Latest 10 Papers - March 10, 2025

by ADMIN 34 views

Visual SLAM

Visual SLAM (Simultaneous Localization and Mapping) is a technique used in robotics and computer vision to enable a robot or a camera to build a map of its environment while simultaneously localizing itself within that map. Here are some recent papers on Visual SLAM:

1. OpenGV 2.0: Motion prior-assisted calibration and SLAM with vehicle-mounted surround-view systems

  • Date: 2025-03-05
  • Comment: This paper presents a new version of the OpenGV library, which is a popular open-source library for computer vision and robotics. The new version includes a motion prior-assisted calibration and SLAM (Simultaneous Localization and Mapping) algorithm for vehicle-mounted surround-view systems.

2. Monocular visual simultaneous localization and mapping: (r)evolution from geometry to deep learning-based pipelines

  • Date: 2025-03-04
  • Comment: This paper provides a comprehensive review of the evolution of monocular visual SLAM from traditional geometry-based pipelines to deep learning-based pipelines. The authors discuss the advantages and limitations of each approach and provide a roadmap for future research.

3. vS-Graphs: Integrating Visual SLAM and Situational Graphs through Multi-level Scene Understanding

  • Date: 2025-03-03
  • Comment: This paper presents a new approach to integrating visual SLAM and situational graphs through multi-level scene understanding. The authors propose a new framework that combines visual SLAM and situational graphs to enable more accurate and robust scene understanding.

4. MUSt3R: Multi-view Network for Stereo 3D Reconstruction

  • Date: 2025-03-03
  • Comment: This paper presents a new multi-view network for stereo 3D reconstruction. The authors propose a novel architecture that combines multiple views of the scene to enable more accurate and robust 3D reconstruction.

5. HMD^2: Environment-aware Motion Generation from Single Egocentric Head-Mounted Device

  • Date: 2025-03-02
  • Comment: This paper presents a new approach to environment-aware motion generation from a single egocentric head-mounted device. The authors propose a novel algorithm that takes into account the environment and the user's motion to generate more realistic and natural motion.

6. Action-Consistent Decentralized Belief Space Planning with Inconsistent Beliefs and Limited Data Sharing: Framework and Simplification Algorithms with Formal Guarantees

  • Date: 2025-03-02
  • Comment: This paper presents a new framework for action-consistent decentralized belief space planning with inconsistent beliefs and limited data sharing. The authors propose a novel algorithm that provides formal guarantees for the correctness and optimality of the plan.

7. AirSLAM: An Efficient and Illumination-Robust Point-Line Visual SLAM System

  • Date: 2025-02-27
  • Comment: This paper presents a new efficient and illumination-robust point-line visual SLAM system. The authors propose a novel algorithm that combines point-line features and illumination-invariant features to enable more accurate and robust SLAM.

8. Increasing the Task Flexibility of Heavy-Duty Manipulators Using Visual 6D Pose Estimation of Objects

  • Date: 2025-02-26
  • Comment: This paper presents a new approach to increasing the task flexibility of heavy-duty manipulators using visual 6D pose estimation of objects. The authors propose a novel algorithm that enables more accurate and robust pose estimation of objects in 6D space.

9. SLAM in the Dark: Self-Supervised Learning of Pose, Depth and Loop-Closure from Thermal Images

  • Date: 2025-02-26
  • Comment: This paper presents a new approach to SLAM in the dark using self-supervised learning of pose, depth, and loop-closure from thermal images. The authors propose a novel algorithm that enables more accurate and robust SLAM in low-light environments.

10. GSORB-SLAM: Gaussian Splatting SLAM benefits from ORB features and Transmittance information

  • Date: 2025-02-22
  • Comment: This paper presents a new approach to Gaussian splatting SLAM that benefits from ORB features and transmittance information. The authors propose a novel algorithm that combines ORB features and transmittance information to enable more accurate and robust SLAM.

Visual Inertial SLAM

Visual Inertial SLAM (VINS) is a technique used in robotics and computer vision to enable a robot or a camera to build a map of its environment while simultaneously localizing itself within that map using both visual and inertial data. Here are some recent papers on Visual Inertial SLAM:

1. Efficient Submap-based Autonomous MAV Exploration using Visual-Inertial SLAM Configurable for LiDARs or Depth Cameras

  • Date: 2025-03-05
  • Comment: This paper presents a new approach to efficient submap-based autonomous MAV exploration using visual-inertial SLAM configurable for LiDARs or depth cameras. The authors propose a novel algorithm that enables more accurate and robust exploration of the environment.

2. MLINE-VINS: Robust Monocular Visual-Inertial SLAM With Flow Manhattan and Line Features

  • Date: 2025-03-03
  • Comment: This paper presents a new approach to robust monocular visual-inertial SLAM with flow Manhattan and line features. The authors propose a novel algorithm that combines flow Manhattan and line features to enable more accurate and robust SLAM.

3. RUSSO: Robust Underwater SLAM with Sonar Optimization against Visual Degradation

  • Date: 2025-03-03
  • Comment: This paper presents a new approach to robust underwater SLAM with sonar optimization against visual degradation. The authors propose a novel algorithm that combines sonar and visual data to enable more accurate and robust SLAM in underwater environments.

4. LVI-GS: Tightly-coupled LiDAR-Visual-Inertial SLAM using 3D Gaussian Splatting

  • Date: 2024-11-05
  • Comment: This paper presents a new approach to tightly-coupled LiDAR-visual-inertial SLAM using 3D Gaussian splatting. The authors propose a novel algorithm that combines LiDAR and visual data using 3D Gaussian splatting to enable more accurate and robust SLAM.

5. SuperVINS: A Real-Time Visual-Inertial SLAM Framework for Challenging Imaging Conditions

  • Date: 2024-11-03
  • Comment: This paper presents a new approach to real-time visual-inertial SLAM for challenging imaging conditions. The authors propose a novel algorithm that enables more accurate and robust SLAM in challenging imaging conditions.

6. Uncertainty-Aware Visual-Inertial SLAM with Volumetric Occupancy Mapping

  • Date: 2024-09-23
  • Comment: This paper presents a new approach to uncertainty-aware visual-inertial SLAM with volumetric occupancy mapping. The authors propose a novel algorithm that combines uncertainty-aware SLAM and volumetric occupancy mapping to enable more accurate and robust SLAM.

7. Visual-Inertial SLAM as Simple as A, B, VINS

  • Date: 2024-09-22
  • Comment: This paper presents a new approach to visual-inertial SLAM as simple as A, B, VINS. The authors propose a novel algorithm that simplifies the visual-inertial SLAM process to enable more accurate and robust SLAM.

8. Enhancing Visual Inertial SLAM with Magnetic Measurements

  • Date: 2024-09-16
  • Comment: This paper presents a new approach to enhancing visual-inertial SLAM with magnetic measurements. The authors propose a novel algorithm that combines visual-inertial SLAM and magnetic measurements to enable more accurate and robust SLAM.

9. Advancements in Translation Accuracy for Stereo Visual-Inertial Initialization

  • Date: 2024-08-18
  • Comment: This paper presents a new approach to advancements in translation accuracy for stereo visual-inertial initialization. The authors propose a novel algorithm that improves the translation accuracy for stereo visual-inertial initialization.

10. Visual-Inertial SLAM for Agricultural Robotics: Benchmarking the Benefits and Computational Costs of Loop Closing

  • Date: 2024-08-03
  • Comment: This paper presents a new approach to visual-inertial SLAM for agricultural robotics. The authors propose a novel algorithm that benchmarks the benefits and computational costs of loop closing for visual-inertial SLAM in agricultural robotics.

Visual Inertial Odometry

Visual Inertial Odometry (VIO) is a technique used in robotics and computer vision to enable a robot or a camera to estimate its motion and orientation using both visual and inertial data. Here are some recent papers on Visual Inertial Odometry:

1. MLINE-VINS: Robust Monocular Visual-Inertial SLAM With Flow Manhattan and Line Features

  • Date: 2025-03-03
  • Comment: This paper presents a new approach to robust monocular visual-inertial SLAM with flow Manhattan and line features. The authors propose a novel algorithm that combines flow Manhattan and line features to enable more accurate and robust VIO.

2. ADUGS-VINS: Generalized Visual-Inertial Odometry for Robust Navigation in Highly Dynamic and Complex Environments

  • Date: 2025-03-03
  • Comment: This paper presents a new approach to generalized visual-inertial odometry for robust navigation in highly
    Q&A: Latest 10 Papers - March 10, 2025

Q: What is Visual SLAM?

A: Visual SLAM (Simultaneous Localization and Mapping) is a technique used in robotics and computer vision to enable a robot or a camera to build a map of its environment while simultaneously localizing itself within that map.

Q: What are some recent papers on Visual SLAM?

A: Some recent papers on Visual SLAM include:

  • OpenGV 2.0: Motion prior-assisted calibration and SLAM with vehicle-mounted surround-view systems
  • Monocular visual simultaneous localization and mapping: (r)evolution from geometry to deep learning-based pipelines
  • vS-Graphs: Integrating Visual SLAM and Situational Graphs through Multi-level Scene Understanding
  • MUSt3R: Multi-view Network for Stereo 3D Reconstruction
  • HMD^2: Environment-aware Motion Generation from Single Egocentric Head-Mounted Device
  • Action-Consistent Decentralized Belief Space Planning with Inconsistent Beliefs and Limited Data Sharing: Framework and Simplification Algorithms with Formal Guarantees
  • AirSLAM: An Efficient and Illumination-Robust Point-Line Visual SLAM System
  • Increasing the Task Flexibility of Heavy-Duty Manipulators Using Visual 6D Pose Estimation of Objects
  • SLAM in the Dark: Self-Supervised Learning of Pose, Depth and Loop-Closure from Thermal Images
  • GSORB-SLAM: Gaussian Splatting SLAM benefits from ORB features and Transmittance information

Q: What is Visual Inertial SLAM?

A: Visual Inertial SLAM (VINS) is a technique used in robotics and computer vision to enable a robot or a camera to build a map of its environment while simultaneously localizing itself within that map using both visual and inertial data.

Q: What are some recent papers on Visual Inertial SLAM?

A: Some recent papers on Visual Inertial SLAM include:

  • Efficient Submap-based Autonomous MAV Exploration using Visual-Inertial SLAM Configurable for LiDARs or Depth Cameras
  • MLINE-VINS: Robust Monocular Visual-Inertial SLAM With Flow Manhattan and Line Features
  • RUSSO: Robust Underwater SLAM with Sonar Optimization against Visual Degradation
  • LVI-GS: Tightly-coupled LiDAR-Visual-Inertial SLAM using 3D Gaussian Splatting
  • SuperVINS: A Real-Time Visual-Inertial SLAM Framework for Challenging Imaging Conditions
  • Uncertainty-Aware Visual-Inertial SLAM with Volumetric Occupancy Mapping
  • Visual-Inertial SLAM as Simple as A, B, VINS
  • Enhancing Visual Inertial SLAM with Magnetic Measurements
  • Advancements in Translation Accuracy for Stereo Visual-Inertial Initialization
  • Visual-Inertial SLAM for Agricultural Robotics: Benchmarking the Benefits and Computational Costs of Loop Closing

Q: What is Visual Inertial Odometry?

A: Visual Inertial Odometry (VIO) is a technique used in robotics and computer vision to enable a robot or a camera to estimate its motion and orientation using both visual and inertial data.

Q: What are some recent papers on Visual Inertial Odometry?

A: Some recent papers on Visual Inertial Odometry include:

  • MLINE-VINS: Robust Monocular Visual-Inertial SLAM With Flow Manhattan and Line Features
  • ADUGS-VINS: Generalized Visual-Inertial Odometry for Robust Navigation in Highly Dynamic and Complex Environments
  • ESVO2: Direct Visual-Inertial Odometry with Stereo Event Cameras
  • XIRVIO: Critic-guided Iterative Refinement for Visual-Inertial Odometry with Explainable Adaptive Weighting
  • Improving Monocular Visual-Inertial Initialization with Structureless Visual-Inertial Bundle Adjustment
  • A Robust and Efficient Visual-Inertial Initialization with Probabilistic Normal Epipolar Constraint
  • HelmetPoser: A Helmet-Mounted IMU Dataset for Data-Driven Estimation of Human Head Motion in Diverse Conditions
  • DeepVL: Dynamics and Inertial Measurements-based Deep Velocity Learning for Underwater Odometry
  • A Transformation-based Consistent Estimation Framework: Analysis, Design and Applications
  • XR-VIO: High-precision Visual Inertial Odometry with Fast Initialization for XR Applications

Q: What is Lidar SLAM?

A: Lidar SLAM (Simultaneous Localization and Mapping) is a technique used in robotics and computer vision to enable a robot or a camera to build a map of its environment while simultaneously localizing itself within that map using Lidar data.

Q: What are some recent papers on Lidar SLAM?

A: Some recent papers on Lidar SLAM include:

  • Online Tree Reconstruction and Forest Inventory on a Mobile Robotic System
  • Balancing Accuracy and Efficiency for Large-Scale SLAM: A Minimal Subset Approach for Scalable Loop Closures
  • Anti-Degeneracy Scheme for Lidar SLAM based on Particle Filter in Geometry Feature-Less Environments
  • SiLVR: Scalable Lidar-Visual Radiance Field Reconstruction with Uncertainty Quantification
  • LiDAR Loop Closure Detection using Semantic Graphs with Graph Attention Networks
  • Lifelong 3D Mapping Framework for Hand-held & Robot-mounted LiDAR Mapping Systems
  • Unified Few-shot Crack Segmentation and its Precise 3D Automatic Measurement in Concrete Structures
  • Informed, Constrained, Aligned: A Field Analysis on Degeneracy-aware Point Cloud Registration in the Wild
  • ROLO-SLAM: Rotation-Optimized LiDAR-Only SLAM in Uneven Terrain with Ground Vehicle
  • A flexible framework for accurate LiDAR odometry, map manipulation, and localization

Q: What is Lidar Odometry?

A: Lidar Odometry is a technique used in robotics and computer vision to enable a robot or a camera to estimate its motion and orientation using Lidar data.

Q: What are some recent papers on Lidar Odometry?

A: Some recent papers on Lidar Odometry include:

  • Robustness of LiDAR-Based Pose Estimation: Evaluating and Improving Odometry and Localization Under Common Point Cloud Corruptions
  • CAO-RONet: A Robust 4D Radar Odometry with Exploring More Information from Low-Quality Points
  • CAR-LOAM: Color-Assisted Robust LiDAR Odometry and Mapping
  • Ground-Optimized 4D Radar-Inertial Odometry via Continuous Velocity Integration using Gaussian Process
  • Doppler Correspondence: Non-Iterative Scan Matching With Doppler Velocity-Based Correspondence
  • CTE-MLO: Continuous-time and Efficient Multi-LiDAR Odometry with Localizability-aware Point Cloud Sampling
  • Kinematic-ICP: Enhancing LiDAR Odometry with Kinematic Constraints for Wheeled Mobile Robots Moving on Planar Surfaces
  • DOC-Depth: A novel approach for dense depth ground truth generation
  • OpenLiDARMap: Zero-Drift Point Cloud Mapping using Map Priors
  • Performance Assessment of Lidar Odometry Frameworks: A Case Study at the Australian Botanic Garden Mount Annan

Q: What is SLAMMOT?

A: SLAMMOT (Simultaneous Localization and Mapping with Multiple Objects and Tasks) is a technique used in robotics and computer vision to enable a robot or a camera to build a map of its environment while simultaneously localizing itself within that map and performing multiple tasks.

Q: What are some recent papers on SLAMMOT?

A: Some recent papers on SLAMMOT include:

  • LiDAR SLAMMOT based on Confidence-guided Data Association
  • Visual SLAMMOT Considering Multiple Motion Models
  • Communication-Efficient Cooperative SLAMMOT via Determining the Number of Collaboration Vehicles

Q: What is GNSS?

A: GNSS (Global Navigation Satellite System) is a technique used in robotics and computer vision to enable a robot or a camera to determine its location and orientation using satellite signals.

Q: What are some recent papers on GNSS?

A: Some recent papers on GNSS include:

  • Evaluating ML Robustness in GNSS Interference Classification, Characterization & Localization
  • GS-GVINS: A Tightly-integrated GNSS-Visual-Inertial Navigation System Augmented by 3D Gaussian Splatting
  • Time-based GNSS attack detection
  • Open-Source Factor Graph Optimization Package for GNSS: Examples and Applications
  • Consumer INS Coupled with Carrier Phase Measurements for GNSS Spoofing Detection
  • Drone Carrier: An Integrated Unmanned Surface Vehicle for Autonomous Inspection and Intervention in GNSS-Denied Maritime Environment
  • PO-GVINS: Tightly Coupled GNSS-Visual-Inertial Integration with Pose-Only Representation
  • Multimodal-to-Text Prompt Engineering in Large Language Models Using Feature Embeddings for GNSS Interference Characterization
  • Wheel-GINS: A GNSS/INS Integrated Navigation System with a Wheel-mounted IMU
  • GNSS/GPS Spoofing and Jamming Identification Using Machine Learning and Deep Learning

Q: What is Graph Optimization?

A: Graph Optimization is a technique used in robotics and computer vision to enable a robot or a camera to optimize its motion and orientation using graph-based methods.

Q: What are some recent papers on Graph Optimization?

A: Some recent papers on Graph Optimization include:

  • Distributed Certifiably Correct Range-Aided SLAM
  • Impact of Temporal Delay on Radar-Inertial Odometry
  • ecg2o: A Seamless Extension of g2o for