Joint Scene Flow Estimation and Moving Object Segmentation on Rotational LiDAR Data

LiDAR-based scene flow estimation (SFE) and moving object segmentation (MOS) are important tasks with broad-ranging applications in autonomous driving, such as traffic surveillance, motion analysis, obstacle avoidance, etc. Most existing works address SFE and MOS separately, ignoring the underlying shared geometric constraints and their inherent correlation. This article rethinks LiDAR-based SFE and MOS tasks, providing our key insight that jointly addressing them can tackle challenges in both tasks, and their solutions can reinforce one another to improve the performance of both. Based on this insight, we introduce a novel framework that exploits shared geometric constraints by explicitly partitioning the scene into static and moving regions and subsequently estimating flow differently for these regions. A lightweight and interpretable neural network dubbed SFEMOS is proposed. It employs an encoder and two specially designed head modules for each task, achieving MOS without relying on prior poses and online point-wise flow estimation for 360-degree point clouds. Due to the absence of public datasets for concurrently evaluating both tasks, we generate ground truth flow data using MOS labels from SemanticKITTI. Additionally, we establish a new dataset using a rotational LiDAR mounted on our own autonomous vehicle. Evaluation results on both datasets validate the superior performance of our proposed SFEMOS. Our dataset and label generation method are released at https://github.com/nubot-nudt/SFEMOS.

• Publication year

  2024

• Venue

  IEEE transactions on intelligent transportation systems (Print)

• Publication date

  2024-11-01

• Fields of study

  Computer Science, Engineering, Environmental Science

• Identifiers
  DOI 10.1109/TITS.2024.3432755
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• MarS3D: A Plug-and-Play Motion-Aware Model for Semantic Segmentation on Multi-Scan 3D Point Clouds

  2023cited by this paper
• DifFlow3D: Toward Robust Uncertainty-Aware Scene Flow Estimation with Diffusion Model

  2023cited by this paper
• InsMOS: Instance-Aware Moving Object Segmentation in LiDAR Data

  2023cited by this paper
• PT-FlowNet: Scene Flow Estimation on Point Clouds With Point Transformer

  2023cited by this paper
• RDMNet: Reliable Dense Matching Based Point Cloud Registration for Autonomous Driving

  2023influential reference
• Building Volumetric Beliefs for Dynamic Environments Exploiting Map-Based Moving Object Segmentation

  2023cited by this paper
• Receding Moving Object Segmentation in 3D LiDAR Data Using Sparse 4D Convolutions

  2022cited by this paper
• Automatic Labeling to Generate Training Data for Online LiDAR-Based Moving Object Segmentation

  2022cited by this paper
• Abstract Flow for Temporal Semantic Segmentation on the Permutohedral Lattice

  2022cited by this paper
• Patchwork++: Fast and Robust Ground Segmentation Solving Partial Under-Segmentation Using 3D Point Cloud

  2022cited by this paper
• RigidFlow: Self-Supervised Scene Flow Learning on Point Clouds by Local Rigidity Prior

  2022cited by this paper
• Efficient Spatial-Temporal Information Fusion for LiDAR-Based 3D Moving Object Segmentation

  2022cited by this paper
• RVMOS: Range-View Moving Object Segmentation leveraged by Semantic and Motion Features

  2022cited by this paper
• Unsupervised 4D LiDAR Moving Object Segmentation in Stationary Settings with Multivariate Occupancy Time Series

  2022cited by this paper
• SCOOP: Self-Supervised Correspondence and Optimization-Based Scene Flow

  2022cited by this paper
• FH-Net: A Fast Hierarchical Network for Scene Flow Estimation on Real-World Point Clouds

  2022cited by this paper
• SLIM: Self-Supervised LiDAR Scene Flow and Motion Segmentation

  2021cited by this paper
• LiMoSeg: Real-time Bird's Eye View based LiDAR Motion Segmentation

  2021cited by this paper
• Moving Object Segmentation in 3D LiDAR Data: A Learning-Based Approach Exploiting Sequential Data

  2021influential reference
• Self-Point-Flow: Self-Supervised Scene Flow Estimation from Point Clouds with Optimal Transport and Random Walk

  2021cited by this paper
• SCTN: Sparse Convolution-Transformer Network for Scene Flow Estimation

  2021cited by this paper
• Moving Object Detection by 3D Flow Field Analysis

  2021cited by this paper
• PointMoSeg: Sparse Tensor-Based End-to-End Moving-Obstacle Segmentation in 3-D Lidar Point Clouds for Autonomous Driving

  2021influential reference
• Weakly Supervised Learning of Rigid 3D Scene Flow

  2021cited by this paper
• PSE-Match: A Viewpoint-Free Place Recognition Method With Parallel Semantic Embedding

  2021influential reference
• LiDAR-based Recurrent 3D Semantic Segmentation with Temporal Memory Alignment

  2020cited by this paper
• 3D LiDAR-Based Global Localization Using Siamese Neural Network

  2020influential reference
• 3D LiDAR Map Compression for Efficient Localization on Resource Constrained Vehicles

  2020influential reference
• PolarNet: An Improved Grid Representation for Online LiDAR Point Clouds Semantic Segmentation

  2020cited by this paper
• SpSequenceNet: Semantic Segmentation Network on 4D Point Clouds

  2020cited by this paper
• Pseudo-LiDAR Point Cloud Interpolation Based on 3D Motion Representation and Spatial Supervision

  2020cited by this paper
• SalsaNext: Fast, Uncertainty-Aware Semantic Segmentation of LiDAR Point Clouds

  2020cited by this paper
• FLOT: Scene Flow on Point Clouds Guided by Optimal Transport

  2020influential reference
• Multi-Scale Interaction for Real-Time LiDAR Data Segmentation on an Embedded Platform

  2020cited by this paper
• PointPWC-Net: Cost Volume on Point Clouds for (Self-)Supervised Scene Flow Estimation

  2020cited by this paper
• Cylindrical and Asymmetrical 3D Convolution Networks for LiDAR Segmentation

  2020cited by this paper
• SuMa++: Efficient LiDAR-based Semantic SLAM

  2019cited by this paper
• RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds

  2019cited by this paper
• SemanticKITTI: A Dataset for Semantic Scene Understanding of LiDAR Sequences

  2019influential reference
• Motion Guided Attention for Video Salient Object Detection

  2019cited by this paper
• FlowNet3D: Learning Scene Flow in 3D Point Clouds

  2018cited by this paper
• Object scene flow

  2017cited by this paper
• Comprehensive Data Set for Automatic Single Camera Visual Speed Measurement

  2017cited by this paper
• SGDR: Stochastic Gradient Descent with Warm Restarts

  2016influential reference
• PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation

  2016cited by this paper
• A Large Dataset to Train Convolutional Networks for Disparity, Optical Flow, and Scene Flow Estimation

  2015cited by this paper
• Sinkhorn Distances: Lightspeed Computation of Optimal Transport

  2013cited by this paper
• A Method for Registration of 3-D Shapes

  1992cited by this paper
• The Hungarian method for the assignment problem

  1955cited by this paper
• Ieee Transactions on Intelligent Transportation Systems 1 Autonomous Visual Navigation and Laser-based Moving Obstacle Avoidance

  year unknowninfluential reference

• Toward network clustering using graph-boosting and graph-factorization

  2026cites this paper
• LighterBEV: LiDAR Global Localization Meets Online Learning

  2026cites this paper
• Precise path planning based on improved A* and DWA algorithms for complex agricultural environments

  2025cites this paper
• A novel deep neural architecture for efficient and scalable multidomain image classification

  2025cites this paper
• Computer vision-based laser communication system for robust optical beam tracking and alignment

  2025cites this paper
• Detection of DDoS attacks in IoT networks using a majority voting mechanism

  2025cites this paper
• Proximal Policy Optimization for Vehicular Big Data Offloading Across Edge, Regional, and Cloud Layers

  2025cites this paper
• Design of Data Driven Automated Driving Control Algorithm for Enhanced Human Acceptance

  2025cites this paper
• A Novel Security Defense and Economic Assessment Algorithm for mmWave‐Vehicular Network Based on Deep Reinforcement Learning

  2025cites this paper
• Lightweight Surface Moving Object Detection with Noise Suppression Based on Lidar

  2025cites this paper
• STC-Tracker: Spatiotemporal-Consistent Multi-Robot Collaboration Framework for Long-Term Dynamic Object Tracking

  2025cites this paper
• Real-Time Long-Range Object Tracking Based on Ensembled Model

  2025cites this paper
• Transportation mode detection through spatial attention-based transductive long short-term memory and off-policy feature selection

  2025cites this paper
• A Novel Remote Sensing Recognition Using Modified GMM Segmentation and DenseNet

  2025cites this paper
• Environment-Driven Online LiDAR-Camera Extrinsic Calibration

  2025cites this paper
• Mitigation of Carbon Footprint in Hybrid Vehicle-Dominated Distribution Systems via P2G Technology and Virtual Assets

  2025cites this paper
• A multi objective optimization framework for smart parking using digital twin pareto front MDP and PSO for smart cities

  2025cites this paper
• Intelligent optimization method for hazardous materials transportation routing with multi-mode and multi-criterion collaborative constraints

  2025cites this paper
• SemanticFlow: A Self-Supervised Framework for Joint Scene Flow Prediction and Instance Segmentation in Dynamic Environments

  2025cites this paper
• BiAttentionNet: a dual-branch automatic driving image segmentation network integrating spatial and channel attention mechanisms

  2025cites this paper
• Self-Supervised Diffusion-Based Scene Flow Estimation and Motion Segmentation With 4D Radar

  2025cites this paper
• Alternating interaction fusion of Image-Point cloud for Multi-Modal 3D object detection

  2025cites this paper
• Multiple target detection using photonic radar for autonomous vehicles under atmospheric rain conditions

  2025cites this paper
• Synergistic integration of refined pelican optimization algorithm and deep neural networks for autonomous vehicle control in edge computing architectures

  2025cites this paper
• Deep fusion for enhancing Internet of Vehicles security in sustainable smart vehicular communication with LSTM and GRU-guided artificial deep neural networks

  2025cites this paper
• Integrating cyber-physical systems with embedding technology for controlling autonomous vehicle driving

  2025cites this paper
• ELMAR: Enhancing LiDAR Detection with 4D Radar Motion Awareness and Cross-modal Uncertainty

  2025cites this paper
• Single-Frame Point-Pixel Registration via Supervised Cross-Modal Feature Matching

  2025cites this paper
• IoT-Enhanced Smart Parking Management With IncepDenseMobileNet for Improved Classification

  2025cites this paper
• Spinning dynamics of stress-dependent viscosity of generalized Cross-nonlinear materials affected by gravitationally swirling disk

  2025cites this paper
• WIN: Variable-View Implicit LiDAR Upsampling Network

  2025cites this paper
• Multi-Target Coverage Trajectory Planning for Ceiling Painting Robot Chassis via Two-Stage Optimization

  2025cites this paper
• HSFANet: Hierarchical Scale-Sensitive Feature Aggregation Network for Small Object Detection in UAV Aerial Images

  2025cites this paper
• Machine learning-based prediction analysis for Australia sustainable transport: Optimal transportation mix equation for long-term policy planning and environmental impact mitigation

  2025cites this paper
• iDigiChat: intelligent digital marketing service chatbot for providing efficient customer services using artificial intelligence

  2025cites this paper
• CNTs-coated stretchable fabric as ultrasensitive e-skin for rotational motion monitoring in humanoid robots

  2024cites this paper
• I-rod: an ensemble of CNNs for object detection in unconstrained road scenarios

  2024cites this paper
• Solving Maxwell Equations in Hybrid Electrical Vehicle to Optimal Design by Multi-Objective Optimization Algorithms

  2024cites this paper
• BEVPlace++: Fast, Robust, and Lightweight LiDAR Global Localization for Unmanned Ground Vehicles

  2024cites this paper
• Advanced Human Pose Estimation and Event Classification Using Context-Aware Features and XGBoost Classifier

  2024cites this paper
• Fine-tuned depth-augmented U-Net for enhanced semantic segmentation in indoor autonomous vision systems

  2024cites this paper
• LiDAR-BEVMTN: Real-Time LiDAR Bird’s-Eye View Multi-Task Perception Network for Autonomous Driving

  2023cites this paper
• The social reactive and/or reasoned acceptance of the intelligent speed adaptation system in Türkiye and Israel

  year unknowncites this paper