This repository contains simple usage explanations of how the RangeNet++ inference works with the TensorRT and C++ interface.

Developed by Xieyuanli Chen, Andres Milioto and Jens Behley.

For more details about RangeNet++, one could find in LiDAR-Bonnetal.

For more details about rangenet_lib, one could find in rangenet_lib.

Add dynamic point cloud culling module, which can be used in low frame rate point cloud (2-3Hz) .

Remove all pseudo-dynamic point clouds, including static cars and people.

cd catkin_ws
source devel/setup.bash
cd src/semantic_slam

./run_taozi.sh 

cd catkin_ws
source devel/setup.bash
cd src/semantic_slam

./run_kitti_mapping.sh  or   ./run_taozi_mapping.sh

rosbag play taozihu_1.bag --pause -r0.5

The difference from the first is that the real dynamic point cloud is removed, excluding static cars and people. For specific operations, please read "aloam_pose_velodyne_32_mapping.launch" files， and Follow the prompts to modify

The dynamic point cloud removal of the KITTI dataset has not been tested for the time being, and will be added later.

cd catkin_ws
source devel/setup.bash
cd src/semantic_slam

./run_taozi_mapping.sh

rosbag play taozihu_1.bag --pause -r0.2

For specific operations, please read "run_kitti_mapping.sh" files， and Follow the prompts to modify

Use the tum format, pay attention to modify the position of the pose file in the code

source activate pytorch

#roslaunch semantic_slam aloam_pose_velodyne_64_mapping.launch
roslaunch semantic_slam aloam_pose_velodyne_64_mapping_loop.launch

First you need to install the nvidia driver and CUDA.

• CUDA Installation guide: Link

• Then you can do the other dependencies:

  $ sudo apt-get update 
  $ sudo apt-get install -yqq  build-essential python3-dev python3-pip apt-utils git cmake libboost-all-dev libyaml-cpp-dev libopencv-dev

• Then install the Python packages needed:

  conda create -n tensorflow python=3.6
  source activate tensorflow
  conda install tensorflow-gpu==1.11.0
  
  sudo apt install python-empy
  sudo pip install catkin_tools trollius numpy

In order to infer with TensorRT during inference with the C++ libraries:

run version : Ubuntu16.04 + CUDA10.0 + Cudnn7.5.0 + TensorRT-5.1.5.0

• Install TensorRT: Link.
• Our code and the pretrained model now only works with TensorRT version 5 (Note that you need at least version 5.1.0).
• To make the code also works for higher versions of TensorRT, one could have a look at here.

We use the catkin tool to build the library.

mkdir -p ~/catkin_wind/src
cd ~/catkin_wind/src
catkin_init_workspace

git clone https://github.com/GuoFeng-X/semantic_slam.git
cd .. 
catkin_make -j8

仅对A-LOAM中的两个launch文件进行改动，增加了激光雷达的话题参数。具体可参考A-LOAM. 使用mapping后的位姿作为最终位姿。

cd ~/catkin_wind/src
git clone https://github.com/GuoFeng-X/A-LOAM.git
cd .. 
catkin_make -j8

一段kitti数据集运行视频可查看哔哩哔哩，semantic_slam

To run the demo, you need a pre-trained model, which can be downloaded here, model.

将下载好的模型文件放到model中，所有模型文件放到darknet53_64和darknet53_32中。

# go to the root path of the catkin workspace
cd catkin_wind/src/rangenet_lib
source activate tensorflow
roslaunch semantic_slam aloam_pose_velodyne_64.launch
rosbag play -r0.5 kitti_odometry_sequence_00.bag

To run the demo, you need a pre-trained model, which can be downloaded here.

链接: https://pan.baidu.com/s/1EG27yE0Q_1gi3UD_Ta_vZw 提取码: riqc

仅仅改变了其中的 arch_cfg.yaml文件, 注意需要将两者的文件夹名字更换. 将下载的 arch_cfg.yaml 文件copy到darknet53_32中，替换掉原来的文件。

# go to the root path of the catkin workspace
cd catkin_wind/src/rangenet_lib
source activate pytorch
roslaunch semantic_slam aloam_pose_velodyne_32.launch
rosbag play -r0.5 lidar_cam_imu55.bag

For more details about how to train and evaluate a model, please refer to LiDAR-Bonnetal.

Using rangenet_lib, we built a LiDAR-based Semantic SLAM system, called SuMa++.

You could find more implementation details in SuMa++.

OverlapNet is a LiDAR-based loop closure detection method, which uses multiple cues generated from LiDAR scans.

More information about our OverlapNet could be found here.

One could use our rangenet_lib to generate probabilities over semantic classes for training OverlapNet.

More detailed steps and discussion could be found here.

If you use this library for any academic work, please cite the original paper.

@inproceedings{milioto2019iros,
  author    = {A. Milioto and I. Vizzo and J. Behley and C. Stachniss},
  title     = {{RangeNet++: Fast and Accurate LiDAR Semantic Segmentation}},
  booktitle = {IEEE/RSJ Intl.~Conf.~on Intelligent Robots and Systems (IROS)},
  year      = 2019,
  codeurl   = {https://github.com/PRBonn/lidar-bonnetal},
  videourl  = {https://youtu.be/wuokg7MFZyU},
}

If you use SuMa++, please cite the corresponding paper:

@inproceedings{chen2019iros, 
  author    = {X. Chen and A. Milioto and E. Palazzolo and P. Giguère and J. Behley and C. Stachniss},
  title     = {{SuMa++: Efficient LiDAR-based Semantic SLAM}},
  booktitle = {Proceedings of the IEEE/RSJ Int. Conf. on Intelligent Robots and Systems (IROS)},
  year      = {2019},
  codeurl   = {https://github.com/PRBonn/semantic_suma/},
  videourl  = {https://youtu.be/uo3ZuLuFAzk},
}

Copyright 2019, Xieyuanli Chen, Andres Milioto, Jens Behley, Cyrill Stachniss, University of Bonn.

This project is free software made available under the MIT License. For details see the LICENSE file.