search

tudelft3d / SUMS-Semantic-Urban-Mesh-Segmentation-public

SUMS: Semantic Urban Mesh Segmentation.
GNU General Public License v3.0
56 stars 14 forks source link

Building CGAL problem #18

Closed Roy2xc closed 1 year ago

Roy2xc commented 1 year ago

Hello, I'm using the CGAL provided by your code, and it sames not work. 20230804165047

I'm very confused what's the problem, do you have any advise?

Thank you.

WeixiaoGao commented 1 year ago

Hi, Have you configured the CGAL path in CMake using the CGAL we provided? If so, please also check the boost iostreams and serialization packages.

Roy2xc commented 1 year ago

Thank you! That's the reason,. and now I tried to run the code, my command is : .\semantic_urban_mesh_segmentation.exe F:\my\SUMS-Semantic-Urban-Mesh-Segmentation-public-main\build\semantic_urban_mesh_segmentation\Release\config.txt But the program just end without any report...

Roy2xc commented 1 year ago

Should I build the folders like your datasets?

WeixiaoGao commented 1 year ago

You only need to keep the path without the filename "config.txt" as it always stays the same. Yes, you need to create folders.

Roy2xc commented 1 year ago

Hi, Dr.Gao, I got segments result but cannot find the prediction output... there are only some textures'jpg file in my output folder...

And another question is how to batch or packag download from the site you provide? I'm only able to download them one by one...

Thank you for your reply!

Roy2xc commented 1 year ago

I mean what should i do to get the final result? It seems i finally got a intermediate results of segments rather than predictions.

Roy2xc commented 1 year ago

I set the mode to "Test_config" and get a fairly good result, but it still cannot do batch operation, only the first file in the folder was processed。

WeixiaoGao commented 1 year ago

Hi, If you put your data correctly in the input folder (.ply with .jpg), the program will find and process all files there. You also need to make sure the type of folder in the config.txt is turned on like process_train = true. You can use some batch link download tools to download all folders and files.