Noising the point cloud.

The package is tested in Ubuntu 16.04, ROS kinetic 1.12.14, Python 3.6.

Requirements:

numpy 1.17.2

scikit-learn 0.23.1

tensorflow 1.14.0

keras 2.2.4

Anaconda3 is recommended. With Anaconda3, only tensorflow and keras need to be installed.

To make ROS and Anaconda3 compatible, in a new terminal:

gedit ~/.bashrc

add: source /opt/ros/kinetic/setup.bash

delete: export PATH="/home/tyang/anaconda3/bin:$PATH"

source ~/.bashrc

before launch the package:

export PATH="/home/tyang/anaconda3/bin:$PATH"

example:

download the lanoising package and decompress in ./src of your catkin workspace (e.g. catkin_ws).

in a new terminal:

cd ./catkin_ws

catkin_make

download the models and put all the files in ./catkin_ws/src/lanoising/models:

https://drive.google.com/file/d/1CoVrr3dVQ5DY4WpF7xCM9z6Vx7PYKW1w/view?usp=sharing

or: https://pan.baidu.com/s/1ZFhiuWFYNuSCThR02bLO8A with the code: ptio

in the terminal:

roscore

in a new terminal:

rviz

play the reference rosbag (point clouds recorded by velodyne LiDAR under clear weather conditions):

rosbag play -l --clock 2019-02-19-17-13-37.bag

in rviz, change the Fixed frame to "velodyne".

add the topic "/velodyne_points" in rviz to show the reference data.

set the visibility in lanoising.py.

in a new terminal:

cd ./catkin_ws

source devel/setup.bash

export PATH="/home/tyang/anaconda3/bin:$PATH"

roslaunch lanoising lanoising.launch

add the topic "/filtered_points" in rviz to show the noising point cloud.

If you publish work based on, or using, this code, we would appreciate citations to the following:

@artical{yt20iros,
    author       = {Tao Yang, You Li, Yassine Ruichek, and Zhi Yan}},
    title        = {LaNoising: A Data-driven Approach for 903nm ToF LiDAR Performance Modeling under Fog},
    booktitle    = {In Proceedings of the 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
    month        = {October},
    year         = {2020},
    }