search

meidachen / STPLS3D

🔥 Synthetic and real-world 2d/3d dataset for semantic and instance segmentation (BMVC 2022 Oral)
236 stars 20 forks source link

training with randla #25

Open hongzhiqiang67 opened 1 year ago

hongzhiqiang67 commented 1 year ago

Hi, I would like to use your RandLA-Net code to achiev，In main_STPLS3D.py you write to RealWorldData in data_folder,ResidentialArea_GT in test_area.ResidentialArea_GT Which partial dataset is this? I try to train with RealWorldData four regions, but the input_0.300 and original_ply folders I get after preprocessing are empty。 Can you answer me, thanks a lot！thank you

meidachen commented 1 year ago

Hi @hongzhiqiang67 ,

RandLA was implemented using our old data pipeline, basically, it used the data in txt format instead of the current ply format. I will later try to update it to use the ply files. But in the meanwhile, please try the txt data in the following link: https://drive.google.com/file/d/17nUphM7LpmR3tT9eX0M1SXfIVjidr0-T/view?usp=sharing

Best,

Meida Chen

hongzhiqiang67 commented 1 year ago

你好@hongzhiqiang67,

RandLA 是使用我们旧的数据管道实现的，基本上，它使用 txt 格式的数据而不是当前的 ply 格式。我稍后会尝试更新它以使用 ply 文件。但同时，请尝试以下链接中的 txt 数据： https ://drive.google.com/file/d/17nUphM7LpmR3tT9eX0M1SXfIVjidr0-T/view?usp=sharing

最好的，

陈美达

When i run the main file i get this error，AttributeError: module 'utils.nearest_neighbors.lib.python' has no attribute 'knn_batch'。The error is in randlanet file line 162。

The environment I use is tensorflow2.2.0，CUDA10.1，python3.8，Does this affect？

meidachen commented 1 year ago

Were you able to run the official RandLA using your environment?

hongzhiqiang67 commented 1 year ago

ok,thank you

Meida Chen @.***> 于2022年8月12日周五 00:30写道：

Hi @hongzhiqiang67 https://github.com/hongzhiqiang67 ,

RandLA was implemented using our old data pipeline, basically, it used the data in txt format instead of the current ply format. I will later try to update it to use the ply files. But in the meanwhile, please try the txt data in the following link:

https://drive.google.com/file/d/17nUphM7LpmR3tT9eX0M1SXfIVjidr0-T/view?usp=sharing

Best,

Meida Chen

— Reply to this email directly, view it on GitHub https://github.com/meidachen/STPLS3D/issues/25#issuecomment-1212216729, or unsubscribe https://github.com/notifications/unsubscribe-auth/AXBEF3VHTSX37WHMDU2EYY3VYUTC7ANCNFSM56ILEX7A . You are receiving this because you were mentioned.Message ID: @.***>

hongzhiqiang67 commented 1 year ago

Hello, I want to cite your dataset, can you tell me where in the United States your data was collected, I want to make a graph

Meida Chen @.***> 于2022年8月14日周日 04:44写道：

Were you able to run the official RandLA using your environment?

— Reply to this email directly, view it on GitHub https://github.com/meidachen/STPLS3D/issues/25#issuecomment-1214220107, or unsubscribe https://github.com/notifications/unsubscribe-auth/AXBEF3VAZRT66443UCP5PULVZACKNANCNFSM56ILEX7A . You are receiving this because you were mentioned.Message ID: @.***>

meidachen commented 1 year ago

Sure, USC: 34.020178, -118.288826 WMSC: 33.444650, -118.481350 OCCC: 28.429107, -81.461848 RA: 33.874179, -118.073325

hongzhiqiang67 commented 1 year ago

Thank you very much, your data is point cloud data reconstructed by drone aerial photogrammetry, do you still have original photo data, because I need to do accuracy verification, which involves specific building coordinates. Can you send me a photo?

meidachen commented 1 year ago

Which data set are you interested in? (USC, RA, synthetic ones etc)

From: hongzhiqiang67 @.> Sent: Saturday, November 5, 2022 2:20 AM To: meidachen/STPLS3D @.> Cc: Meida Chen @.>; Comment @.> Subject: Re: [meidachen/STPLS3D] training with randla (Issue #25)

Thank you very much, your data is point cloud data reconstructed by drone aerial photogrammetry, do you still have original photo data, because I need to do accuracy verification, which involves specific building coordinates. Can you send me a photo?

— Reply to this email directly, view it on GitHubhttps://urldefense.com/v3/__https://github.com/meidachen/STPLS3D/issues/25*issuecomment-1304453134__;Iw!!LIr3w8kk_Xxm!udQEQEeMiXZO1HULRZP-EdhOmE3dKXgfM4vCz6t8krXcaD64bX0hWP-LVoSYVdDBj1qW_Ii1wUGhuJZx9vpZ1fk$, or unsubscribehttps://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/AGDV6LTPRB43EQW2LVMCZCLWGYRFNANCNFSM56ILEX7A__;!!LIr3w8kk_Xxm!udQEQEeMiXZO1HULRZP-EdhOmE3dKXgfM4vCz6t8krXcaD64bX0hWP-LVoSYVdDBj1qW_Ii1wUGhuJZxT5nuUbk$. You are receiving this because you commented.Message ID: @.***>

hongzhiqiang67 commented 1 year ago

these four，thanks

USC: 34.020178, -118.288826 WMSC: 33.444650, -118.481350 OCCC: 28.429107, -81.461848 RA: 33.874179, -118.073325

Meida Chen @.***> 于2022年11月8日周二 03:04写道：

Which data set are you interested in? (USC, RA, synthetic ones etc)

From: hongzhiqiang67 @.> Sent: Saturday, November 5, 2022 2:20 AM To: meidachen/STPLS3D @.> Cc: Meida Chen @.>; Comment @.> Subject: Re: [meidachen/STPLS3D] training with randla (Issue #25)

Thank you very much, your data is point cloud data reconstructed by drone aerial photogrammetry, do you still have original photo data, because I need to do accuracy verification, which involves specific building coordinates. Can you send me a photo?

— Reply to this email directly, view it on GitHub< https://urldefense.com/v3/__https://github.com/meidachen/STPLS3D/issues/25*issuecomment-1304453134__;Iw!!LIr3w8kk_Xxm!udQEQEeMiXZO1HULRZP-EdhOmE3dKXgfM4vCz6t8krXcaD64bX0hWP-LVoSYVdDBj1qW_Ii1wUGhuJZx9vpZ1fk$>, or unsubscribe< https://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/AGDV6LTPRB43EQW2LVMCZCLWGYRFNANCNFSM56ILEX7A__;!!LIr3w8kk_Xxm!udQEQEeMiXZO1HULRZP-EdhOmE3dKXgfM4vCz6t8krXcaD64bX0hWP-LVoSYVdDBj1qW_Ii1wUGhuJZxT5nuUbk$

. You are receiving this because you commented.Message ID: @.***>

— Reply to this email directly, view it on GitHub https://github.com/meidachen/STPLS3D/issues/25#issuecomment-1306056815, or unsubscribe https://github.com/notifications/unsubscribe-auth/AXBEF3WYBD7TYA6JGMITYIDWHFHCFANCNFSM56ILEX7A . You are receiving this because you were mentioned.Message ID: @.***>

meidachen commented 1 year ago

Hi @hongzhiqiang67 , I have uploaded all source images; please check the main page for the download link.

hongzhiqiang67 commented 1 year ago

Thank you!

Meida Chen @.***> 于2022年11月30日周三 08:01写道：

Hi @hongzhiqiang67 https://github.com/hongzhiqiang67 , I have uploaded all source images; please check the main page for the download link.

— Reply to this email directly, view it on GitHub https://github.com/meidachen/STPLS3D/issues/25#issuecomment-1331473537, or unsubscribe https://github.com/notifications/unsubscribe-auth/AXBEF3VLDKQAFEOS3YFB6S3WK2KMBANCNFSM56ILEX7A . You are receiving this because you were mentioned.Message ID: @.***>