T-LOAM: Truncated Least Squares Lidar-only Odometry and Mapping in Real-Time

Overview

T-LOAM: Truncated Least Squares Lidar-only Odometry and Mapping in Real-Time

The first Lidar-only odometry framework with high performance based on truncated least squares and Open3D point cloud library, The foremost improvement include:

  • Fast and precision pretreatment module, multi-region ground extraction and dynamic curved-voxel clustering perform ground point extraction and category segmentation.
  • Feature extraction based on principal component analysis(PCA) elaborate four distinctive feature,including: planar features, ground features, edge features, sphere features
  • There are three kinds of residual functions based on truncated least squares method for directly processing above features which are point-to-point, point-to-line, and point-to-plane.
  • Open3d point cloud library is integrated into SLAM algorithm framework for the first time. We extend more functions and implemented the message interface related to ROS.

[Demo Video] [Preprint Paper]

drawing

drawing drawing drawing drawing

Note that regard to pure odometry without corrections through loop closures, T-LOAM delivers much less drift than F-LOAM.

Framework overview

drawing

Each frame of the 3D LiDAR is processed as input. Four main processing modules are introduced to construct the backbone of the algorithm: (a) multi-region ground extraction module, (b) dynamic curved-voxel clustering module, (c) feature extraction module, (d) pose optimization module.

Evaluation

KITTI Sequence 00 F-LOAM T-LOAM
Translational Error(%) 1.11 0.98
Relative Error(°/100m) 0.40 0.60

Graphic Result(Path and Translation)

F-LOAM

drawing

T-LOAM

drawing

F-LOAM

drawing

T-LOAM

drawing

Dependency

-ROS(Melodic Ubuntu18.04)

sudo apt-get install python-catkin-tools ros-melodic-ecl-threads ros-melodic-jsk-recognition-msgs ros-melodic-jsk-visualization ros-melodic-velodyne-msgs

-YAML(0.6.3) Note that you must build a shared library due to we utilize the ros nodelet package.

tar -zxvf yaml-cpp-yaml-cpp-0.6.3.tar.gz
cd yaml-2.3.0 && mkdir build && cd build
cmake [-G generator] [-DYAML_BUILD_SHARED_LIBS=ON] ..
make 
sudo make install

-Open3D(A Modern Library for 3D Data Processing 0.12.0)

Please note that open3d installation will be a slightly troublesome process, please be patient. Another problem that needs attention is that Open3D-ML cannot be used in ROS at the same time due to the link error2286 and error3432. In order to fix this, you need to specify the cmake flag -DGLIBCXX_USE_CXX11_ABI=ON. However, the latest Tensorflow2.4 installed through conda(not pip) already supports the C++11 API, you can check the API with print(tensorflow.__cxx11_abi_flag__). If the flag is true, you can set the compile flag -DBUILD_TENSORFLOW_OPS=ON Next, you can complete the installation according to the instructions

cd Open3D
util/scripts/install-deps-ubuntu.sh
mkdir build && cd build 
cmake \
    -DBUILD_SHARED_LIBS=ON \
    -DPYTHON_EXECUTABLE=$(which python3) \
    -DBUILD_CUDA_MODULE=ON \
    -DGLIBCXX_USE_CXX11_ABI=ON \
    -DBUILD_LIBREALSENSE=ON  \
    -DCMAKE_BUILD_TYPE=Release \
    -DCMAKE_INSTALL_PREFIX=/usr/local \
    -DBUILD_PYTORCH_OPS=OFF \
    -DBUILD_TENSORFLOW_OPS=OFF \
    -DBUNDLE_OPEN3D_ML=ON \
    -DOPEN3D_ML_ROOT=${replace with own Open3D-ML path} \
    ../
make -j4
sudo make install 

If you have clone problems, you can download it directly from the link below.

Baidu Disk code: khy9 or Google Drive

-Ceres Solver(A large scale non-linear optimization library 2.0) you can complete the installation according to the guide

Installation

Now create the Catkin Environment:

mkdir -p ~/tloam_ws/src
cd ~/tloam_ws
catkin init
catkin config --merge-devel
catkin config --cmake-args -DCMAKE_BUILD_TYPE=Release

And clone the project:

cd src
git clone https://github.com/zpw6106/tloam.git
catkin build

Usage

Download the KITTI Odometry Dataset (Graviti can provide faster download speed in China), then organize it according to the following structure, and modify the read path in the config/kitti/kitti_reader.yaml

drawing

-Example for running T-LOAM using the KITTI Dataset

roslaunch tloam tloam_kitti.launch

Contributors

Pengwei Zhou (Email: [email protected])

BibTex Citation

Thank you for citing our T-LOAM paper on IEEEif you use any of this code:

@ARTICLE{9446309,
  author={Zhou, Pengwei and Guo, Xuexun and Pei, Xiaofei and Chen, Ci},
  journal={IEEE Transactions on Geoscience and Remote Sensing}, 
  title={T-LOAM: Truncated Least Squares LiDAR-Only Odometry and Mapping in Real Time}, 
  year={2021},
  volume={},
  number={},
  pages={1-13},
  doi={10.1109/TGRS.2021.3083606}
  }

Credits

We hereby recommend reading A-LOAM ,floam and TEASER for reference and thank them for making their work public.

License

The source code is released under GPLv3 license.

I am constantly working on improving this code. For any technical issues or commercial use, please contact me([email protected]).

Owner
Pengwei Zhou
Lidar SLAM & Sensor Fusion
Pengwei Zhou
UMPNet: Universal Manipulation Policy Network for Articulated Objects

UMPNet: Universal Manipulation Policy Network for Articulated Objects Zhenjia Xu, Zhanpeng He, Shuran Song Columbia University Robotics and Automation

Columbia Artificial Intelligence and Robotics Lab 33 Dec 03, 2022
Citation Intent Classification in scientific papers using the Scicite dataset an Pytorch

Citation Intent Classification Table of Contents About the Project Built With Installation Usage Acknowledgments About The Project Citation Intent Cla

Federico Nocentini 4 Mar 04, 2022
Finite-temperature variational Monte Carlo calculation of uniform electron gas using neural canonical transformation.

CoulombGas This code implements the neural canonical transformation approach to the thermodynamic properties of uniform electron gas. Building on JAX,

FermiFlow 9 Mar 03, 2022
DeepLM: Large-scale Nonlinear Least Squares on Deep Learning Frameworks using Stochastic Domain Decomposition (CVPR 2021)

DeepLM DeepLM: Large-scale Nonlinear Least Squares on Deep Learning Frameworks using Stochastic Domain Decomposition (CVPR 2021) Run Please install th

Jingwei Huang 130 Dec 02, 2022
On-device speech-to-intent engine powered by deep learning

Rhino Made in Vancouver, Canada by Picovoice Rhino is Picovoice's Speech-to-Intent engine. It directly infers intent from spoken commands within a giv

Picovoice 510 Dec 30, 2022
Fast and robust certifiable relative pose estimation

Fast and Robust Relative Pose Estimation for Calibrated Cameras This repository contains the code for the relative pose estimation between two central

42 Dec 06, 2022
Code for training and evaluation of the model from "Language Generation with Recurrent Generative Adversarial Networks without Pre-training"

Language Generation with Recurrent Generative Adversarial Networks without Pre-training Code for training and evaluation of the model from "Language G

Amir Bar 253 Sep 14, 2022
NeoPlay is the project dedicated to ESport events.

NeoPlay is the project dedicated to ESport events. On this platform users can participate in tournaments with prize pools as well as create their own tournaments.

3 Dec 18, 2021
How to Train a GAN? Tips and tricks to make GANs work

(this list is no longer maintained, and I am not sure how relevant it is in 2020) How to Train a GAN? Tips and tricks to make GANs work While research

Soumith Chintala 10.8k Dec 31, 2022
Code for ICCV2021 paper PARE: Part Attention Regressor for 3D Human Body Estimation

PARE: Part Attention Regressor for 3D Human Body Estimation [ICCV 2021] PARE: Part Attention Regressor for 3D Human Body Estimation, Muhammed Kocabas,

Muhammed Kocabas 277 Jan 03, 2023
ICLR2021 (Under Review)

Self-Supervised Time Series Representation Learning by Inter-Intra Relational Reasoning This repository contains the official PyTorch implementation o

Haoyi Fan 58 Dec 30, 2022
Official implementation of our paper "Learning to Bootstrap for Combating Label Noise"

Learning to Bootstrap for Combating Label Noise This repo is the official implementation of our paper "Learning to Bootstrap for Combating Label Noise

21 Apr 09, 2022
Mesh Graphormer is a new transformer-based method for human pose and mesh reconsruction from an input image

MeshGraphormer ✨ ✨ This is our research code of Mesh Graphormer. Mesh Graphormer is a new transformer-based method for human pose and mesh reconsructi

Microsoft 251 Jan 08, 2023
PyTorch implementation of neural style randomization for data augmentation

README Augment training images for deep neural networks by randomizing their visual style, as described in our paper: https://arxiv.org/abs/1809.05375

84 Nov 23, 2022
A Real-Time-Strategy game for Deep Learning research

Description DeepRTS is a high-performance Real-TIme strategy game for Reinforcement Learning research. It is written in C++ for performance, but provi

Centre for Artificial Intelligence Research (CAIR) 156 Dec 19, 2022
Seeing All the Angles: Learning Multiview Manipulation Policies for Contact-Rich Tasks from Demonstrations

Seeing All the Angles: Learning Multiview Manipulation Policies for Contact-Rich Tasks from Demonstrations Trevor Ablett, Daniel (Yifan) Zhai, Jonatha

STARS Laboratory 3 Feb 01, 2022
A DeepStack custom model for detecting common objects in dark/night images and videos.

DeepStack_ExDark This repository provides a custom DeepStack model that has been trained and can be used for creating a new object detection API for d

MOSES OLAFENWA 98 Dec 24, 2022
Adaptive Denoising Training (ADT) for Recommendation.

DenoisingRec Adaptive Denoising Training for Recommendation. This is the pytorch implementation of our paper at WSDM 2021: Denoising Implicit Feedback

Wenjie Wang 51 Dec 30, 2022
A PyTorch Reimplementation of TecoGAN: Temporally Coherent GAN for Video Super-Resolution

TecoGAN-PyTorch Introduction This is a PyTorch reimplementation of TecoGAN: Temporally Coherent GAN for Video Super-Resolution (VSR). Please refer to

165 Dec 17, 2022
Pytorch implementation of Learning Rate Dropout.

Learning-Rate-Dropout Pytorch implementation of Learning Rate Dropout. Paper Link: https://arxiv.org/pdf/1912.00144.pdf Train ResNet-34 for Cifar10: r

42 Nov 25, 2022