Towards Semantic Segmentation of Urban-Scale 3D Point Clouds: A Dataset, Benchmarks and Challenges

This is the official repository of the SensatUrban dataset. For technical details, please refer to:

Towards Semantic Segmentation of Urban-Scale 3D Point Clouds: A Dataset, Benchmarks and Challenges
Qingyong Hu, Bo Yang*, Sheikh Khalid, Wen Xiao, Niki Trigoni, Andrew Markham.
[Paper] [Blog] [Video] [Project page] [Download] [Evaluation]

🔥 原始说明此处不再赘述。本文档仅用于记录配置使用情况。

Environment Configuration （Ubuntu18）

0、环境说明

本人环境是UBUNTU18 + CUDA10.0 + cudnn7.4 + tensorflow1.14

🔥 注意：
1、python3.6+cuda10+tensorflow1.14验证跑通CPU运行环境。
2、python3.6+cuda11（RTX3090）+tensorflow2.6+keras2.6下配置跑通GPU运行环境，代码需要修改tensorflow库调用，改用tensorflow1版本的书写方式调用。RTX30系显卡支持最低cuda版本为cuda11，tensorflow1.14不支持cuda11。

1、下载数据集

Download the files named "data_release.zip" here. 本人目录结构：Randlanet（即代码根目录）/Dataset/SensatUrban/train & test.

2、创建环境

conda create -n randlanet python=3.5
source activate randlanet
pip install -r helper_requirements.txt

🔥 这里没找到pyyaml的5.4版本，直接执行的“pip install pyyaml”，默认安装了5.3.1版本。

执行以下批处理，处理cpp库

sh compile_op.sh

🔥 注意：
1、安装g++和gcc，保证两者版本一致。
2、utils文件夹需要在正确位置上，否则生成库有问题。
3、经过验证，python版本不同编译的库不可通用，如python3.5版本编译的不可再python3.6下使用；但ubuntu18编译的库可在ubuntu20下使用。

3、数据预处理

执行批处理之前，需要安装以下库，否则会出错：

pip install cython
pip install numpy
pip install sklearn
pip install open3d
pip install tensorflow-gpu==1.14

如下运行，执行数据预处理。

python input_preparation.py --dataset_path $YOURPATH
cd $YOURPATH
cd ../
mkdir original_block_ply
mv data_release/train/* original_block_ply
mv data_release/test/* original_block_ply
mv data_release/grid* ./

The data should organized in the following format:

/Dataset/SensatUrban/
          └── original_block_ply/
                  ├── birmingham_block_0.ply
                  ├── birmingham_block_1.ply 
		  ...
	    	  └── cambridge_block_34.ply 
          └── grid_0.200/
	     	  ├── birmingham_block_0_KDTree.pkl
                  ├── birmingham_block_0.ply
		  ├── birmingham_block_0_proj.pkl 
		  ...
	    	  └── cambridge_block_34.ply 

4、训练

• Start training: (Please first modified the root_path)

python main_SensatUrban.py --mode train --gpu 0 

🔥 训练前，记得修改class SensatUrban的root_path参数，ubuntu下本人地址为'/home/username/桌面/Randlanet/Dataset'
🔥 使用cpu则修改参数为“--gpu -1”
🔥 本机上tool.py中注释了cpu和gpu不同训练超参数。
🔥 batchsize过小，结果显著较差，miou基本维持在10%左右。

5、测试

• Evaluation:

python main_SensatUrban.py --mode test --gpu 0 

• Submit the results to the server: The compressed results can be found in /test/Log_*/test_preds/submission.zip. Then, feel free to submit this results to the evaluation server.

6、可视化

增加了可视化部分的代码，只需要命令行运行：

python visualize2ply.py

🔥 可视化处理前请自行需要根据需求修改文件地址。
🔥 可视化结果可以通过cloudcompare读入ply文件，scalar选择class一项。

Citation

If you find our work useful in your research, please consider citing:

@inproceedings{hu2020towards,
  title={Towards Semantic Segmentation of Urban-Scale 3D Point Clouds: A Dataset, Benchmarks and Challenges},
  author={Hu, Qingyong and Yang, Bo and Khalid, Sheikh and Xiao, Wen and Trigoni, Niki and Markham, Andrew},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year={2021}
}

Updates

• 01/03/2021: The SensatUrban has been accepted by CVPR 2021!
• 11/02/2021: The dataset is available for download!
• 07/09/2020: Initial release!

Related Repos

1. RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds
2. SoTA-Point-Cloud: Deep Learning for 3D Point Clouds: A Survey
3. 3D-BoNet: Learning Object Bounding Boxes for 3D Instance Segmentation on Point Clouds
4. SpinNet: Learning a General Surface Descriptor for 3D Point Cloud Registration
5. SQN: Weakly-Supervised Semantic Segmentation of Large-Scale 3D Point Clouds with 1000x Fewer Labels