (required on Linux & Unix [not OSX] if SuiteSparse is compiled with timing).
-- Found required Ceres dependency: Eigen version 3.4.0 in /home/weientai18/anaconda3/envs/immatch1/share/eigen3/cmake
-- Found required Ceres dependency: glog
-- Found required Ceres dependency: gflags
-- Found Ceres version: 2.1.0 installed in: /home/weientai18/anaconda3/envs/immatch1 with components: [EigenSparse, SparseLinearAlgebraLibrary, LAPACK, SuiteSparse, CXSparse, SchurSpecializations, Multithreading]
-- Found AMD headers in: /usr/include/suitesparse
-- Found AMD library: /usr/lib/x86_64-linux-gnu/libamd.so
-- Found CAMD headers in: /usr/include/suitesparse
-- Found CAMD library: /usr/lib/x86_64-linux-gnu/libcamd.so
-- Found CCOLAMD headers in: /usr/include/suitesparse
-- Found CCOLAMD library: /usr/lib/x86_64-linux-gnu/libccolamd.so
-- Found CHOLMOD headers in: /usr/include/suitesparse
-- Found CHOLMOD library: /usr/lib/x86_64-linux-gnu/libcholmod.so
-- Found COLAMD headers in: /usr/include/suitesparse
-- Found COLAMD library: /usr/lib/x86_64-linux-gnu/libcolamd.so
-- Found SPQR headers in: /usr/include/suitesparse
-- Found SPQR library: /usr/lib/x86_64-linux-gnu/libspqr.so
-- Found Config headers in: /usr/include/suitesparse
-- Found Config library: /usr/lib/x86_64-linux-gnu/libsuitesparseconfig.so
-- Did not find Intel TBB library, assuming SuiteSparseQR was not compiled with TBB.
-- Adding librt to SuiteSparse_config libraries (required on Linux & Unix [not OSX] if SuiteSparse is compiled with timing).
-- Found required Ceres dependency: Eigen version 3.4.0 in /home/weientai18/anaconda3/envs/immatch1/share/eigen3/cmake
-- Found required Ceres dependency: glog
-- Found required Ceres dependency: gflags
-- Found Ceres version: 2.1.0 installed in: /home/weientai18/anaconda3/envs/immatch1 with components: [EigenSparse, SparseLinearAlgebraLibrary, LAPACK, SuiteSparse, CXSparse, SchurSpecializations, Multithreading]
-- Found Boost: /home/weientai18/anaconda3/envs/immatch1/lib/cmake/Boost-1.78.0/BoostConfig.cmake (found version "1.78.0") found components: program_options filesystem system unit_test_framework
-- Found Eigen
-- Includes : /usr/include/eigen3
-- Found FLANN
-- Includes : /usr/include
-- Libraries : /usr/lib/x86_64-linux-gnu/libflann.so
-- Found LZ4
-- Includes : /usr/include
-- Libraries : /usr/lib/x86_64-linux-gnu/liblz4.so
-- Found FreeImage
-- Includes : /usr/include
-- Libraries : /usr/lib/x86_64-linux-gnu/libfreeimage.so
-- Found Glog
-- Includes : /usr/include
-- Libraries : /usr/lib/x86_64-linux-gnu/libglog.so
-- Found SQLite3
-- Includes : /usr/include
-- Libraries : /usr/lib/x86_64-linux-gnu/libsqlite3.so
-- Found Glew
-- Includes : /usr/include
-- Libraries : /usr/lib/x86_64-linux-gnu/libGLEW.so
-- Using header-only CGAL
-- Targetting Unix Makefiles
-- Using /usr/bin/c++ compiler.
-- Found Boost: /home/weientai18/anaconda3/envs/immatch1/lib/cmake/Boost-1.78.0/BoostConfig.cmake (found suitable version "1.78.0", minimum required is "1.48")
-- Boost include dirs: /home/weientai18/anaconda3/envs/immatch1/include
-- Boost libraries:
-- Using gcc version 4 or later. Adding -frounding-math
-- Configuring done
-- Generating done
-- Build files have been written to: /home/weientai18/meshloc_release/build/temp.linux-x86_64-cpython-38
Consolidate compiler generated dependencies of target meshloc
[100%] Built target meshloc
Creating /home/weientai18/anaconda3/envs/immatch1/lib/python3.8/site-packages/meshloc.egg-link (link to .)
meshloc 0.1.0 is already the active version in easy-install.pth
Installed /home/weientai18/meshloc_release
Processing dependencies for meshloc==0.1.0
Finished processing dependencies for meshloc==0.1.0

NetVLAD Image Retrieval Code Part

Hello,

I was just wondering do you plan to release the netvlad-based code part that was used for image retrieval ?

Thanks!

The accuracy of the final acquired pose of MeshLoc

Hello, dear author! This work is very wonderful.
I tried to experiment with the pose of some images in the 12scens_kitchen file.
I found that the final meshloc obtained quaternions are basically the same as the true value (but the first W appears negative， XYZ is similar).
However, the position difference is large.
I tried to use d2net and LOFTER, and the final result is similar, with a big position gap.
Here is my parameter configuration,

what is wrong ?

Here are the quaternions obtained from the rotation matrix of pose.txt of frame-000000-colar.jpg

Unable to achieve accuracy shown in Benchmark

Hi tsattler, The results obtained by using parameters given in README (match method is LoFTR) can not achieve the accuracy shown in the benchmark, especially at (0.25m, 2 degrees) level. Would you please help us solve it?
Looking forward to your reply.

Apply for render algorithm code

Hi Tsattler,

It is an impressive works. I see the render code in meshloc is not available anymore(https://arxiv.org/pdf/1601.06950.pdf), but its very important for my reaserch. I would very appriciate, if you could provide me the MeshLoc's render algorithm. Or, there is another render engine/algorithms is avaliable (as you know).

Thank you so much! Looking forward to your reply.

How to run MeshLoc on own data

Hi，tsattler
The content of your research is exciting.
How to run MeshLoc on own data?

Question about how to generate the mesh about aachen?

error when run code on Aachen

thanks to your excellent work!

I am trying to run meshloc on Aachen dataset, which was downloaded using the script
bash download_meshloc_data.sh -n aachen

and the error below occur:

Traceback (most recent call last):
File "../meshloc_release/localize.py", line 624, in
main()
File "../meshloc_release/localize.py", line 450, in main
img2_id = map_db_name_to_id[retrieved_db[j]]
KeyError: 'db_1883.jpg'

Could you please help me fix it?

How to render a large-scale scene in real-time?

Hi Tsattler,

Thanks for your good work. I notice you mention that large 3D models can be rendered in less than a millisecond in paper.
However, when I used pyrender and pytorch3d to render a large-scale scene, it usually took several seconds or longer.
Would you please tell me what tool do you use to render such a large-scale scene?

Looking forward to your reply.

I read the paper mentioning Position averaging, but found nothing about it in the code?

@v-pnk @tsattler
Hi,I'm confused about the underlined part, since C_i_ is a sampling point based on C (virtual), how do I go about counting the number of inliers points in 2d-3D？

how the depth map is obtained？

@tsattler @v-pnk
Hi,Seeing that the code is essentially based on the relocation of the depth map, I am very confused by how the depth map is obtained?Suppose that even if the lidar back-projection image obtains the depth map, the depth map is sparse or has holes. It is difficult to ensure that the matched feature points have depth. So I want to know how to get the depth map and whether to use the algorithm of depth map completion?
Thanks